system
stringclasses
7 values
input
stringlengths
10
94.5k
output
stringlengths
14
55.5k
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include<stdio.h> int main(){ printf("Hello World!!!"); return 0; }
.file "tmpxft_00168bba_00000000-6_hello.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Hello World!!!" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2057: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include<stdio.h> int main(){ printf("Hello World!!!"); return 0; }
#include <hip/hip_runtime.h> #include<stdio.h> int main(){ printf("Hello World!!!"); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include<stdio.h> int main(){ printf("Hello World!!!"); return 0; }
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include<stdio.h> int main(){ printf("Hello World!!!"); return 0; }
.text .file "hello.hip" .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movl $.L.str, %edi xorl %eax, %eax callq printf xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size main, .Lfunc_end0-main .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Hello World!!!" .size .L.str, 15 .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00168bba_00000000-6_hello.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Hello World!!!" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2057: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello.hip" .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movl $.L.str, %edi xorl %eax, %eax callq printf xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size main, .Lfunc_end0-main .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Hello World!!!" .size .L.str, 15 .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <cuda_runtime.h> #include <cuda.h> #include <curand.h> #include <cuda_runtime_api.h> #include <device_functions.h> #include <thrust/host_vector.h> #include <thrust/device_vector.h> #include <thrust/sort.h> #include<sys/time.h> #include <sstream> #include <iostream> #include <fstream> #include <iostream> #include <stdio.h> // printf() is only supported // for devices of compute capability 2.0 and higher #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ < 200) #define printf(f, ...) ((void)(f, __VA_ARGS__),0) #endif #define CURAND_CALL(x) do { if((x)!=CURAND_STATUS_SUCCESS) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) /// The number of points to generate within 0,1 #define NUM_BOIDS 1000 #define NUM_FRAMES 150 __device__ float vectorMag_kernal(float _vector1, float _vector2, float _vector3) { float mag; mag = sqrtf((_vector1*_vector1) + (_vector2*_vector2) + (_vector3*_vector3)); return mag; } __device__ void steerBoid_kernal(float * _targetX, float * _targetZ, float * _currentX, float * _currentZ, float * _sourceX, float *_sourceZ) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float steeringForce = 0.02; float diffX[NUM_BOIDS]; float diffZ[NUM_BOIDS]; diffX[idx] = _targetX[idx] - _currentX[idx]; diffZ[idx] = _targetZ[idx] - _currentZ[idx]; _sourceX[idx] =( (diffX[idx]/vectorMag_kernal(diffX[idx], 0, diffZ[idx]))*steeringForce); _sourceZ[idx] =( (diffZ[idx]/vectorMag_kernal(diffX[idx], 0, diffZ[idx]))*steeringForce); } __global__ void avoidBoundaries_kernal(float * _posx, float * _posz, float * _velx, float * _velz, int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float desiredVelX[NUM_BOIDS]; float desiredVelZ[NUM_BOIDS]; float * desiredVelX_ptr = &desiredVelX[0]; float * desiredVelZ_ptr = &desiredVelZ[0]; if(idx<_noBoids) { if(_posz[idx] >= 2 && _velz[idx] >0) { desiredVelX[idx] = _velx[idx]; desiredVelZ[idx] = -_velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); // _velz[idx] = -_velz[idx]; } else if(_posz[idx] <= -2 && _velz[idx] <0) { desiredVelX[idx] = _velx[idx]; desiredVelZ[idx] = -_velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velz[idx] = -_velz[idx]; } else if(_posx[idx] >= 2 && _velx[idx] >0) { desiredVelX[idx] = -_velx[idx]; desiredVelZ[idx] = _velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velx[idx] = -_velx[idx]; } else if(_posx[idx] <= -2 && _velx[idx] <0) { desiredVelX[idx] = -_velx[idx]; desiredVelZ[idx] = _velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velx[idx] = -_velx[idx]; } } } __global__ void updatePos_kernal(float * _posx, float * _posz, float * _velx, float * _velz, int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; if(idx< _noBoids) { _posx[idx] += _velx[idx]; _posz[idx] += _velz[idx]; } } __device__ float distance_kernal(float _posx, float _posz, float _otherPosx, float _otherPosz) { float distance = sqrtf(((_posx-_otherPosx)*(_posx-_otherPosx)) + ((_posz-_otherPosz)*(_posz-_otherPosz))); return distance; } __global__ void limitVel_kernal(float _limit, float * _posx, float * _posz, float * _velx, float * _velz, const int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float mag[NUM_BOIDS]; if(idx < _noBoids) { mag[idx] = sqrtf((_velx[idx]*_velx[idx]) + (_velz[idx]*_velz[idx])); if( mag[idx] > _limit) { _velx[idx] = (_velx[idx]/mag[idx])*_limit; _velz[idx] = (_velz[idx]/mag[idx])*_limit; } } } __device__ void alignment_kernal(float * _alignmentVectorX, float * _alignmentVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float tmpX[NUM_BOIDS]; float tmpZ[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.15) { atomicAdd(&(_alignmentVectorX[idx]), _velx[idy]); atomicAdd(&(_alignmentVectorZ[idx]), _velz[idy]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { //find average position _alignmentVectorX[idx] = _alignmentVectorX[idx]/numberOfNeighbours[idx]; _alignmentVectorZ[idx] = _alignmentVectorZ[idx]/numberOfNeighbours[idx]; // normalize mag[idx] = norm3d(_alignmentVectorX[idx], 0.0f, _alignmentVectorZ[idx]); if(mag[idx] > 0) { _alignmentVectorX[idx] = (_alignmentVectorX[idx] / mag[idx]); _alignmentVectorZ[idx] = (_alignmentVectorZ[idx] / mag[idx]); } //steer steerBoid_kernal(_alignmentVectorX, _alignmentVectorZ, _velx, _velz, _alignmentVectorX, _alignmentVectorZ); } } } __device__ void seperation_kernal(float * _seperationVectorX, float * _seperationVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ float _diffVectorX[NUM_BOIDS]; __shared__ float _diffVectorZ[NUM_BOIDS]; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float tmpX[NUM_BOIDS]; float tmpZ[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; _diffVectorX[idx] = 0; _diffVectorZ[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.1) { atomicAdd(&(_diffVectorX[idx]), (_posx[idy]-_posx[idx])); atomicAdd(&(_diffVectorZ[idx]), (_posz[idy]-_posz[idx])); // normalise (make atomic) //_diffVectorX[idx] = _diffVectorX[idx] / norm3d(_diffVectorX[idx], 0.0f, _diffVectorZ[idx]); //_diffVectorZ[idx] = _diffVectorZ[idx] / norm3d(_diffVectorX[idx], 0.0f, _diffVectorZ[idx]); // add neighbours position to current boids part of the seperation vector atomicAdd(&(_seperationVectorX[idx]), _diffVectorX[idx]); atomicAdd(&(_seperationVectorZ[idx]), _diffVectorZ[idx]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { // tmpX[idx] = _seperationVectorX[idx]/numberOfNeighbours[idx]; //tmpZ[idx] = _seperationVectorZ[idx]/numberOfNeighbours[idx]; //find average position _seperationVectorX[idx] = _seperationVectorX[idx]/numberOfNeighbours[idx]; _seperationVectorZ[idx] = _seperationVectorZ[idx]/numberOfNeighbours[idx]; _seperationVectorX[idx] = ( _seperationVectorX[idx] * -1); _seperationVectorZ[idx] = ( _seperationVectorZ[idx] * -1); mag[idx] = norm3d(_seperationVectorX[idx], 0.0f, _seperationVectorZ[idx]); if(mag[idx]>0) { _seperationVectorX[idx] = (_seperationVectorX[idx] / mag[idx]); _seperationVectorZ[idx] = (_seperationVectorZ[idx] / mag[idx]); } steerBoid_kernal(_seperationVectorX, _seperationVectorZ, _velx, _velz, _seperationVectorX, _seperationVectorZ); } } } __device__ void cohesion_kernal(float * _cohesionVectorX, float * _cohesionVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; _cohesionVectorX[idx] = 0; _cohesionVectorZ[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.2) { // add neighbours position to current boids part of the cohesion vector atomicAdd(&(_cohesionVectorX[idx]), _posx[idy]); atomicAdd(&(_cohesionVectorZ[idx]), _posz[idy]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { float tmpX = _cohesionVectorX[idx]/numberOfNeighbours[idx]; float tmpZ = _cohesionVectorZ[idx]/numberOfNeighbours[idx]; //find average position _cohesionVectorX[idx] = tmpX; _cohesionVectorZ[idx] = tmpZ; _cohesionVectorX[idx] = ( _cohesionVectorX[idx] - _posx[idx]); _cohesionVectorZ[idx] = ( _cohesionVectorZ[idx] - _posz[idx]); mag[idx] = norm3d(_cohesionVectorX[idx], 0.0f, _cohesionVectorZ[idx]); if(mag[idx] > 0) { _cohesionVectorX[idx] = (_cohesionVectorX[idx] / mag[idx]); _cohesionVectorZ[idx] = (_cohesionVectorZ[idx] / mag[idx]); } steerBoid_kernal(_cohesionVectorX, _cohesionVectorZ, _velx, _velz, _cohesionVectorX, _cohesionVectorZ); } } } __global__ void flock_kernal(float * _cohesionVectorX, float * _cohesionVectorZ,float * _seperationVectorX, float * _seperationVectorZ, float * _alignmentVectorX, float * _alignmentVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; uint idy = blockIdx.y * blockDim.y + threadIdx.y; float mag[NUM_BOIDS]; if( idx <_numBoids) { // calculate cohesion cohesion_kernal(_cohesionVectorX, _cohesionVectorZ, _posx, _posz, _velx, _velz, _numBoids); seperation_kernal(_seperationVectorX, _seperationVectorZ, _posx, _posz, _velx, _velz, _numBoids); alignment_kernal(_alignmentVectorX, _alignmentVectorZ, _posx, _posz, _velx, _velz, _numBoids); // wait for threads to sync (dont add cohesion vector until calculated) __syncthreads(); if(idy == 0) { _velx[idx]+= _cohesionVectorX[idx] + _seperationVectorX[idx] + _alignmentVectorX[idx]; _velz[idx]+= _cohesionVectorZ[idx] + _seperationVectorZ[idx] + _alignmentVectorZ[idx]; } } } void dumpGeo(uint _frameNumber, thrust::device_vector <float> _posX, thrust::device_vector <float> _posZ) { char fname[150]; std::sprintf(fname,"geo/flock_gpu.%03d.geo",++_frameNumber); // we will use a stringstream as it may be more efficient std::stringstream ss; std::ofstream file; file.open(fname); if (!file.is_open()) { std::cerr << "failed to Open file "<<fname<<'\n'; exit(EXIT_FAILURE); } // write header see here http://www.sidefx.com/docs/houdini15.0/io/formats/geo ss << "PGEOMETRY V5\n"; ss << "NPoints " << NUM_BOIDS << " NPrims 1\n"; ss << "NPointGroups 0 NPrimGroups 1\n"; // this is hard coded but could be flexible we have 1 attrib which is Colour ss << "NPointAttrib 1 NVertexAttrib 0 NPrimAttrib 2 NAttrib 0\n"; // now write out our point attrib this case Cd for diffuse colour ss <<"PointAttrib \n"; // default the colour to white ss <<"Cd 3 float 1 1 1\n"; // now we write out the particle data in the format // x y z 1 (attrib so in this case colour) for(unsigned int i=0; i<NUM_BOIDS; ++i) { ss<<_posX[i]<<" "<<0<<" "<<_posZ[i] << " 1 "; ss<<"("<<std::abs(1)<<" "<<std::abs(1)<<" "<<std::abs(1)<<")\n"; } // now write out the index values ss<<"PrimitiveAttrib\n"; ss<<"generator 1 index 1 location1\n"; ss<<"dopobject 1 index 1 /obj/AutoDopNetwork:1\n"; ss<<"Part "<<NUM_BOIDS<<" "; for(size_t i=0; i<NUM_BOIDS; ++i) { ss<<i<<" "; } ss<<" [0 0]\n"; ss<<"box_object1 unordered\n"; ss<<"1 1\n"; ss<<"beginExtra\n"; ss<<"endExtra\n"; // dump string stream to disk; file<<ss.rdbuf(); file.close(); } /** * Fill an array with random floats using the CURAND function. * \param devData The chunk of memory you want to fill with floats within the range (0,1] * \param n The size of the chunk of data * \author Richard Southern */ int randFloats(float *&devData, const size_t n) { // The generator, used for random numbers curandGenerator_t gen; // Create pseudo-random number generator CURAND_CALL(curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT)); // Set seed to be the current time (note that calls close together will have same seed!) CURAND_CALL(curandSetPseudoRandomGeneratorSeed(gen, time(NULL))); // Generate n floats on device CURAND_CALL(curandGenerateUniform(gen, devData, n)); // Cleanup CURAND_CALL(curandDestroyGenerator(gen)); return EXIT_SUCCESS; } int main() { // stores point pos thrust::device_vector<float> m_dBoidsPosX; thrust::device_vector<float> m_dBoidsPosZ; // stores point pos thrust::device_vector<float> m_dBoidsVelX; thrust::device_vector<float> m_dBoidsVelZ; // stores flocking vectors thrust::device_vector<float> m_dCohesionX; thrust::device_vector<float> m_dCohesionZ; thrust::device_vector<float> m_dSeperationX; thrust::device_vector<float> m_dSeperationZ; thrust::device_vector<float> m_dAlignmentX; thrust::device_vector<float> m_dAlignmentZ; //thrust::device_vector<float> d_Pos(NUM_BOIDS*3); // cant set size when constructing as member variable so resize here instead m_dBoidsPosX.resize(NUM_BOIDS); m_dBoidsPosZ.resize(NUM_BOIDS); m_dBoidsVelX.resize(NUM_BOIDS); m_dBoidsVelZ.resize(NUM_BOIDS); m_dCohesionX.resize(NUM_BOIDS); m_dCohesionZ.resize(NUM_BOIDS); m_dSeperationX.resize(NUM_BOIDS); m_dSeperationZ.resize(NUM_BOIDS); m_dAlignmentX.resize(NUM_BOIDS); m_dAlignmentZ.resize(NUM_BOIDS); // fill vector with random values for pos thrust::device_vector <float> tmp_PosPnts(NUM_BOIDS*4); float * tmp_PosPnts_ptr = thrust::raw_pointer_cast(&tmp_PosPnts[0]); randFloats(tmp_PosPnts_ptr, NUM_BOIDS*4); // give random start positions m_dBoidsPosX.assign(tmp_PosPnts.begin(), tmp_PosPnts.begin() + NUM_BOIDS); m_dBoidsPosZ.assign(tmp_PosPnts.begin() + NUM_BOIDS, tmp_PosPnts.begin() + 2*NUM_BOIDS); // give random start vel m_dBoidsVelX.assign(tmp_PosPnts.begin() + 2*NUM_BOIDS, tmp_PosPnts.begin() + 3*NUM_BOIDS); m_dBoidsVelZ.assign(tmp_PosPnts.begin() + 3*NUM_BOIDS, tmp_PosPnts.begin() + 4*NUM_BOIDS); // create pointers pointing to the device vectors float * m_dBoidsPosX_ptr= thrust::raw_pointer_cast(&m_dBoidsPosX[0]); float * m_dBoidsPosZ_ptr= thrust::raw_pointer_cast(&m_dBoidsPosZ[0]); float * m_dBoidsVelX_ptr= thrust::raw_pointer_cast(&m_dBoidsVelX[0]); float * m_dBoidsVelZ_ptr= thrust::raw_pointer_cast(&m_dBoidsVelZ[0]); float * m_dCohesionX_ptr= thrust::raw_pointer_cast(&m_dCohesionX[0]); float * m_dCohesionZ_ptr= thrust::raw_pointer_cast(&m_dCohesionZ[0]); float * m_dSeperationX_ptr= thrust::raw_pointer_cast(&m_dSeperationX[0]); float * m_dSeperationZ_ptr= thrust::raw_pointer_cast(&m_dSeperationZ[0]); float * m_dAlignmentX_ptr= thrust::raw_pointer_cast(&m_dAlignmentX[0]); float * m_dAlignmentZ_ptr= thrust::raw_pointer_cast(&m_dAlignmentZ[0]); //UPDATE----------------------------------------------------------------------------- unsigned int nThreads = 1024; unsigned int nBlocks = NUM_BOIDS/ nThreads + 1; //thrust::device_vector<unsigned int> d_numNeighbourBoids(GRID_RESOLUTION*GRID_RESOLUTION, NULL_CELL); //unsigned int * d_numNeighbourBoids_ptr = thrust::raw_pointer_cast(&d_numNeighbourBoids[0]); // for nearest neighbour unsigned int blockN = NUM_BOIDS / 32 + 1; dim3 block2(32, 32); // block of (X,Y) threads dim3 grid2(blockN, 1); // grid blockN * blockN blocks for(int i = 0; i<NUM_FRAMES; i++) { // reset vectors thrust::fill(m_dCohesionX.begin(), m_dCohesionX.begin() + NUM_BOIDS, 0); thrust::fill(m_dCohesionZ.begin(), m_dCohesionZ.begin() + NUM_BOIDS, 0); thrust::fill(m_dSeperationX.begin(), m_dSeperationX.begin() + NUM_BOIDS, 0); thrust::fill(m_dSeperationZ.begin(), m_dSeperationZ.begin() + NUM_BOIDS, 0); thrust::fill(m_dAlignmentX.begin(), m_dAlignmentX.begin() + NUM_BOIDS, 0); thrust::fill(m_dAlignmentZ.begin(), m_dAlignmentZ.begin() + NUM_BOIDS, 0); flock_kernal<<<grid2,block2>>>(m_dCohesionX_ptr, m_dCohesionZ_ptr, m_dSeperationX_ptr, m_dSeperationZ_ptr, m_dAlignmentX_ptr, m_dAlignmentZ_ptr, m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); cudaThreadSynchronize(); limitVel_kernal<<<nBlocks,nThreads>>>(0.02, m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); cudaThreadSynchronize(); avoidBoundaries_kernal<<<nBlocks,1024>>>(m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); cudaThreadSynchronize(); updatePos_kernal<<<nBlocks,1024>>>(m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); cudaThreadSynchronize(); dumpGeo(i, m_dBoidsPosX, m_dBoidsPosZ); } }
#include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #include <hip/hip_runtime_api.h> #include <hip/device_functions.h> #include <thrust/host_vector.h> #include <thrust/device_vector.h> #include <thrust/sort.h> #include<sys/time.h> #include <sstream> #include <iostream> #include <fstream> #include <iostream> #include <stdio.h> // printf() is only supported // for devices of compute capability 2.0 and higher #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ < 200) #define printf(f, ...) ((void)(f, __VA_ARGS__),0) #endif #define CURAND_CALL(x) do { if((x)!=HIPRAND_STATUS_SUCCESS) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) /// The number of points to generate within 0,1 #define NUM_BOIDS 1000 #define NUM_FRAMES 150 __device__ float vectorMag_kernal(float _vector1, float _vector2, float _vector3) { float mag; mag = sqrtf((_vector1*_vector1) + (_vector2*_vector2) + (_vector3*_vector3)); return mag; } __device__ void steerBoid_kernal(float * _targetX, float * _targetZ, float * _currentX, float * _currentZ, float * _sourceX, float *_sourceZ) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float steeringForce = 0.02; float diffX[NUM_BOIDS]; float diffZ[NUM_BOIDS]; diffX[idx] = _targetX[idx] - _currentX[idx]; diffZ[idx] = _targetZ[idx] - _currentZ[idx]; _sourceX[idx] =( (diffX[idx]/vectorMag_kernal(diffX[idx], 0, diffZ[idx]))*steeringForce); _sourceZ[idx] =( (diffZ[idx]/vectorMag_kernal(diffX[idx], 0, diffZ[idx]))*steeringForce); } __global__ void avoidBoundaries_kernal(float * _posx, float * _posz, float * _velx, float * _velz, int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float desiredVelX[NUM_BOIDS]; float desiredVelZ[NUM_BOIDS]; float * desiredVelX_ptr = &desiredVelX[0]; float * desiredVelZ_ptr = &desiredVelZ[0]; if(idx<_noBoids) { if(_posz[idx] >= 2 && _velz[idx] >0) { desiredVelX[idx] = _velx[idx]; desiredVelZ[idx] = -_velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); // _velz[idx] = -_velz[idx]; } else if(_posz[idx] <= -2 && _velz[idx] <0) { desiredVelX[idx] = _velx[idx]; desiredVelZ[idx] = -_velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velz[idx] = -_velz[idx]; } else if(_posx[idx] >= 2 && _velx[idx] >0) { desiredVelX[idx] = -_velx[idx]; desiredVelZ[idx] = _velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velx[idx] = -_velx[idx]; } else if(_posx[idx] <= -2 && _velx[idx] <0) { desiredVelX[idx] = -_velx[idx]; desiredVelZ[idx] = _velz[idx]; steerBoid_kernal(desiredVelX_ptr, desiredVelZ_ptr, _velx, _velz, _velx, _velz); //_velx[idx] = -_velx[idx]; } } } __global__ void updatePos_kernal(float * _posx, float * _posz, float * _velx, float * _velz, int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; if(idx< _noBoids) { _posx[idx] += _velx[idx]; _posz[idx] += _velz[idx]; } } __device__ float distance_kernal(float _posx, float _posz, float _otherPosx, float _otherPosz) { float distance = sqrtf(((_posx-_otherPosx)*(_posx-_otherPosx)) + ((_posz-_otherPosz)*(_posz-_otherPosz))); return distance; } __global__ void limitVel_kernal(float _limit, float * _posx, float * _posz, float * _velx, float * _velz, const int _noBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; float mag[NUM_BOIDS]; if(idx < _noBoids) { mag[idx] = sqrtf((_velx[idx]*_velx[idx]) + (_velz[idx]*_velz[idx])); if( mag[idx] > _limit) { _velx[idx] = (_velx[idx]/mag[idx])*_limit; _velz[idx] = (_velz[idx]/mag[idx])*_limit; } } } __device__ void alignment_kernal(float * _alignmentVectorX, float * _alignmentVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float tmpX[NUM_BOIDS]; float tmpZ[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.15) { atomicAdd(&(_alignmentVectorX[idx]), _velx[idy]); atomicAdd(&(_alignmentVectorZ[idx]), _velz[idy]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { //find average position _alignmentVectorX[idx] = _alignmentVectorX[idx]/numberOfNeighbours[idx]; _alignmentVectorZ[idx] = _alignmentVectorZ[idx]/numberOfNeighbours[idx]; // normalize mag[idx] = norm3d(_alignmentVectorX[idx], 0.0f, _alignmentVectorZ[idx]); if(mag[idx] > 0) { _alignmentVectorX[idx] = (_alignmentVectorX[idx] / mag[idx]); _alignmentVectorZ[idx] = (_alignmentVectorZ[idx] / mag[idx]); } //steer steerBoid_kernal(_alignmentVectorX, _alignmentVectorZ, _velx, _velz, _alignmentVectorX, _alignmentVectorZ); } } } __device__ void seperation_kernal(float * _seperationVectorX, float * _seperationVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ float _diffVectorX[NUM_BOIDS]; __shared__ float _diffVectorZ[NUM_BOIDS]; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float tmpX[NUM_BOIDS]; float tmpZ[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; _diffVectorX[idx] = 0; _diffVectorZ[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.1) { atomicAdd(&(_diffVectorX[idx]), (_posx[idy]-_posx[idx])); atomicAdd(&(_diffVectorZ[idx]), (_posz[idy]-_posz[idx])); // normalise (make atomic) //_diffVectorX[idx] = _diffVectorX[idx] / norm3d(_diffVectorX[idx], 0.0f, _diffVectorZ[idx]); //_diffVectorZ[idx] = _diffVectorZ[idx] / norm3d(_diffVectorX[idx], 0.0f, _diffVectorZ[idx]); // add neighbours position to current boids part of the seperation vector atomicAdd(&(_seperationVectorX[idx]), _diffVectorX[idx]); atomicAdd(&(_seperationVectorZ[idx]), _diffVectorZ[idx]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { // tmpX[idx] = _seperationVectorX[idx]/numberOfNeighbours[idx]; //tmpZ[idx] = _seperationVectorZ[idx]/numberOfNeighbours[idx]; //find average position _seperationVectorX[idx] = _seperationVectorX[idx]/numberOfNeighbours[idx]; _seperationVectorZ[idx] = _seperationVectorZ[idx]/numberOfNeighbours[idx]; _seperationVectorX[idx] = ( _seperationVectorX[idx] * -1); _seperationVectorZ[idx] = ( _seperationVectorZ[idx] * -1); mag[idx] = norm3d(_seperationVectorX[idx], 0.0f, _seperationVectorZ[idx]); if(mag[idx]>0) { _seperationVectorX[idx] = (_seperationVectorX[idx] / mag[idx]); _seperationVectorZ[idx] = (_seperationVectorZ[idx] / mag[idx]); } steerBoid_kernal(_seperationVectorX, _seperationVectorZ, _velx, _velz, _seperationVectorX, _seperationVectorZ); } } } __device__ void cohesion_kernal(float * _cohesionVectorX, float * _cohesionVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { int const noBoids = _numBoids; __shared__ unsigned int numberOfNeighbours[NUM_BOIDS]; float mag[NUM_BOIDS]; // current boid whos neighbours were looking for uint idx = blockIdx.x * blockDim.x + threadIdx.x; // neighbours of current boid uint idy = blockIdx.y * blockDim.y + threadIdx.y; if(idx < noBoids && idy < noBoids) { // reset values numberOfNeighbours[idx] = 0; _cohesionVectorX[idx] = 0; _cohesionVectorZ[idx] = 0; // wait for threads to sync __syncthreads(); if(idx != idy) { if(distance_kernal(_posx[idx], _posz[idx], _posx[idy], _posz[idy]) < 0.2) { // add neighbours position to current boids part of the cohesion vector atomicAdd(&(_cohesionVectorX[idx]), _posx[idy]); atomicAdd(&(_cohesionVectorZ[idx]), _posz[idy]); atomicAdd(&numberOfNeighbours[idx], 1); } } } // wait for threads to sync __syncthreads(); //limit to 1D if(idy == 0 && idx< noBoids) { //avoid dividing by zero if(numberOfNeighbours[idx] > 0) { float tmpX = _cohesionVectorX[idx]/numberOfNeighbours[idx]; float tmpZ = _cohesionVectorZ[idx]/numberOfNeighbours[idx]; //find average position _cohesionVectorX[idx] = tmpX; _cohesionVectorZ[idx] = tmpZ; _cohesionVectorX[idx] = ( _cohesionVectorX[idx] - _posx[idx]); _cohesionVectorZ[idx] = ( _cohesionVectorZ[idx] - _posz[idx]); mag[idx] = norm3d(_cohesionVectorX[idx], 0.0f, _cohesionVectorZ[idx]); if(mag[idx] > 0) { _cohesionVectorX[idx] = (_cohesionVectorX[idx] / mag[idx]); _cohesionVectorZ[idx] = (_cohesionVectorZ[idx] / mag[idx]); } steerBoid_kernal(_cohesionVectorX, _cohesionVectorZ, _velx, _velz, _cohesionVectorX, _cohesionVectorZ); } } } __global__ void flock_kernal(float * _cohesionVectorX, float * _cohesionVectorZ,float * _seperationVectorX, float * _seperationVectorZ, float * _alignmentVectorX, float * _alignmentVectorZ, float * _posx, float * _posz, float * _velx, float * _velz, int _numBoids) { uint idx = blockIdx.x * blockDim.x + threadIdx.x; uint idy = blockIdx.y * blockDim.y + threadIdx.y; float mag[NUM_BOIDS]; if( idx <_numBoids) { // calculate cohesion cohesion_kernal(_cohesionVectorX, _cohesionVectorZ, _posx, _posz, _velx, _velz, _numBoids); seperation_kernal(_seperationVectorX, _seperationVectorZ, _posx, _posz, _velx, _velz, _numBoids); alignment_kernal(_alignmentVectorX, _alignmentVectorZ, _posx, _posz, _velx, _velz, _numBoids); // wait for threads to sync (dont add cohesion vector until calculated) __syncthreads(); if(idy == 0) { _velx[idx]+= _cohesionVectorX[idx] + _seperationVectorX[idx] + _alignmentVectorX[idx]; _velz[idx]+= _cohesionVectorZ[idx] + _seperationVectorZ[idx] + _alignmentVectorZ[idx]; } } } void dumpGeo(uint _frameNumber, thrust::device_vector <float> _posX, thrust::device_vector <float> _posZ) { char fname[150]; std::sprintf(fname,"geo/flock_gpu.%03d.geo",++_frameNumber); // we will use a stringstream as it may be more efficient std::stringstream ss; std::ofstream file; file.open(fname); if (!file.is_open()) { std::cerr << "failed to Open file "<<fname<<'\n'; exit(EXIT_FAILURE); } // write header see here http://www.sidefx.com/docs/houdini15.0/io/formats/geo ss << "PGEOMETRY V5\n"; ss << "NPoints " << NUM_BOIDS << " NPrims 1\n"; ss << "NPointGroups 0 NPrimGroups 1\n"; // this is hard coded but could be flexible we have 1 attrib which is Colour ss << "NPointAttrib 1 NVertexAttrib 0 NPrimAttrib 2 NAttrib 0\n"; // now write out our point attrib this case Cd for diffuse colour ss <<"PointAttrib \n"; // default the colour to white ss <<"Cd 3 float 1 1 1\n"; // now we write out the particle data in the format // x y z 1 (attrib so in this case colour) for(unsigned int i=0; i<NUM_BOIDS; ++i) { ss<<_posX[i]<<" "<<0<<" "<<_posZ[i] << " 1 "; ss<<"("<<std::abs(1)<<" "<<std::abs(1)<<" "<<std::abs(1)<<")\n"; } // now write out the index values ss<<"PrimitiveAttrib\n"; ss<<"generator 1 index 1 location1\n"; ss<<"dopobject 1 index 1 /obj/AutoDopNetwork:1\n"; ss<<"Part "<<NUM_BOIDS<<" "; for(size_t i=0; i<NUM_BOIDS; ++i) { ss<<i<<" "; } ss<<" [0 0]\n"; ss<<"box_object1 unordered\n"; ss<<"1 1\n"; ss<<"beginExtra\n"; ss<<"endExtra\n"; // dump string stream to disk; file<<ss.rdbuf(); file.close(); } /** * Fill an array with random floats using the CURAND function. * \param devData The chunk of memory you want to fill with floats within the range (0,1] * \param n The size of the chunk of data * \author Richard Southern */ int randFloats(float *&devData, const size_t n) { // The generator, used for random numbers hiprandGenerator_t gen; // Create pseudo-random number generator CURAND_CALL(hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_DEFAULT)); // Set seed to be the current time (note that calls close together will have same seed!) CURAND_CALL(hiprandSetPseudoRandomGeneratorSeed(gen, time(NULL))); // Generate n floats on device CURAND_CALL(hiprandGenerateUniform(gen, devData, n)); // Cleanup CURAND_CALL(hiprandDestroyGenerator(gen)); return EXIT_SUCCESS; } int main() { // stores point pos thrust::device_vector<float> m_dBoidsPosX; thrust::device_vector<float> m_dBoidsPosZ; // stores point pos thrust::device_vector<float> m_dBoidsVelX; thrust::device_vector<float> m_dBoidsVelZ; // stores flocking vectors thrust::device_vector<float> m_dCohesionX; thrust::device_vector<float> m_dCohesionZ; thrust::device_vector<float> m_dSeperationX; thrust::device_vector<float> m_dSeperationZ; thrust::device_vector<float> m_dAlignmentX; thrust::device_vector<float> m_dAlignmentZ; //thrust::device_vector<float> d_Pos(NUM_BOIDS*3); // cant set size when constructing as member variable so resize here instead m_dBoidsPosX.resize(NUM_BOIDS); m_dBoidsPosZ.resize(NUM_BOIDS); m_dBoidsVelX.resize(NUM_BOIDS); m_dBoidsVelZ.resize(NUM_BOIDS); m_dCohesionX.resize(NUM_BOIDS); m_dCohesionZ.resize(NUM_BOIDS); m_dSeperationX.resize(NUM_BOIDS); m_dSeperationZ.resize(NUM_BOIDS); m_dAlignmentX.resize(NUM_BOIDS); m_dAlignmentZ.resize(NUM_BOIDS); // fill vector with random values for pos thrust::device_vector <float> tmp_PosPnts(NUM_BOIDS*4); float * tmp_PosPnts_ptr = thrust::raw_pointer_cast(&tmp_PosPnts[0]); randFloats(tmp_PosPnts_ptr, NUM_BOIDS*4); // give random start positions m_dBoidsPosX.assign(tmp_PosPnts.begin(), tmp_PosPnts.begin() + NUM_BOIDS); m_dBoidsPosZ.assign(tmp_PosPnts.begin() + NUM_BOIDS, tmp_PosPnts.begin() + 2*NUM_BOIDS); // give random start vel m_dBoidsVelX.assign(tmp_PosPnts.begin() + 2*NUM_BOIDS, tmp_PosPnts.begin() + 3*NUM_BOIDS); m_dBoidsVelZ.assign(tmp_PosPnts.begin() + 3*NUM_BOIDS, tmp_PosPnts.begin() + 4*NUM_BOIDS); // create pointers pointing to the device vectors float * m_dBoidsPosX_ptr= thrust::raw_pointer_cast(&m_dBoidsPosX[0]); float * m_dBoidsPosZ_ptr= thrust::raw_pointer_cast(&m_dBoidsPosZ[0]); float * m_dBoidsVelX_ptr= thrust::raw_pointer_cast(&m_dBoidsVelX[0]); float * m_dBoidsVelZ_ptr= thrust::raw_pointer_cast(&m_dBoidsVelZ[0]); float * m_dCohesionX_ptr= thrust::raw_pointer_cast(&m_dCohesionX[0]); float * m_dCohesionZ_ptr= thrust::raw_pointer_cast(&m_dCohesionZ[0]); float * m_dSeperationX_ptr= thrust::raw_pointer_cast(&m_dSeperationX[0]); float * m_dSeperationZ_ptr= thrust::raw_pointer_cast(&m_dSeperationZ[0]); float * m_dAlignmentX_ptr= thrust::raw_pointer_cast(&m_dAlignmentX[0]); float * m_dAlignmentZ_ptr= thrust::raw_pointer_cast(&m_dAlignmentZ[0]); //UPDATE----------------------------------------------------------------------------- unsigned int nThreads = 1024; unsigned int nBlocks = NUM_BOIDS/ nThreads + 1; //thrust::device_vector<unsigned int> d_numNeighbourBoids(GRID_RESOLUTION*GRID_RESOLUTION, NULL_CELL); //unsigned int * d_numNeighbourBoids_ptr = thrust::raw_pointer_cast(&d_numNeighbourBoids[0]); // for nearest neighbour unsigned int blockN = NUM_BOIDS / 32 + 1; dim3 block2(32, 32); // block of (X,Y) threads dim3 grid2(blockN, 1); // grid blockN * blockN blocks for(int i = 0; i<NUM_FRAMES; i++) { // reset vectors thrust::fill(m_dCohesionX.begin(), m_dCohesionX.begin() + NUM_BOIDS, 0); thrust::fill(m_dCohesionZ.begin(), m_dCohesionZ.begin() + NUM_BOIDS, 0); thrust::fill(m_dSeperationX.begin(), m_dSeperationX.begin() + NUM_BOIDS, 0); thrust::fill(m_dSeperationZ.begin(), m_dSeperationZ.begin() + NUM_BOIDS, 0); thrust::fill(m_dAlignmentX.begin(), m_dAlignmentX.begin() + NUM_BOIDS, 0); thrust::fill(m_dAlignmentZ.begin(), m_dAlignmentZ.begin() + NUM_BOIDS, 0); flock_kernal<<<grid2,block2>>>(m_dCohesionX_ptr, m_dCohesionZ_ptr, m_dSeperationX_ptr, m_dSeperationZ_ptr, m_dAlignmentX_ptr, m_dAlignmentZ_ptr, m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); hipDeviceSynchronize(); limitVel_kernal<<<nBlocks,nThreads>>>(0.02, m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); hipDeviceSynchronize(); avoidBoundaries_kernal<<<nBlocks,1024>>>(m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); hipDeviceSynchronize(); updatePos_kernal<<<nBlocks,1024>>>(m_dBoidsPosX_ptr, m_dBoidsPosZ_ptr, m_dBoidsVelX_ptr, m_dBoidsVelZ_ptr, NUM_BOIDS); hipDeviceSynchronize(); dumpGeo(i, m_dBoidsPosX, m_dBoidsPosZ); } }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "cuda.h" #include "cuda_runtime.h" #include "device_launch_parameters.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; cudaMalloc((void **)&dA, size); cudaMalloc((void **)&dB, size); cudaMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); cudaMemcpy(dA, hA, size, cudaMemcpyHostToDevice); cudaMemcpy(dB, hB, size, cudaMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); cudaMemcpy(hC2, dC, size, cudaMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free cudaFree(dA); cudaFree(dB); cudaFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
code for sm_80 Function : _Z7dMatAddPfS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R6, SR_TID.X ; /* 0x0000000000067919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe40000000a00 */ /*0040*/ UMOV UR5, 0x989680 ; /* 0x0098968000057882 */ /* 0x000fe40000000000 */ /*0050*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fca0008000000 */ /*0060*/ IMAD.WIDE R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x001fc800078e0207 */ /*0070*/ IMAD.WIDE R4, R6, R7, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x000fc800078e0207 */ /*0080*/ IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fc800078e0207 */ /*0090*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*00a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001ea4000c1e1900 */ /*00b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x004fca0000000000 */ /*00c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*00d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*00e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x000ea4000c1e1900 */ /*00f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x004fca0000000000 */ /*0100*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0110*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x000ea4000c1e1900 */ /*0130*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x004fca0000000000 */ /*0140*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0150*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ee8000c1e1900 */ /*0160*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x000ee4000c1e1900 */ /*0170*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x008fca0000000000 */ /*0180*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0190*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*01a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*01b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*01c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*01d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*01e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*01f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0200*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0210*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0220*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0230*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0240*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0250*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0260*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0270*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0280*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0290*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*02a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*02b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*02c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*02d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*02e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*02f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0300*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0310*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0320*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0330*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0340*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0350*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0360*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0370*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0380*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0390*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*03a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*03b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*03c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*03d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*03e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*03f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0400*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0410*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0420*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0430*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0440*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0450*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0460*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0470*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0480*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0490*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*04a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*04b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*04c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*04d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*04e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*04f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0500*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0510*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0520*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0530*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0540*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0550*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0560*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0570*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0580*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0590*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*05a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*05b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*05c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*05d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*05e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*05f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0600*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0610*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0620*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0630*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0640*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0650*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0660*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0670*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0680*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0690*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*06a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*06b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*06c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*06d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*06e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*06f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0700*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0710*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0720*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0730*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0740*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0750*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0760*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0770*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0780*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0790*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*07a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*07b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*07c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*07d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*07e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*07f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0800*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0001e8000c101906 */ /*0810*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0820*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f22000c1e1900 */ /*0830*/ UIADD3 UR5, UP0, UR5, -0x20, URZ ; /* 0xffffffe005057890 */ /* 0x000fc8000ff1e03f */ /*0840*/ UIADD3.X UR4, UR4, -0x1, URZ, UP0, !UPT ; /* 0xffffffff04047890 */ /* 0x000fe200087fe43f */ /*0850*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0860*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0001e8000c101906 */ /*0870*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*0880*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008ea2000c1e1900 */ /*0890*/ ISETP.NE.U32.AND P0, PT, RZ, UR5, PT ; /* 0x00000005ff007c0c */ /* 0x000fc8000bf05070 */ /*08a0*/ ISETP.NE.AND.EX P0, PT, RZ, UR4, PT, P0 ; /* 0x00000004ff007c0c */ /* 0x000fe2000bf05300 */ /*08b0*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x004fca0000000000 */ /*08c0*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0001ee000c101906 */ /*08d0*/ @P0 BRA 0x90 ; /* 0xfffff7b000000947 */ /* 0x000fea000383ffff */ /*08e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*08f0*/ BRA 0x8f0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0900*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0910*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0920*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0930*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0940*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0950*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0960*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0970*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "cuda.h" #include "cuda_runtime.h" #include "device_launch_parameters.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; cudaMalloc((void **)&dA, size); cudaMalloc((void **)&dB, size); cudaMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); cudaMemcpy(dA, hA, size, cudaMemcpyHostToDevice); cudaMemcpy(dB, hB, size, cudaMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); cudaMemcpy(hC2, dC, size, cudaMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free cudaFree(dA); cudaFree(dB); cudaFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
.file "tmpxft_00065296_00000000-6_00.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2062: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z7hMatAddPfS_S_ .type _Z7hMatAddPfS_S_, @function _Z7hMatAddPfS_S_: .LFB2057: .cfi_startproc endbr64 movq %rdx, %rcx movq %rdi, %rdx addq $4096, %rdi .L4: movl $10000000, %eax .L5: movss (%rdx), %xmm0 addss (%rsi), %xmm0 movss %xmm0, (%rcx) subq $1, %rax jne .L5 addq $4, %rdx addq $4, %rsi addq $4, %rcx cmpq %rdi, %rdx jne .L4 ret .cfi_endproc .LFE2057: .size _Z7hMatAddPfS_S_, .-_Z7hMatAddPfS_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n" .text .globl _Z7comparePfS_ .type _Z7comparePfS_, @function _Z7comparePfS_: .LFB2058: .cfi_startproc endbr64 movl $0, %edx .L12: movss (%rdi,%rdx,4), %xmm0 movss (%rsi,%rdx,4), %xmm1 ucomiss %xmm1, %xmm0 jp .L13 jne .L13 addq $1, %rdx cmpq $1024, %rdx jne .L12 movl $1, %eax ret .L13: subq $8, %rsp .cfi_def_cfa_offset 16 cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 leaq .LC0(%rip), %rsi movl $2, %edi movl $2, %eax call __printf_chk@PLT movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z7comparePfS_, .-_Z7comparePfS_ .globl _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ .type _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_, @function _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_: .LFB2084: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L22 .L18: movq 120(%rsp), %rax subq %fs:40, %rax jne .L23 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L22: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7dMatAddPfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L18 .L23: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_, .-_Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ .globl _Z7dMatAddPfS_S_ .type _Z7dMatAddPfS_S_, @function _Z7dMatAddPfS_S_: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z7dMatAddPfS_S_, .-_Z7dMatAddPfS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "true" .LC2: .string "false" .LC3: .string "Malloc error!\n" .LC4: .string "Running Host function\n" .LC5: .string "Running Device function\n" .LC7: .string "Running Compare function\n" .LC8: .string "Vectors are equal?: %s\n" .LC9: .string "CPU: %f seconds\n" .LC10: .string "GPU: %f seconds\n" .text .globl main .type main, @function main: .LFB2059: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $88, %rsp .cfi_def_cfa_offset 144 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $4096, %edi call malloc@PLT movq %rax, %rbp movl $4096, %edi call malloc@PLT movq %rax, %rbx movl $4096, %edi call malloc@PLT movq %rax, %r13 movl $4096, %edi call malloc@PLT testq %rbp, %rbp je .L27 movq %rax, %r12 testq %rbx, %rbx je .L27 testq %r13, %r13 je .L27 testq %rax, %rax je .L27 leaq 24(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 40(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT movl $0, %eax movl $1024, %ecx .L30: pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 movss %xmm0, 0(%rbp,%rax,4) movl %ecx, %edx subl %eax, %edx pxor %xmm0, %xmm0 cvtsi2ssl %edx, %xmm0 movss %xmm0, (%rbx,%rax,4) addq $1, %rax cmpq $1024, %rax jne .L30 leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call clock@PLT movq %rax, %r15 movq %r13, %rdx movq %rbx, %rsi movq %rbp, %rdi call _Z7hMatAddPfS_S_ call clock@PLT movq %rax, 8(%rsp) leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call clock@PLT movq %rax, %r14 movl $1, %ecx movl $4096, %edx movq %rbp, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4096, %edx movq %rbx, %rsi movq 32(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 60(%rsp) movl $1, 64(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $0, %r9d movl $0, %r8d movq 60(%rsp), %rdx movl $1, %ecx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L37 .L31: movl $2, %ecx movl $4096, %edx movq 40(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT call clock@PLT movq 8(%rsp), %rdx subq %r15, %rdx pxor %xmm1, %xmm1 cvtsi2sdq %rdx, %xmm1 movsd .LC6(%rip), %xmm0 divsd %xmm0, %xmm1 movq %xmm1, %r15 subq %r14, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 divsd %xmm0, %xmm1 movq %xmm1, %r14 leaq .LC7(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rsi movq %r13, %rdi call _Z7comparePfS_ testb %al, %al leaq .LC2(%rip), %rdx leaq .LC1(%rip), %rax cmovne %rax, %rdx leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r15, %xmm0 leaq .LC9(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq %r14, %xmm0 leaq .LC10(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rdi call cudaFree@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq %r13, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 72(%rsp), %rax subq %fs:40, %rax jne .L38 movl $0, %eax addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L27: .cfi_restore_state leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %edi call exit@PLT .L37: movq 40(%rsp), %rdx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ jmp .L31 .L38: call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1 .LC11: .string "_Z7dMatAddPfS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _Z7dMatAddPfS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC6: .long 0 .long 1093567616 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "cuda.h" #include "cuda_runtime.h" #include "device_launch_parameters.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; cudaMalloc((void **)&dA, size); cudaMalloc((void **)&dB, size); cudaMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); cudaMemcpy(dA, hA, size, cudaMemcpyHostToDevice); cudaMemcpy(dB, hB, size, cudaMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); cudaMemcpy(hC2, dC, size, cudaMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free cudaFree(dA); cudaFree(dB); cudaFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "hip/hip_runtime.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; hipMalloc((void **)&dA, size); hipMalloc((void **)&dB, size); hipMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); hipMemcpy(dA, hA, size, hipMemcpyHostToDevice); hipMemcpy(dB, hB, size, hipMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); hipMemcpy(hC2, dC, size, hipMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free hipFree(dA); hipFree(dB); hipFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "hip/hip_runtime.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; hipMalloc((void **)&dA, size); hipMalloc((void **)&dB, size); hipMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); hipMemcpy(dA, hA, size, hipMemcpyHostToDevice); hipMemcpy(dB, hB, size, hipMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); hipMemcpy(hC2, dC, size, hipMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free hipFree(dA); hipFree(dB); hipFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z7dMatAddPfS_S_ .globl _Z7dMatAddPfS_S_ .p2align 8 .type _Z7dMatAddPfS_S_,@function _Z7dMatAddPfS_S_: s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[2:3], s[0:1], 0x10 v_lshlrev_b32_e32 v4, 2, v0 s_mov_b64 s[0:1], 0x989680 s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v0, s4, s4, v4 v_add_co_ci_u32_e64 v1, null, s5, 0, s4 v_add_co_u32 v2, s4, s6, v4 v_add_co_u32 v4, s2, s2, v4 v_add_co_ci_u32_e64 v3, null, s7, 0, s4 v_add_co_ci_u32_e64 v5, null, s3, 0, s2 .LBB0_1: global_load_b32 v6, v[0:1], off global_load_b32 v7, v[2:3], off s_add_u32 s0, s0, -1 s_addc_u32 s1, s1, -1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u64 s[0:1], 0 s_waitcnt vmcnt(0) v_add_f32_e32 v6, v6, v7 global_store_b32 v[4:5], v6, off s_cbranch_scc0 .LBB0_1 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7dMatAddPfS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 8 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z7dMatAddPfS_S_, .Lfunc_end0-_Z7dMatAddPfS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z7dMatAddPfS_S_ .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z7dMatAddPfS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 8 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "hip/hip_runtime.h" #define N_THREADS 1024 #define LOOP_TIMES 10000000 /* The functions run too fast to see the difference, so to compare we run then in a loop. In my environment */ /** * Sum numbers of two vectors on GPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ __global__ void dMatAdd(float *A, float *B, float *C) { int i = threadIdx.x; for (long j = 0; j < LOOP_TIMES; j++) C[i] = A[i] + B[i]; } /** * Sum numbers of two vectors on CPU. * * The result is stored on a third vector. * * @param A First input vector pointer. * @param B Second input vector pointer. * @param C Output vector pointer. * @return void */ void hMatAdd(float *A, float *B, float *C) { for (int i = 0; i < N_THREADS; i++) { for (long j = 0; j < LOOP_TIMES; j++) { C[i] = A[i] + B[i]; } } } /** * Compare two float vectors * * @param A First input vector pointer. * @param B Second input vector pointer. * @return true if vectors are equal, false otherwise */ bool compare(float *A, float *B) { for (int i = 0; i < N_THREADS; i++) { if (A[i] != B[i]) { printf( "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n", i, A[i], B[i]); return false; } } return true; } int main(int argc, char const *argv[]) { // Timing variables clock_t h_begin, h_end, d_begin, d_end; double h_time, d_time; // Size auxiliar size_t size = N_THREADS * sizeof(float); // Host memory allocation float *hA = (float *)malloc(size); float *hB = (float *)malloc(size); float *hC = (float *)malloc(size); float *hC2 = (float *)malloc(size); if (hA == NULL || hB == NULL || hC == NULL || hC2 == NULL) { printf("Malloc error!\n"); exit(1); } // Device memory allocation float *dA, *dB, *dC; hipMalloc((void **)&dA, size); hipMalloc((void **)&dB, size); hipMalloc((void **)&dC, size); // Vectors initialization with some values for (int i = 0, j = N_THREADS; i < N_THREADS; i++, j--) { hA[i] = float(i); //printf("hA[%d] = %f\n", i, hA[i]); hB[i] = float(j); } // Host function printf("Running Host function\n"); h_begin = clock(); hMatAdd(hA, hB, hC); h_end = clock(); // Device function printf("Running Device function\n"); d_begin = clock(); hipMemcpy(dA, hA, size, hipMemcpyHostToDevice); hipMemcpy(dB, hB, size, hipMemcpyHostToDevice); dMatAdd<<<1, N_THREADS>>>(dA, dB, dC); hipMemcpy(hC2, dC, size, hipMemcpyDeviceToHost); d_end = clock(); // Results h_time = (double)(h_end - h_begin) / CLOCKS_PER_SEC; d_time = (double)(d_end - d_begin) / CLOCKS_PER_SEC; printf("Running Compare function\n"); printf("Vectors are equal?: %s\n", compare(hC, hC2) ? "true" : "false"); printf("CPU: %f seconds\n", h_time); printf("GPU: %f seconds\n", d_time); // Memory free hipFree(dA); hipFree(dB); hipFree(dC); free(hA); free(hB); free(hC); free(hC2); return 0; }
.text .file "00.hip" .globl _Z22__device_stub__dMatAddPfS_S_ # -- Begin function _Z22__device_stub__dMatAddPfS_S_ .p2align 4, 0x90 .type _Z22__device_stub__dMatAddPfS_S_,@function _Z22__device_stub__dMatAddPfS_S_: # @_Z22__device_stub__dMatAddPfS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7dMatAddPfS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z22__device_stub__dMatAddPfS_S_, .Lfunc_end0-_Z22__device_stub__dMatAddPfS_S_ .cfi_endproc # -- End function .globl _Z7hMatAddPfS_S_ # -- Begin function _Z7hMatAddPfS_S_ .p2align 4, 0x90 .type _Z7hMatAddPfS_S_,@function _Z7hMatAddPfS_S_: # @_Z7hMatAddPfS_S_ .cfi_startproc # %bb.0: xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_2 Depth 2 movl $10000000, %ecx # imm = 0x989680 .p2align 4, 0x90 .LBB1_2: # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 movss (%rdi,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss (%rsi,%rax,4), %xmm0 movss %xmm0, (%rdx,%rax,4) decq %rcx jne .LBB1_2 # %bb.3: # in Loop: Header=BB1_1 Depth=1 incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB1_1 # %bb.4: retq .Lfunc_end1: .size _Z7hMatAddPfS_S_, .Lfunc_end1-_Z7hMatAddPfS_S_ .cfi_endproc # -- End function .globl _Z7comparePfS_ # -- Begin function _Z7comparePfS_ .p2align 4, 0x90 .type _Z7comparePfS_,@function _Z7comparePfS_: # @_Z7comparePfS_ .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 movq %rsi, %rax movss (%rdi), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%rsi), %xmm1 # xmm1 = mem[0],zero,zero,zero xorl %esi, %esi ucomiss %xmm1, %xmm0 movl $0, %ebx jne .LBB2_5 jp .LBB2_5 # %bb.1: # %.lr.ph.preheader xorl %esi, %esi .p2align 4, 0x90 .LBB2_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 cmpq $1023, %rsi # imm = 0x3FF je .LBB2_6 # %bb.3: # in Loop: Header=BB2_2 Depth=1 movss 4(%rdi,%rsi,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss 4(%rax,%rsi,4), %xmm1 # xmm1 = mem[0],zero,zero,zero incq %rsi ucomiss %xmm1, %xmm0 jne .LBB2_4 jnp .LBB2_2 .LBB2_4: # %._crit_edge leaq -1(%rsi), %rax cmpq $1023, %rax # imm = 0x3FF setae %bl .LBB2_5: cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi # kill: def $esi killed $esi killed $rsi movb $2, %al callq printf .LBB2_7: # %.loopexit movl %ebx, %eax popq %rbx .cfi_def_cfa_offset 8 retq .LBB2_6: # %.loopexit.loopexit .cfi_def_cfa_offset 16 setae %bl jmp .LBB2_7 .Lfunc_end2: .size _Z7comparePfS_, .Lfunc_end2-_Z7comparePfS_ .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI3_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $152, %rsp .cfi_def_cfa_offset 208 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %rbx movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r14 movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r15 movl $4096, %edi # imm = 0x1000 callq malloc testq %rbx, %rbx je .LBB3_18 # %bb.1: testq %r14, %r14 je .LBB3_18 # %bb.2: testq %r15, %r15 je .LBB3_18 # %bb.3: testq %rax, %rax je .LBB3_18 # %bb.4: movq %rax, 16(%rsp) # 8-byte Spill leaq 40(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 32(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 24(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc movl $1024, %eax # imm = 0x400 xorl %ecx, %ecx .p2align 4, 0x90 .LBB3_5: # =>This Inner Loop Header: Depth=1 xorps %xmm0, %xmm0 cvtsi2ss %ecx, %xmm0 movss %xmm0, (%rbx,%rcx,4) xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 movss %xmm0, (%r14,%rcx,4) incq %rcx decl %eax cmpq $1024, %rcx # imm = 0x400 jne .LBB3_5 # %bb.6: movl $.Lstr, %edi callq puts@PLT xorl %r12d, %r12d callq clock movq %rax, 8(%rsp) # 8-byte Spill .p2align 4, 0x90 .LBB3_7: # %.preheader.i # =>This Inner Loop Header: Depth=1 movss (%rbx,%r12,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss (%r14,%r12,4), %xmm0 movss %xmm0, (%r15,%r12,4) incq %r12 cmpq $1024, %r12 # imm = 0x400 jne .LBB3_7 # %bb.8: # %_Z7hMatAddPfS_S_.exit callq clock movq %rax, 48(%rsp) # 8-byte Spill movl $.Lstr.1, %edi callq puts@PLT callq clock movq %rax, (%rsp) # 8-byte Spill movq 40(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 32(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdi # imm = 0x100000001 leaq 1023(%rdi), %rdx xorl %r13d, %r13d movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_10 # %bb.9: movq 40(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 120(%rsp) movq %rcx, 112(%rsp) movq %rdx, 104(%rsp) leaq 120(%rsp), %rax movq %rax, 128(%rsp) leaq 112(%rsp), %rax movq %rax, 136(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 88(%rsp), %rdi leaq 72(%rsp), %rsi leaq 64(%rsp), %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 88(%rsp), %rsi movl 96(%rsp), %edx movq 72(%rsp), %rcx movl 80(%rsp), %r8d leaq 128(%rsp), %r9 movl $_Z7dMatAddPfS_S_, %edi pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 72(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_10: movq 24(%rsp), %rsi movl $4096, %edx # imm = 0x1000 movq 16(%rsp), %r12 # 8-byte Reload movq %r12, %rdi movl $2, %ecx callq hipMemcpy callq clock movq %rax, %rbp movl $.Lstr.2, %edi callq puts@PLT movss (%r15), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%r12), %xmm1 # xmm1 = mem[0],zero,zero,zero ucomiss %xmm1, %xmm0 movl $0, %r12d jne .LBB3_15 jp .LBB3_15 # %bb.11: # %.lr.ph.preheader xorl %r13d, %r13d movq 16(%rsp), %rax # 8-byte Reload .p2align 4, 0x90 .LBB3_12: # %.lr.ph # =>This Inner Loop Header: Depth=1 cmpq $1023, %r13 # imm = 0x3FF je .LBB3_16 # %bb.13: # in Loop: Header=BB3_12 Depth=1 movss 4(%r15,%r13,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss 4(%rax,%r13,4), %xmm1 # xmm1 = mem[0],zero,zero,zero incq %r13 ucomiss %xmm1, %xmm0 jne .LBB3_14 jnp .LBB3_12 .LBB3_14: # %._crit_edge.loopexit leaq -1(%r13), %rax cmpq $1023, %rax # imm = 0x3FF setae %r12b .LBB3_15: # %._crit_edge cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi movl %r13d, %esi movb $2, %al callq printf .LBB3_17: # %_Z7comparePfS_.exit subq (%rsp), %rbp # 8-byte Folded Reload xorps %xmm1, %xmm1 cvtsi2sd %rbp, %xmm1 movsd .LCPI3_0(%rip), %xmm0 # xmm0 = mem[0],zero divsd %xmm0, %xmm1 movsd %xmm1, (%rsp) # 8-byte Spill movq 48(%rsp), %rax # 8-byte Reload subq 8(%rsp), %rax # 8-byte Folded Reload xorps %xmm1, %xmm1 cvtsi2sd %rax, %xmm1 divsd %xmm0, %xmm1 movsd %xmm1, 8(%rsp) # 8-byte Spill movl $.L.str.6, %eax movl $.L.str.7, %esi testb %r12b, %r12b cmovneq %rax, %rsi movl $.L.str.5, %edi xorl %eax, %eax callq printf movl $.L.str.8, %edi movsd 8(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movb $1, %al callq printf movl $.L.str.9, %edi movsd (%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movb $1, %al callq printf movq 40(%rsp), %rdi callq hipFree movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free movq 16(%rsp), %rdi # 8-byte Reload callq free xorl %eax, %eax addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB3_16: # %_Z7comparePfS_.exit.loopexit .cfi_def_cfa_offset 208 setae %r12b jmp .LBB3_17 .LBB3_18: movl $.Lstr.3, %edi callq puts@PLT movl $1, %edi callq exit .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7dMatAddPfS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type _Z7dMatAddPfS_S_,@object # @_Z7dMatAddPfS_S_ .section .rodata,"a",@progbits .globl _Z7dMatAddPfS_S_ .p2align 3, 0x0 _Z7dMatAddPfS_S_: .quad _Z22__device_stub__dMatAddPfS_S_ .size _Z7dMatAddPfS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n" .size .L.str, 47 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "Vectors are equal?: %s\n" .size .L.str.5, 24 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "true" .size .L.str.6, 5 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "false" .size .L.str.7, 6 .type .L.str.8,@object # @.str.8 .L.str.8: .asciz "CPU: %f seconds\n" .size .L.str.8, 17 .type .L.str.9,@object # @.str.9 .L.str.9: .asciz "GPU: %f seconds\n" .size .L.str.9, 17 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7dMatAddPfS_S_" .size .L__unnamed_1, 17 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Running Host function" .size .Lstr, 22 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Running Device function" .size .Lstr.1, 24 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Running Compare function" .size .Lstr.2, 25 .type .Lstr.3,@object # @str.3 .Lstr.3: .asciz "Malloc error!" .size .Lstr.3, 14 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__dMatAddPfS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7dMatAddPfS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z7dMatAddPfS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R6, SR_TID.X ; /* 0x0000000000067919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe40000000a00 */ /*0040*/ UMOV UR5, 0x989680 ; /* 0x0098968000057882 */ /* 0x000fe40000000000 */ /*0050*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fca0008000000 */ /*0060*/ IMAD.WIDE R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x001fc800078e0207 */ /*0070*/ IMAD.WIDE R4, R6, R7, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x000fc800078e0207 */ /*0080*/ IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fc800078e0207 */ /*0090*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*00a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001ea4000c1e1900 */ /*00b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x004fca0000000000 */ /*00c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*00d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*00e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x000ea4000c1e1900 */ /*00f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x004fca0000000000 */ /*0100*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0110*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x000ea4000c1e1900 */ /*0130*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x004fca0000000000 */ /*0140*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0150*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ee8000c1e1900 */ /*0160*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x000ee4000c1e1900 */ /*0170*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x008fca0000000000 */ /*0180*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0190*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*01a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*01b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*01c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*01d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*01e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*01f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0200*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0210*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0220*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0230*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0240*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0250*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0260*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0270*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0280*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0290*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*02a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*02b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*02c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*02d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*02e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*02f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0300*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0310*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0320*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0330*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0340*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0350*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0360*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0370*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0380*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0390*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*03a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*03b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*03c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*03d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*03e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*03f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0400*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0410*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0420*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0430*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0440*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0450*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0460*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0470*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0480*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0490*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*04a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*04b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*04c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*04d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*04e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*04f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0500*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0510*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0520*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0530*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0540*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0550*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0560*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0570*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0580*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0590*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*05a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*05b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*05c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*05d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*05e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*05f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0600*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0610*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0620*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0630*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0640*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0650*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0660*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0670*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0680*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0690*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*06a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*06b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*06c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*06d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*06e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*06f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0700*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0003e8000c101906 */ /*0710*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0720*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f24000c1e1900 */ /*0730*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0740*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0005e8000c101906 */ /*0750*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0760*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008f24000c1e1900 */ /*0770*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x010fca0000000000 */ /*0780*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0007e8000c101906 */ /*0790*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*07a0*/ LDG.E R9, [R2.64] ; /* 0x0000000602097981 */ /* 0x001f24000c1e1900 */ /*07b0*/ FADD R9, R0, R9 ; /* 0x0000000900097221 */ /* 0x010fca0000000000 */ /*07c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x0001e8000c101906 */ /*07d0*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*07e0*/ LDG.E R11, [R2.64] ; /* 0x00000006020b7981 */ /* 0x002f24000c1e1900 */ /*07f0*/ FADD R11, R0, R11 ; /* 0x0000000b000b7221 */ /* 0x010fca0000000000 */ /*0800*/ STG.E [R6.64], R11 ; /* 0x0000000b06007986 */ /* 0x0001e8000c101906 */ /*0810*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000f28000c1e1900 */ /*0820*/ LDG.E R13, [R2.64] ; /* 0x00000006020d7981 */ /* 0x004f22000c1e1900 */ /*0830*/ UIADD3 UR5, UP0, UR5, -0x20, URZ ; /* 0xffffffe005057890 */ /* 0x000fc8000ff1e03f */ /*0840*/ UIADD3.X UR4, UR4, -0x1, URZ, UP0, !UPT ; /* 0xffffffff04047890 */ /* 0x000fe200087fe43f */ /*0850*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x010fca0000000000 */ /*0860*/ STG.E [R6.64], R13 ; /* 0x0000000d06007986 */ /* 0x0001e8000c101906 */ /*0870*/ LDG.E R0, [R4.64] ; /* 0x0000000604007981 */ /* 0x000ea8000c1e1900 */ /*0880*/ LDG.E R15, [R2.64] ; /* 0x00000006020f7981 */ /* 0x008ea2000c1e1900 */ /*0890*/ ISETP.NE.U32.AND P0, PT, RZ, UR5, PT ; /* 0x00000005ff007c0c */ /* 0x000fc8000bf05070 */ /*08a0*/ ISETP.NE.AND.EX P0, PT, RZ, UR4, PT, P0 ; /* 0x00000004ff007c0c */ /* 0x000fe2000bf05300 */ /*08b0*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x004fca0000000000 */ /*08c0*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x0001ee000c101906 */ /*08d0*/ @P0 BRA 0x90 ; /* 0xfffff7b000000947 */ /* 0x000fea000383ffff */ /*08e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*08f0*/ BRA 0x8f0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0900*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0910*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0920*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0930*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0940*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0950*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0960*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0970*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z7dMatAddPfS_S_ .globl _Z7dMatAddPfS_S_ .p2align 8 .type _Z7dMatAddPfS_S_,@function _Z7dMatAddPfS_S_: s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[2:3], s[0:1], 0x10 v_lshlrev_b32_e32 v4, 2, v0 s_mov_b64 s[0:1], 0x989680 s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v0, s4, s4, v4 v_add_co_ci_u32_e64 v1, null, s5, 0, s4 v_add_co_u32 v2, s4, s6, v4 v_add_co_u32 v4, s2, s2, v4 v_add_co_ci_u32_e64 v3, null, s7, 0, s4 v_add_co_ci_u32_e64 v5, null, s3, 0, s2 .LBB0_1: global_load_b32 v6, v[0:1], off global_load_b32 v7, v[2:3], off s_add_u32 s0, s0, -1 s_addc_u32 s1, s1, -1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u64 s[0:1], 0 s_waitcnt vmcnt(0) v_add_f32_e32 v6, v6, v7 global_store_b32 v[4:5], v6, off s_cbranch_scc0 .LBB0_1 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7dMatAddPfS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 8 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z7dMatAddPfS_S_, .Lfunc_end0-_Z7dMatAddPfS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z7dMatAddPfS_S_ .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z7dMatAddPfS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 8 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00065296_00000000-6_00.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2062: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z7hMatAddPfS_S_ .type _Z7hMatAddPfS_S_, @function _Z7hMatAddPfS_S_: .LFB2057: .cfi_startproc endbr64 movq %rdx, %rcx movq %rdi, %rdx addq $4096, %rdi .L4: movl $10000000, %eax .L5: movss (%rdx), %xmm0 addss (%rsi), %xmm0 movss %xmm0, (%rcx) subq $1, %rax jne .L5 addq $4, %rdx addq $4, %rsi addq $4, %rcx cmpq %rdi, %rdx jne .L4 ret .cfi_endproc .LFE2057: .size _Z7hMatAddPfS_S_, .-_Z7hMatAddPfS_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n" .text .globl _Z7comparePfS_ .type _Z7comparePfS_, @function _Z7comparePfS_: .LFB2058: .cfi_startproc endbr64 movl $0, %edx .L12: movss (%rdi,%rdx,4), %xmm0 movss (%rsi,%rdx,4), %xmm1 ucomiss %xmm1, %xmm0 jp .L13 jne .L13 addq $1, %rdx cmpq $1024, %rdx jne .L12 movl $1, %eax ret .L13: subq $8, %rsp .cfi_def_cfa_offset 16 cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 leaq .LC0(%rip), %rsi movl $2, %edi movl $2, %eax call __printf_chk@PLT movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z7comparePfS_, .-_Z7comparePfS_ .globl _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ .type _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_, @function _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_: .LFB2084: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L22 .L18: movq 120(%rsp), %rax subq %fs:40, %rax jne .L23 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L22: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7dMatAddPfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L18 .L23: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_, .-_Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ .globl _Z7dMatAddPfS_S_ .type _Z7dMatAddPfS_S_, @function _Z7dMatAddPfS_S_: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z7dMatAddPfS_S_, .-_Z7dMatAddPfS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "true" .LC2: .string "false" .LC3: .string "Malloc error!\n" .LC4: .string "Running Host function\n" .LC5: .string "Running Device function\n" .LC7: .string "Running Compare function\n" .LC8: .string "Vectors are equal?: %s\n" .LC9: .string "CPU: %f seconds\n" .LC10: .string "GPU: %f seconds\n" .text .globl main .type main, @function main: .LFB2059: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $88, %rsp .cfi_def_cfa_offset 144 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $4096, %edi call malloc@PLT movq %rax, %rbp movl $4096, %edi call malloc@PLT movq %rax, %rbx movl $4096, %edi call malloc@PLT movq %rax, %r13 movl $4096, %edi call malloc@PLT testq %rbp, %rbp je .L27 movq %rax, %r12 testq %rbx, %rbx je .L27 testq %r13, %r13 je .L27 testq %rax, %rax je .L27 leaq 24(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 40(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT movl $0, %eax movl $1024, %ecx .L30: pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 movss %xmm0, 0(%rbp,%rax,4) movl %ecx, %edx subl %eax, %edx pxor %xmm0, %xmm0 cvtsi2ssl %edx, %xmm0 movss %xmm0, (%rbx,%rax,4) addq $1, %rax cmpq $1024, %rax jne .L30 leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call clock@PLT movq %rax, %r15 movq %r13, %rdx movq %rbx, %rsi movq %rbp, %rdi call _Z7hMatAddPfS_S_ call clock@PLT movq %rax, 8(%rsp) leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call clock@PLT movq %rax, %r14 movl $1, %ecx movl $4096, %edx movq %rbp, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4096, %edx movq %rbx, %rsi movq 32(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 60(%rsp) movl $1, 64(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $0, %r9d movl $0, %r8d movq 60(%rsp), %rdx movl $1, %ecx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L37 .L31: movl $2, %ecx movl $4096, %edx movq 40(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT call clock@PLT movq 8(%rsp), %rdx subq %r15, %rdx pxor %xmm1, %xmm1 cvtsi2sdq %rdx, %xmm1 movsd .LC6(%rip), %xmm0 divsd %xmm0, %xmm1 movq %xmm1, %r15 subq %r14, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 divsd %xmm0, %xmm1 movq %xmm1, %r14 leaq .LC7(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rsi movq %r13, %rdi call _Z7comparePfS_ testb %al, %al leaq .LC2(%rip), %rdx leaq .LC1(%rip), %rax cmovne %rax, %rdx leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r15, %xmm0 leaq .LC9(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq %r14, %xmm0 leaq .LC10(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rdi call cudaFree@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq %r13, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 72(%rsp), %rax subq %fs:40, %rax jne .L38 movl $0, %eax addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L27: .cfi_restore_state leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %edi call exit@PLT .L37: movq 40(%rsp), %rdx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z30__device_stub__Z7dMatAddPfS_S_PfS_S_ jmp .L31 .L38: call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1 .LC11: .string "_Z7dMatAddPfS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _Z7dMatAddPfS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC6: .long 0 .long 1093567616 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "00.hip" .globl _Z22__device_stub__dMatAddPfS_S_ # -- Begin function _Z22__device_stub__dMatAddPfS_S_ .p2align 4, 0x90 .type _Z22__device_stub__dMatAddPfS_S_,@function _Z22__device_stub__dMatAddPfS_S_: # @_Z22__device_stub__dMatAddPfS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7dMatAddPfS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z22__device_stub__dMatAddPfS_S_, .Lfunc_end0-_Z22__device_stub__dMatAddPfS_S_ .cfi_endproc # -- End function .globl _Z7hMatAddPfS_S_ # -- Begin function _Z7hMatAddPfS_S_ .p2align 4, 0x90 .type _Z7hMatAddPfS_S_,@function _Z7hMatAddPfS_S_: # @_Z7hMatAddPfS_S_ .cfi_startproc # %bb.0: xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_2 Depth 2 movl $10000000, %ecx # imm = 0x989680 .p2align 4, 0x90 .LBB1_2: # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 movss (%rdi,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss (%rsi,%rax,4), %xmm0 movss %xmm0, (%rdx,%rax,4) decq %rcx jne .LBB1_2 # %bb.3: # in Loop: Header=BB1_1 Depth=1 incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB1_1 # %bb.4: retq .Lfunc_end1: .size _Z7hMatAddPfS_S_, .Lfunc_end1-_Z7hMatAddPfS_S_ .cfi_endproc # -- End function .globl _Z7comparePfS_ # -- Begin function _Z7comparePfS_ .p2align 4, 0x90 .type _Z7comparePfS_,@function _Z7comparePfS_: # @_Z7comparePfS_ .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 movq %rsi, %rax movss (%rdi), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%rsi), %xmm1 # xmm1 = mem[0],zero,zero,zero xorl %esi, %esi ucomiss %xmm1, %xmm0 movl $0, %ebx jne .LBB2_5 jp .LBB2_5 # %bb.1: # %.lr.ph.preheader xorl %esi, %esi .p2align 4, 0x90 .LBB2_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 cmpq $1023, %rsi # imm = 0x3FF je .LBB2_6 # %bb.3: # in Loop: Header=BB2_2 Depth=1 movss 4(%rdi,%rsi,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss 4(%rax,%rsi,4), %xmm1 # xmm1 = mem[0],zero,zero,zero incq %rsi ucomiss %xmm1, %xmm0 jne .LBB2_4 jnp .LBB2_2 .LBB2_4: # %._crit_edge leaq -1(%rsi), %rax cmpq $1023, %rax # imm = 0x3FF setae %bl .LBB2_5: cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi # kill: def $esi killed $esi killed $rsi movb $2, %al callq printf .LBB2_7: # %.loopexit movl %ebx, %eax popq %rbx .cfi_def_cfa_offset 8 retq .LBB2_6: # %.loopexit.loopexit .cfi_def_cfa_offset 16 setae %bl jmp .LBB2_7 .Lfunc_end2: .size _Z7comparePfS_, .Lfunc_end2-_Z7comparePfS_ .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI3_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $152, %rsp .cfi_def_cfa_offset 208 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %rbx movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r14 movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r15 movl $4096, %edi # imm = 0x1000 callq malloc testq %rbx, %rbx je .LBB3_18 # %bb.1: testq %r14, %r14 je .LBB3_18 # %bb.2: testq %r15, %r15 je .LBB3_18 # %bb.3: testq %rax, %rax je .LBB3_18 # %bb.4: movq %rax, 16(%rsp) # 8-byte Spill leaq 40(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 32(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 24(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc movl $1024, %eax # imm = 0x400 xorl %ecx, %ecx .p2align 4, 0x90 .LBB3_5: # =>This Inner Loop Header: Depth=1 xorps %xmm0, %xmm0 cvtsi2ss %ecx, %xmm0 movss %xmm0, (%rbx,%rcx,4) xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 movss %xmm0, (%r14,%rcx,4) incq %rcx decl %eax cmpq $1024, %rcx # imm = 0x400 jne .LBB3_5 # %bb.6: movl $.Lstr, %edi callq puts@PLT xorl %r12d, %r12d callq clock movq %rax, 8(%rsp) # 8-byte Spill .p2align 4, 0x90 .LBB3_7: # %.preheader.i # =>This Inner Loop Header: Depth=1 movss (%rbx,%r12,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss (%r14,%r12,4), %xmm0 movss %xmm0, (%r15,%r12,4) incq %r12 cmpq $1024, %r12 # imm = 0x400 jne .LBB3_7 # %bb.8: # %_Z7hMatAddPfS_S_.exit callq clock movq %rax, 48(%rsp) # 8-byte Spill movl $.Lstr.1, %edi callq puts@PLT callq clock movq %rax, (%rsp) # 8-byte Spill movq 40(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 32(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdi # imm = 0x100000001 leaq 1023(%rdi), %rdx xorl %r13d, %r13d movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_10 # %bb.9: movq 40(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 120(%rsp) movq %rcx, 112(%rsp) movq %rdx, 104(%rsp) leaq 120(%rsp), %rax movq %rax, 128(%rsp) leaq 112(%rsp), %rax movq %rax, 136(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 88(%rsp), %rdi leaq 72(%rsp), %rsi leaq 64(%rsp), %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 88(%rsp), %rsi movl 96(%rsp), %edx movq 72(%rsp), %rcx movl 80(%rsp), %r8d leaq 128(%rsp), %r9 movl $_Z7dMatAddPfS_S_, %edi pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 72(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_10: movq 24(%rsp), %rsi movl $4096, %edx # imm = 0x1000 movq 16(%rsp), %r12 # 8-byte Reload movq %r12, %rdi movl $2, %ecx callq hipMemcpy callq clock movq %rax, %rbp movl $.Lstr.2, %edi callq puts@PLT movss (%r15), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%r12), %xmm1 # xmm1 = mem[0],zero,zero,zero ucomiss %xmm1, %xmm0 movl $0, %r12d jne .LBB3_15 jp .LBB3_15 # %bb.11: # %.lr.ph.preheader xorl %r13d, %r13d movq 16(%rsp), %rax # 8-byte Reload .p2align 4, 0x90 .LBB3_12: # %.lr.ph # =>This Inner Loop Header: Depth=1 cmpq $1023, %r13 # imm = 0x3FF je .LBB3_16 # %bb.13: # in Loop: Header=BB3_12 Depth=1 movss 4(%r15,%r13,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss 4(%rax,%r13,4), %xmm1 # xmm1 = mem[0],zero,zero,zero incq %r13 ucomiss %xmm1, %xmm0 jne .LBB3_14 jnp .LBB3_12 .LBB3_14: # %._crit_edge.loopexit leaq -1(%r13), %rax cmpq $1023, %rax # imm = 0x3FF setae %r12b .LBB3_15: # %._crit_edge cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi movl %r13d, %esi movb $2, %al callq printf .LBB3_17: # %_Z7comparePfS_.exit subq (%rsp), %rbp # 8-byte Folded Reload xorps %xmm1, %xmm1 cvtsi2sd %rbp, %xmm1 movsd .LCPI3_0(%rip), %xmm0 # xmm0 = mem[0],zero divsd %xmm0, %xmm1 movsd %xmm1, (%rsp) # 8-byte Spill movq 48(%rsp), %rax # 8-byte Reload subq 8(%rsp), %rax # 8-byte Folded Reload xorps %xmm1, %xmm1 cvtsi2sd %rax, %xmm1 divsd %xmm0, %xmm1 movsd %xmm1, 8(%rsp) # 8-byte Spill movl $.L.str.6, %eax movl $.L.str.7, %esi testb %r12b, %r12b cmovneq %rax, %rsi movl $.L.str.5, %edi xorl %eax, %eax callq printf movl $.L.str.8, %edi movsd 8(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movb $1, %al callq printf movl $.L.str.9, %edi movsd (%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movb $1, %al callq printf movq 40(%rsp), %rdi callq hipFree movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free movq 16(%rsp), %rdi # 8-byte Reload callq free xorl %eax, %eax addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB3_16: # %_Z7comparePfS_.exit.loopexit .cfi_def_cfa_offset 208 setae %r12b jmp .LBB3_17 .LBB3_18: movl $.Lstr.3, %edi callq puts@PLT movl $1, %edi callq exit .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7dMatAddPfS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type _Z7dMatAddPfS_S_,@object # @_Z7dMatAddPfS_S_ .section .rodata,"a",@progbits .globl _Z7dMatAddPfS_S_ .p2align 3, 0x0 _Z7dMatAddPfS_S_: .quad _Z22__device_stub__dMatAddPfS_S_ .size _Z7dMatAddPfS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Elements are not equal. Index %d\t\tA: %f\t\tB:%f\n" .size .L.str, 47 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "Vectors are equal?: %s\n" .size .L.str.5, 24 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "true" .size .L.str.6, 5 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "false" .size .L.str.7, 6 .type .L.str.8,@object # @.str.8 .L.str.8: .asciz "CPU: %f seconds\n" .size .L.str.8, 17 .type .L.str.9,@object # @.str.9 .L.str.9: .asciz "GPU: %f seconds\n" .size .L.str.9, 17 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7dMatAddPfS_S_" .size .L__unnamed_1, 17 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Running Host function" .size .Lstr, 22 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Running Device function" .size .Lstr.1, 24 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Running Compare function" .size .Lstr.2, 25 .type .Lstr.3,@object # @str.3 .Lstr.3: .asciz "Malloc error!" .size .Lstr.3, 14 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__dMatAddPfS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7dMatAddPfS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
code for sm_80 Function : _Z13StarRadKernelPdS_S_diiS_S_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0030*/ S2R R11, SR_CTAID.Y ; /* 0x00000000000b7919 */ /* 0x000e680000002600 */ /*0040*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x184], PT ; /* 0x0000610000007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R11, R11, c[0x0][0x4], R2 ; /* 0x000001000b0b7a24 */ /* 0x002fca00078e0202 */ /*0080*/ ISETP.GE.OR P0, PT, R11, c[0x0][0x180], P0 ; /* 0x000060000b007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*00b0*/ IMAD.MOV.U32 R9, RZ, RZ, 0x8 ; /* 0x00000008ff097424 */ /* 0x000fe200078e00ff */ /*00c0*/ UIADD3 UR4, UR4, -0x1, URZ ; /* 0xffffffff04047890 */ /* 0x000fe2000fffe03f */ /*00d0*/ IMAD R2, R0, c[0x0][0x180], R11 ; /* 0x0000600000027a24 */ /* 0x000fc800078e020b */ /*00e0*/ IMAD.WIDE R2, R2, R9, c[0x0][0x198] ; /* 0x0000660002027625 */ /* 0x000fe200078e0209 */ /*00f0*/ ISETP.NE.AND P0, PT, R11, UR4, PT ; /* 0x000000040b007c0c */ /* 0x000fe2000bf05270 */ /*0100*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0110*/ ISETP.EQ.OR P0, PT, R11, RZ, !P0 ; /* 0x000000ff0b00720c */ /* 0x000fda0004702670 */ /*0120*/ @P0 BRA 0x420 ; /* 0x000002f000000947 */ /* 0x000fea0003800000 */ /*0130*/ IMAD.MOV.U32 R14, RZ, RZ, c[0x0][0x184] ; /* 0x00006100ff0e7624 */ /* 0x000fe400078e00ff */ /*0140*/ IMAD R4, R11.reuse, c[0x0][0x184], R0 ; /* 0x000061000b047a24 */ /* 0x040fe400078e0200 */ /*0150*/ IMAD R5, R11.reuse, R14, -c[0x0][0x184] ; /* 0x800061000b057624 */ /* 0x040fe400078e020e */ /*0160*/ IMAD.WIDE R10, R11, R9, c[0x0][0x188] ; /* 0x000062000b0a7625 */ /* 0x000fc800078e0209 */ /*0170*/ IMAD.IADD R6, R0, 0x1, R5 ; /* 0x0000000100067824 */ /* 0x000fe200078e0205 */ /*0180*/ LDG.E.64 R12, [R10.64] ; /* 0x000000040a0c7981 */ /* 0x000ea2000c1e1b00 */ /*0190*/ IMAD.WIDE R4, R4, R9, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fc600078e0209 */ /*01a0*/ LDG.E.64 R18, [R10.64+0x8] ; /* 0x000008040a127981 */ /* 0x000ee2000c1e1b00 */ /*01b0*/ IMAD R14, R14, 0x2, R6 ; /* 0x000000020e0e7824 */ /* 0x000fe400078e0206 */ /*01c0*/ IMAD.WIDE R6, R6, R9.reuse, c[0x0][0x160] ; /* 0x0000580006067625 */ /* 0x080fe200078e0209 */ /*01d0*/ LDG.E.64 R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000f26000c1e1b00 */ /*01e0*/ IMAD.WIDE R14, R14, R9, c[0x0][0x160] ; /* 0x000058000e0e7625 */ /* 0x000fe400078e0209 */ /*01f0*/ LDG.E.64 R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000f28000c1e1b00 */ /*0200*/ LDG.E.64 R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000f62000c1e1b00 */ /*0210*/ DADD R8, -R6, R4 ; /* 0x0000000006087229 */ /* 0x010e880000000104 */ /*0220*/ DADD R16, -R4, R14 ; /* 0x0000000004107229 */ /* 0x020ec8000000010e */ /*0230*/ DMUL R12, R8, R12 ; /* 0x0000000c080c7228 */ /* 0x004fc80000000000 */ /*0240*/ DMUL R16, R16, R18 ; /* 0x0000001210107228 */ /* 0x008e0c0000000000 */ /*0250*/ DMUL R8, R12, R16 ; /* 0x000000100c087228 */ /* 0x001e0c0000000000 */ /*0260*/ DSETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800722a */ /* 0x001e1c0003f04000 */ /*0270*/ @!P0 STG.E.64 [R2.64], RZ ; /* 0x000000ff02008986 */ /* 0x0011e2000c101b04 */ /*0280*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0290*/ DADD R8, R12, R16.reuse ; /* 0x000000000c087229 */ /* 0x100e620000000010 */ /*02a0*/ IMAD.MOV.U32 R4, RZ, RZ, 0x1 ; /* 0x00000001ff047424 */ /* 0x000fe200078e00ff */ /*02b0*/ BSSY B0, 0x400 ; /* 0x0000014000007945 */ /* 0x000fe40003800000 */ /*02c0*/ DADD R16, R16, R16 ; /* 0x0000000010107229 */ /* 0x000ea40000000010 */ /*02d0*/ MUFU.RCP64H R5, R9 ; /* 0x0000000900057308 */ /* 0x002e680000001800 */ /*02e0*/ DMUL R16, R12, R16 ; /* 0x000000100c107228 */ /* 0x004e940000000000 */ /*02f0*/ FSETP.GEU.AND P1, PT, |R17|, 6.5827683646048100446e-37, PT ; /* 0x036000001100780b */ /* 0x004fe20003f2e200 */ /*0300*/ DFMA R6, -R8, R4, 1 ; /* 0x3ff000000806742b */ /* 0x002e4c0000000104 */ /*0310*/ DFMA R6, R6, R6, R6 ; /* 0x000000060606722b */ /* 0x002e4c0000000006 */ /*0320*/ DFMA R6, R4, R6, R4 ; /* 0x000000060406722b */ /* 0x002e4c0000000004 */ /*0330*/ DFMA R4, -R8, R6, 1 ; /* 0x3ff000000804742b */ /* 0x002e4c0000000106 */ /*0340*/ DFMA R4, R6, R4, R6 ; /* 0x000000040604722b */ /* 0x002e4c0000000006 */ /*0350*/ DMUL R6, R16, R4 ; /* 0x0000000410067228 */ /* 0x002e4c0000000000 */ /*0360*/ DFMA R10, -R8, R6, R16 ; /* 0x00000006080a722b */ /* 0x002e4c0000000110 */ /*0370*/ DFMA R4, R4, R10, R6 ; /* 0x0000000a0404722b */ /* 0x002e540000000006 */ /*0380*/ FFMA R0, RZ, R9, R5 ; /* 0x00000009ff007223 */ /* 0x002fca0000000005 */ /*0390*/ FSETP.GT.AND P0, PT, |R0|, 1.469367938527859385e-39, PT ; /* 0x001000000000780b */ /* 0x000fda0003f04200 */ /*03a0*/ @P0 BRA P1, 0x3f0 ; /* 0x0000004000000947 */ /* 0x000fea0000800000 */ /*03b0*/ MOV R0, 0x3d0 ; /* 0x000003d000007802 */ /* 0x000fe40000000f00 */ /*03c0*/ CALL.REL.NOINC 0x440 ; /* 0x0000007000007944 */ /* 0x001fea0003c00000 */ /*03d0*/ IMAD.MOV.U32 R4, RZ, RZ, R12 ; /* 0x000000ffff047224 */ /* 0x000fe400078e000c */ /*03e0*/ IMAD.MOV.U32 R5, RZ, RZ, R13 ; /* 0x000000ffff057224 */ /* 0x000fe400078e000d */ /*03f0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0400*/ STG.E.64 [R2.64], R4 ; /* 0x0000000402007986 */ /* 0x000fe2000c101b04 */ /*0410*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0420*/ STG.E.64 [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe2000c101b04 */ /*0430*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0440*/ FSETP.GEU.AND P0, PT, |R9|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000900780b */ /* 0x040fe20003f0e200 */ /*0450*/ IMAD.MOV.U32 R7, RZ, RZ, R17 ; /* 0x000000ffff077224 */ /* 0x000fe200078e0011 */ /*0460*/ LOP3.LUT R4, R9.reuse, 0x800fffff, RZ, 0xc0, !PT ; /* 0x800fffff09047812 */ /* 0x040fe200078ec0ff */ /*0470*/ IMAD.MOV.U32 R13, RZ, RZ, 0x1ca00000 ; /* 0x1ca00000ff0d7424 */ /* 0x000fe200078e00ff */ /*0480*/ LOP3.LUT R15, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000090f7812 */ /* 0x000fe200078ec0ff */ /*0490*/ IMAD.MOV.U32 R6, RZ, RZ, R16 ; /* 0x000000ffff067224 */ /* 0x000fe200078e0010 */ /*04a0*/ LOP3.LUT R5, R4, 0x3ff00000, RZ, 0xfc, !PT ; /* 0x3ff0000004057812 */ /* 0x000fe200078efcff */ /*04b0*/ IMAD.MOV.U32 R4, RZ, RZ, R8 ; /* 0x000000ffff047224 */ /* 0x000fe200078e0008 */ /*04c0*/ FSETP.GEU.AND P2, PT, |R7|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000700780b */ /* 0x040fe20003f4e200 */ /*04d0*/ IMAD.MOV.U32 R16, RZ, RZ, 0x1 ; /* 0x00000001ff107424 */ /* 0x000fe200078e00ff */ /*04e0*/ LOP3.LUT R12, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000070c7812 */ /* 0x000fe200078ec0ff */ /*04f0*/ IMAD.MOV.U32 R10, RZ, RZ, R6 ; /* 0x000000ffff0a7224 */ /* 0x000fe200078e0006 */ /*0500*/ BSSY B1, 0xa00 ; /* 0x000004f000017945 */ /* 0x000fe20003800000 */ /*0510*/ @!P0 DMUL R4, R8, 8.98846567431157953865e+307 ; /* 0x7fe0000008048828 */ /* 0x000e220000000000 */ /*0520*/ ISETP.GE.U32.AND P1, PT, R12, R15, PT ; /* 0x0000000f0c00720c */ /* 0x000fca0003f26070 */ /*0530*/ MUFU.RCP64H R17, R5 ; /* 0x0000000500117308 */ /* 0x001e240000001800 */ /*0540*/ @!P2 LOP3.LUT R11, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000090ba812 */ /* 0x000fe200078ec0ff */ /*0550*/ @!P2 IMAD.MOV.U32 R20, RZ, RZ, RZ ; /* 0x000000ffff14a224 */ /* 0x000fc600078e00ff */ /*0560*/ @!P2 ISETP.GE.U32.AND P3, PT, R12, R11, PT ; /* 0x0000000b0c00a20c */ /* 0x000fe40003f66070 */ /*0570*/ SEL R11, R13.reuse, 0x63400000, !P1 ; /* 0x634000000d0b7807 */ /* 0x040fe40004800000 */ /*0580*/ @!P2 SEL R21, R13, 0x63400000, !P3 ; /* 0x634000000d15a807 */ /* 0x000fe40005800000 */ /*0590*/ LOP3.LUT R11, R11, 0x800fffff, R7.reuse, 0xf8, !PT ; /* 0x800fffff0b0b7812 */ /* 0x100fe400078ef807 */ /*05a0*/ @!P2 LOP3.LUT R21, R21, 0x80000000, R7, 0xf8, !PT ; /* 0x800000001515a812 */ /* 0x000fe200078ef807 */ /*05b0*/ DFMA R18, R16, -R4, 1 ; /* 0x3ff000001012742b */ /* 0x001e060000000804 */ /*05c0*/ @!P2 LOP3.LUT R21, R21, 0x100000, RZ, 0xfc, !PT ; /* 0x001000001515a812 */ /* 0x000fc600078efcff */ /*05d0*/ DFMA R18, R18, R18, R18 ; /* 0x000000121212722b */ /* 0x001e080000000012 */ /*05e0*/ @!P2 DFMA R10, R10, 2, -R20 ; /* 0x400000000a0aa82b */ /* 0x000fc80000000814 */ /*05f0*/ DFMA R18, R16, R18, R16 ; /* 0x000000121012722b */ /* 0x0010640000000010 */ /*0600*/ IMAD.MOV.U32 R16, RZ, RZ, R12 ; /* 0x000000ffff107224 */ /* 0x001fe400078e000c */ /*0610*/ IMAD.MOV.U32 R17, RZ, RZ, R15 ; /* 0x000000ffff117224 */ /* 0x000fe200078e000f */ /*0620*/ @!P0 LOP3.LUT R17, R5, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000005118812 */ /* 0x000fe200078ec0ff */ /*0630*/ DFMA R20, R18, -R4, 1 ; /* 0x3ff000001214742b */ /* 0x002e220000000804 */ /*0640*/ @!P2 LOP3.LUT R16, R11, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff000000b10a812 */ /* 0x000fe400078ec0ff */ /*0650*/ IADD3 R22, R17, -0x1, RZ ; /* 0xffffffff11167810 */ /* 0x000fc60007ffe0ff */ /*0660*/ DFMA R18, R18, R20, R18 ; /* 0x000000141212722b */ /* 0x0010640000000012 */ /*0670*/ IADD3 R20, R16, -0x1, RZ ; /* 0xffffffff10147810 */ /* 0x001fc80007ffe0ff */ /*0680*/ ISETP.GT.U32.AND P0, PT, R20, 0x7feffffe, PT ; /* 0x7feffffe1400780c */ /* 0x000fe20003f04070 */ /*0690*/ DMUL R14, R18, R10 ; /* 0x0000000a120e7228 */ /* 0x002e060000000000 */ /*06a0*/ ISETP.GT.U32.OR P0, PT, R22, 0x7feffffe, P0 ; /* 0x7feffffe1600780c */ /* 0x000fc60000704470 */ /*06b0*/ DFMA R20, R14, -R4, R10 ; /* 0x800000040e14722b */ /* 0x001e0c000000000a */ /*06c0*/ DFMA R14, R18, R20, R14 ; /* 0x00000014120e722b */ /* 0x001048000000000e */ /*06d0*/ @P0 BRA 0x8a0 ; /* 0x000001c000000947 */ /* 0x000fea0003800000 */ /*06e0*/ LOP3.LUT R7, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000009077812 */ /* 0x003fc800078ec0ff */ /*06f0*/ ISETP.GE.U32.AND P0, PT, R12.reuse, R7, PT ; /* 0x000000070c00720c */ /* 0x040fe20003f06070 */ /*0700*/ IMAD.IADD R6, R12, 0x1, -R7 ; /* 0x000000010c067824 */ /* 0x000fc600078e0a07 */ /*0710*/ SEL R13, R13, 0x63400000, !P0 ; /* 0x634000000d0d7807 */ /* 0x000fe40004000000 */ /*0720*/ IMNMX R6, R6, -0x46a00000, !PT ; /* 0xb960000006067817 */ /* 0x000fc80007800200 */ /*0730*/ IMNMX R6, R6, 0x46a00000, PT ; /* 0x46a0000006067817 */ /* 0x000fca0003800200 */ /*0740*/ IMAD.IADD R16, R6, 0x1, -R13 ; /* 0x0000000106107824 */ /* 0x000fe400078e0a0d */ /*0750*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fc600078e00ff */ /*0760*/ IADD3 R7, R16, 0x7fe00000, RZ ; /* 0x7fe0000010077810 */ /* 0x000fcc0007ffe0ff */ /*0770*/ DMUL R12, R14, R6 ; /* 0x000000060e0c7228 */ /* 0x000e140000000000 */ /*0780*/ FSETP.GTU.AND P0, PT, |R13|, 1.469367938527859385e-39, PT ; /* 0x001000000d00780b */ /* 0x001fda0003f0c200 */ /*0790*/ @P0 BRA 0x9f0 ; /* 0x0000025000000947 */ /* 0x000fea0003800000 */ /*07a0*/ DFMA R4, R14, -R4, R10 ; /* 0x800000040e04722b */ /* 0x000e22000000000a */ /*07b0*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fd200078e00ff */ /*07c0*/ FSETP.NEU.AND P0, PT, R5.reuse, RZ, PT ; /* 0x000000ff0500720b */ /* 0x041fe40003f0d000 */ /*07d0*/ LOP3.LUT R9, R5, 0x80000000, R9, 0x48, !PT ; /* 0x8000000005097812 */ /* 0x000fc800078e4809 */ /*07e0*/ LOP3.LUT R7, R9, R7, RZ, 0xfc, !PT ; /* 0x0000000709077212 */ /* 0x000fce00078efcff */ /*07f0*/ @!P0 BRA 0x9f0 ; /* 0x000001f000008947 */ /* 0x000fea0003800000 */ /*0800*/ IMAD.MOV R5, RZ, RZ, -R16 ; /* 0x000000ffff057224 */ /* 0x000fe200078e0a10 */ /*0810*/ DMUL.RP R6, R14, R6 ; /* 0x000000060e067228 */ /* 0x000e220000008000 */ /*0820*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fcc00078e00ff */ /*0830*/ DFMA R4, R12, -R4, R14 ; /* 0x800000040c04722b */ /* 0x000e46000000000e */ /*0840*/ LOP3.LUT R9, R7, R9, RZ, 0x3c, !PT ; /* 0x0000000907097212 */ /* 0x001fc600078e3cff */ /*0850*/ IADD3 R4, -R16, -0x43300000, RZ ; /* 0xbcd0000010047810 */ /* 0x002fc80007ffe1ff */ /*0860*/ FSETP.NEU.AND P0, PT, |R5|, R4, PT ; /* 0x000000040500720b */ /* 0x000fc80003f0d200 */ /*0870*/ FSEL R12, R6, R12, !P0 ; /* 0x0000000c060c7208 */ /* 0x000fe40004000000 */ /*0880*/ FSEL R13, R9, R13, !P0 ; /* 0x0000000d090d7208 */ /* 0x000fe20004000000 */ /*0890*/ BRA 0x9f0 ; /* 0x0000015000007947 */ /* 0x000fea0003800000 */ /*08a0*/ DSETP.NAN.AND P0, PT, R6, R6, PT ; /* 0x000000060600722a */ /* 0x003e1c0003f08000 */ /*08b0*/ @P0 BRA 0x9d0 ; /* 0x0000011000000947 */ /* 0x001fea0003800000 */ /*08c0*/ DSETP.NAN.AND P0, PT, R8, R8, PT ; /* 0x000000080800722a */ /* 0x000e1c0003f08000 */ /*08d0*/ @P0 BRA 0x9a0 ; /* 0x000000c000000947 */ /* 0x001fea0003800000 */ /*08e0*/ ISETP.NE.AND P0, PT, R16, R17, PT ; /* 0x000000111000720c */ /* 0x000fe20003f05270 */ /*08f0*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0900*/ IMAD.MOV.U32 R13, RZ, RZ, -0x80000 ; /* 0xfff80000ff0d7424 */ /* 0x000fd400078e00ff */ /*0910*/ @!P0 BRA 0x9f0 ; /* 0x000000d000008947 */ /* 0x000fea0003800000 */ /*0920*/ ISETP.NE.AND P0, PT, R16, 0x7ff00000, PT ; /* 0x7ff000001000780c */ /* 0x000fe40003f05270 */ /*0930*/ LOP3.LUT R13, R7, 0x80000000, R9, 0x48, !PT ; /* 0x80000000070d7812 */ /* 0x000fe400078e4809 */ /*0940*/ ISETP.EQ.OR P0, PT, R17, RZ, !P0 ; /* 0x000000ff1100720c */ /* 0x000fda0004702670 */ /*0950*/ @P0 LOP3.LUT R4, R13, 0x7ff00000, RZ, 0xfc, !PT ; /* 0x7ff000000d040812 */ /* 0x000fe200078efcff */ /*0960*/ @!P0 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0c8224 */ /* 0x000fe400078e00ff */ /*0970*/ @P0 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0c0224 */ /* 0x000fe400078e00ff */ /*0980*/ @P0 IMAD.MOV.U32 R13, RZ, RZ, R4 ; /* 0x000000ffff0d0224 */ /* 0x000fe200078e0004 */ /*0990*/ BRA 0x9f0 ; /* 0x0000005000007947 */ /* 0x000fea0003800000 */ /*09a0*/ LOP3.LUT R13, R9, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000090d7812 */ /* 0x000fe200078efcff */ /*09b0*/ IMAD.MOV.U32 R12, RZ, RZ, R8 ; /* 0x000000ffff0c7224 */ /* 0x000fe200078e0008 */ /*09c0*/ BRA 0x9f0 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*09d0*/ LOP3.LUT R13, R7, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000070d7812 */ /* 0x000fe200078efcff */ /*09e0*/ IMAD.MOV.U32 R12, RZ, RZ, R6 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0006 */ /*09f0*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0a00*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0000 */ /*0a10*/ IMAD.MOV.U32 R5, RZ, RZ, 0x0 ; /* 0x00000000ff057424 */ /* 0x000fc800078e00ff */ /*0a20*/ RET.REL.NODEC R4 0x0 ; /* 0xfffff5d004007950 */ /* 0x000fea0003c3ffff */ /*0a30*/ BRA 0xa30; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
.file "tmpxft_00035ec0_00000000-6_StarRadKernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ .type _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_, @function _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_: .LFB2051: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movsd %xmm0, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movq %r9, 16(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq 232(%rsp), %rax movq %rax, (%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 28(%rsp), %rax movq %rax, 160(%rsp) leaq 24(%rsp), %rax movq %rax, 168(%rsp) leaq 16(%rsp), %rax movq %rax, 176(%rsp) leaq 8(%rsp), %rax movq %rax, 184(%rsp) movq %rsp, %rax movq %rax, 192(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $1, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) leaq 72(%rsp), %rcx leaq 64(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 200(%rsp), %rax subq %fs:40, %rax jne .L8 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 72(%rsp) .cfi_def_cfa_offset 232 pushq 72(%rsp) .cfi_def_cfa_offset 240 leaq 144(%rsp), %r9 movq 108(%rsp), %rcx movl 116(%rsp), %r8d movq 96(%rsp), %rsi movl 104(%rsp), %edx leaq _Z13StarRadKernelPdS_S_diiS_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_, .-_Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .type _Z13StarRadKernelPdS_S_diiS_S_S_, @function _Z13StarRadKernelPdS_S_diiS_S_S_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z13StarRadKernelPdS_S_diiS_S_S_, .-_Z13StarRadKernelPdS_S_diiS_S_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z13StarRadKernelPdS_S_diiS_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z13StarRadKernelPdS_S_diiS_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z13StarRadKernelPdS_S_diiS_S_S_ .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .p2align 8 .type _Z13StarRadKernelPdS_S_diiS_S_S_,@function _Z13StarRadKernelPdS_S_diiS_S_S_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x4c s_load_b64 s[4:5], s[0:1], 0x20 v_bfe_u32 v2, v0, 10, 10 v_and_b32_e32 v4, 0x3ff, v0 s_waitcnt lgkmcnt(0) s_and_b32 s3, s2, 0xffff s_lshr_b32 s2, s2, 16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s15, s2, v[2:3] v_mad_u64_u32 v[2:3], null, s14, s3, v[4:5] v_cmp_gt_i32_e32 vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_gt_i32_e64 s2, s5, v2 s_and_b32 s2, s2, vcc_lo s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_6 s_add_i32 s2, s4, -1 v_cmp_ne_u32_e32 vcc_lo, 0, v0 v_cmp_ne_u32_e64 s2, s2, v0 v_mov_b32_e32 v3, 0 v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s3, vcc_lo, s2 s_and_saveexec_b32 s2, s3 s_cbranch_execz .LBB0_5 v_add_nc_u32_e32 v3, 1, v0 v_add_nc_u32_e32 v1, -1, v0 s_clause 0x1 s_load_b64 s[6:7], s[0:1], 0x0 s_load_b64 s[8:9], s[0:1], 0x28 v_mad_u64_u32 v[4:5], null, v0, s5, v[2:3] v_mad_u64_u32 v[6:7], null, v1, s5, v[2:3] v_mad_u64_u32 v[8:9], null, v3, s5, v[2:3] v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_ashrrev_i32_e32 v5, 31, v4 v_ashrrev_i32_e32 v7, 31, v6 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_lshlrev_b64 v[12:13], 3, v[0:1] v_ashrrev_i32_e32 v9, 31, v8 v_lshlrev_b64 v[4:5], 3, v[4:5] s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_lshlrev_b64 v[6:7], 3, v[6:7] v_lshlrev_b64 v[8:9], 3, v[8:9] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_u32 v4, vcc_lo, s6, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v5, vcc_lo s_delay_alu instid0(VALU_DEP_4) v_add_co_u32 v6, vcc_lo, s6, v6 v_add_co_ci_u32_e32 v7, vcc_lo, s7, v7, vcc_lo v_add_co_u32 v8, vcc_lo, s6, v8 v_add_co_ci_u32_e32 v9, vcc_lo, s7, v9, vcc_lo s_clause 0x2 global_load_b64 v[10:11], v[4:5], off global_load_b64 v[5:6], v[6:7], off global_load_b64 v[7:8], v[8:9], off v_ashrrev_i32_e32 v4, 31, v3 v_add_co_u32 v12, vcc_lo, s8, v12 v_add_co_ci_u32_e32 v13, vcc_lo, s9, v13, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 3, v[3:4] v_add_co_u32 v3, vcc_lo, s8, v3 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s9, v4, vcc_lo s_clause 0x1 global_load_b64 v[12:13], v[12:13], off global_load_b64 v[3:4], v[3:4], off s_waitcnt vmcnt(3) v_add_f64 v[5:6], v[10:11], -v[5:6] s_waitcnt vmcnt(2) v_add_f64 v[7:8], v[7:8], -v[10:11] s_waitcnt vmcnt(1) s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_f64 v[5:6], v[5:6], v[12:13] s_waitcnt vmcnt(0) v_mul_f64 v[7:8], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[5:6], v[7:8] v_cmp_lt_f64_e32 vcc_lo, 0, v[3:4] v_mov_b32_e32 v3, 0 v_mov_b32_e32 v4, 0 s_and_saveexec_b32 s3, vcc_lo s_cbranch_execz .LBB0_4 v_add_f64 v[3:4], v[7:8], v[7:8] v_add_f64 v[7:8], v[5:6], v[7:8] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[5:6], v[3:4] v_div_scale_f64 v[5:6], null, v[7:8], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f64_e32 v[9:10], v[5:6] s_waitcnt_depctr 0xfff v_fma_f64 v[11:12], -v[5:6], v[9:10], 1.0 v_fma_f64 v[9:10], v[9:10], v[11:12], v[9:10] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[11:12], -v[5:6], v[9:10], 1.0 v_fma_f64 v[9:10], v[9:10], v[11:12], v[9:10] v_div_scale_f64 v[11:12], vcc_lo, v[3:4], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[13:14], v[11:12], v[9:10] v_fma_f64 v[5:6], -v[5:6], v[13:14], v[11:12] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fmas_f64 v[5:6], v[5:6], v[9:10], v[13:14] v_div_fixup_f64 v[3:4], v[5:6], v[7:8], v[3:4] .LBB0_4: s_or_b32 exec_lo, exec_lo, s3 .LBB0_5: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) s_or_b32 exec_lo, exec_lo, s2 s_load_b64 s[0:1], s[0:1], 0x38 v_mad_u64_u32 v[5:6], null, v2, s4, v[0:1] v_ashrrev_i32_e32 v6, 31, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[5:6] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b64 v[0:1], v[3:4], off .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13StarRadKernelPdS_S_diiS_S_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 320 .amdhsa_user_sgpr_count 14 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z13StarRadKernelPdS_S_diiS_S_S_, .Lfunc_end0-_Z13StarRadKernelPdS_S_diiS_S_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 8 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 48 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 56 .size: 8 .value_kind: global_buffer - .offset: 64 .size: 4 .value_kind: hidden_block_count_x - .offset: 68 .size: 4 .value_kind: hidden_block_count_y - .offset: 72 .size: 4 .value_kind: hidden_block_count_z - .offset: 76 .size: 2 .value_kind: hidden_group_size_x - .offset: 78 .size: 2 .value_kind: hidden_group_size_y - .offset: 80 .size: 2 .value_kind: hidden_group_size_z - .offset: 82 .size: 2 .value_kind: hidden_remainder_x - .offset: 84 .size: 2 .value_kind: hidden_remainder_y - .offset: 86 .size: 2 .value_kind: hidden_remainder_z - .offset: 104 .size: 8 .value_kind: hidden_global_offset_x - .offset: 112 .size: 8 .value_kind: hidden_global_offset_y - .offset: 120 .size: 8 .value_kind: hidden_global_offset_z - .offset: 128 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 320 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z13StarRadKernelPdS_S_diiS_S_S_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13StarRadKernelPdS_S_diiS_S_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void StarRadKernel (double *Qbase2, double *Vrad, double *QStar, double dt, int nrad, int nsec, double *invdiffRmed, double *Rmed, double *dq) { int j = threadIdx.x + blockDim.x*blockIdx.x; int i = threadIdx.y + blockDim.y*blockIdx.y; double dqm, dqp; if (i<nrad && j<nsec){ if ((i == 0 || i == nrad-1)) dq[i + j*nrad] = 0.0; else { dqm = (Qbase2[i*nsec + j] - Qbase2[(i-1)*nsec + j])*invdiffRmed[i]; dqp = (Qbase2[(i+1)*nsec + j] - Qbase2[i*nsec + j])*invdiffRmed[i+1]; if (dqp * dqm > 0.0) dq[i+j*nrad] = 2.0*dqp*dqm/(dqp+dqm); else dq[i+j*nrad] = 0.0; } } }
.text .file "StarRadKernel.hip" .globl _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ # -- Begin function _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .p2align 4, 0x90 .type _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_,@function _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_: # @_Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movsd %xmm0, 64(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) movq %r9, 56(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rax movq %rax, 144(%rsp) leaq 176(%rsp), %rax movq %rax, 152(%rsp) leaq 184(%rsp), %rax movq %rax, 160(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z13StarRadKernelPdS_S_diiS_S_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_, .Lfunc_end0-_Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z13StarRadKernelPdS_S_diiS_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z13StarRadKernelPdS_S_diiS_S_S_,@object # @_Z13StarRadKernelPdS_S_diiS_S_S_ .section .rodata,"a",@progbits .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .p2align 3, 0x0 _Z13StarRadKernelPdS_S_diiS_S_S_: .quad _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .size _Z13StarRadKernelPdS_S_diiS_S_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z13StarRadKernelPdS_S_diiS_S_S_" .size .L__unnamed_1, 33 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z13StarRadKernelPdS_S_diiS_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z13StarRadKernelPdS_S_diiS_S_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0030*/ S2R R11, SR_CTAID.Y ; /* 0x00000000000b7919 */ /* 0x000e680000002600 */ /*0040*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x184], PT ; /* 0x0000610000007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R11, R11, c[0x0][0x4], R2 ; /* 0x000001000b0b7a24 */ /* 0x002fca00078e0202 */ /*0080*/ ISETP.GE.OR P0, PT, R11, c[0x0][0x180], P0 ; /* 0x000060000b007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*00b0*/ IMAD.MOV.U32 R9, RZ, RZ, 0x8 ; /* 0x00000008ff097424 */ /* 0x000fe200078e00ff */ /*00c0*/ UIADD3 UR4, UR4, -0x1, URZ ; /* 0xffffffff04047890 */ /* 0x000fe2000fffe03f */ /*00d0*/ IMAD R2, R0, c[0x0][0x180], R11 ; /* 0x0000600000027a24 */ /* 0x000fc800078e020b */ /*00e0*/ IMAD.WIDE R2, R2, R9, c[0x0][0x198] ; /* 0x0000660002027625 */ /* 0x000fe200078e0209 */ /*00f0*/ ISETP.NE.AND P0, PT, R11, UR4, PT ; /* 0x000000040b007c0c */ /* 0x000fe2000bf05270 */ /*0100*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0110*/ ISETP.EQ.OR P0, PT, R11, RZ, !P0 ; /* 0x000000ff0b00720c */ /* 0x000fda0004702670 */ /*0120*/ @P0 BRA 0x420 ; /* 0x000002f000000947 */ /* 0x000fea0003800000 */ /*0130*/ IMAD.MOV.U32 R14, RZ, RZ, c[0x0][0x184] ; /* 0x00006100ff0e7624 */ /* 0x000fe400078e00ff */ /*0140*/ IMAD R4, R11.reuse, c[0x0][0x184], R0 ; /* 0x000061000b047a24 */ /* 0x040fe400078e0200 */ /*0150*/ IMAD R5, R11.reuse, R14, -c[0x0][0x184] ; /* 0x800061000b057624 */ /* 0x040fe400078e020e */ /*0160*/ IMAD.WIDE R10, R11, R9, c[0x0][0x188] ; /* 0x000062000b0a7625 */ /* 0x000fc800078e0209 */ /*0170*/ IMAD.IADD R6, R0, 0x1, R5 ; /* 0x0000000100067824 */ /* 0x000fe200078e0205 */ /*0180*/ LDG.E.64 R12, [R10.64] ; /* 0x000000040a0c7981 */ /* 0x000ea2000c1e1b00 */ /*0190*/ IMAD.WIDE R4, R4, R9, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fc600078e0209 */ /*01a0*/ LDG.E.64 R18, [R10.64+0x8] ; /* 0x000008040a127981 */ /* 0x000ee2000c1e1b00 */ /*01b0*/ IMAD R14, R14, 0x2, R6 ; /* 0x000000020e0e7824 */ /* 0x000fe400078e0206 */ /*01c0*/ IMAD.WIDE R6, R6, R9.reuse, c[0x0][0x160] ; /* 0x0000580006067625 */ /* 0x080fe200078e0209 */ /*01d0*/ LDG.E.64 R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000f26000c1e1b00 */ /*01e0*/ IMAD.WIDE R14, R14, R9, c[0x0][0x160] ; /* 0x000058000e0e7625 */ /* 0x000fe400078e0209 */ /*01f0*/ LDG.E.64 R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000f28000c1e1b00 */ /*0200*/ LDG.E.64 R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000f62000c1e1b00 */ /*0210*/ DADD R8, -R6, R4 ; /* 0x0000000006087229 */ /* 0x010e880000000104 */ /*0220*/ DADD R16, -R4, R14 ; /* 0x0000000004107229 */ /* 0x020ec8000000010e */ /*0230*/ DMUL R12, R8, R12 ; /* 0x0000000c080c7228 */ /* 0x004fc80000000000 */ /*0240*/ DMUL R16, R16, R18 ; /* 0x0000001210107228 */ /* 0x008e0c0000000000 */ /*0250*/ DMUL R8, R12, R16 ; /* 0x000000100c087228 */ /* 0x001e0c0000000000 */ /*0260*/ DSETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800722a */ /* 0x001e1c0003f04000 */ /*0270*/ @!P0 STG.E.64 [R2.64], RZ ; /* 0x000000ff02008986 */ /* 0x0011e2000c101b04 */ /*0280*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0290*/ DADD R8, R12, R16.reuse ; /* 0x000000000c087229 */ /* 0x100e620000000010 */ /*02a0*/ IMAD.MOV.U32 R4, RZ, RZ, 0x1 ; /* 0x00000001ff047424 */ /* 0x000fe200078e00ff */ /*02b0*/ BSSY B0, 0x400 ; /* 0x0000014000007945 */ /* 0x000fe40003800000 */ /*02c0*/ DADD R16, R16, R16 ; /* 0x0000000010107229 */ /* 0x000ea40000000010 */ /*02d0*/ MUFU.RCP64H R5, R9 ; /* 0x0000000900057308 */ /* 0x002e680000001800 */ /*02e0*/ DMUL R16, R12, R16 ; /* 0x000000100c107228 */ /* 0x004e940000000000 */ /*02f0*/ FSETP.GEU.AND P1, PT, |R17|, 6.5827683646048100446e-37, PT ; /* 0x036000001100780b */ /* 0x004fe20003f2e200 */ /*0300*/ DFMA R6, -R8, R4, 1 ; /* 0x3ff000000806742b */ /* 0x002e4c0000000104 */ /*0310*/ DFMA R6, R6, R6, R6 ; /* 0x000000060606722b */ /* 0x002e4c0000000006 */ /*0320*/ DFMA R6, R4, R6, R4 ; /* 0x000000060406722b */ /* 0x002e4c0000000004 */ /*0330*/ DFMA R4, -R8, R6, 1 ; /* 0x3ff000000804742b */ /* 0x002e4c0000000106 */ /*0340*/ DFMA R4, R6, R4, R6 ; /* 0x000000040604722b */ /* 0x002e4c0000000006 */ /*0350*/ DMUL R6, R16, R4 ; /* 0x0000000410067228 */ /* 0x002e4c0000000000 */ /*0360*/ DFMA R10, -R8, R6, R16 ; /* 0x00000006080a722b */ /* 0x002e4c0000000110 */ /*0370*/ DFMA R4, R4, R10, R6 ; /* 0x0000000a0404722b */ /* 0x002e540000000006 */ /*0380*/ FFMA R0, RZ, R9, R5 ; /* 0x00000009ff007223 */ /* 0x002fca0000000005 */ /*0390*/ FSETP.GT.AND P0, PT, |R0|, 1.469367938527859385e-39, PT ; /* 0x001000000000780b */ /* 0x000fda0003f04200 */ /*03a0*/ @P0 BRA P1, 0x3f0 ; /* 0x0000004000000947 */ /* 0x000fea0000800000 */ /*03b0*/ MOV R0, 0x3d0 ; /* 0x000003d000007802 */ /* 0x000fe40000000f00 */ /*03c0*/ CALL.REL.NOINC 0x440 ; /* 0x0000007000007944 */ /* 0x001fea0003c00000 */ /*03d0*/ IMAD.MOV.U32 R4, RZ, RZ, R12 ; /* 0x000000ffff047224 */ /* 0x000fe400078e000c */ /*03e0*/ IMAD.MOV.U32 R5, RZ, RZ, R13 ; /* 0x000000ffff057224 */ /* 0x000fe400078e000d */ /*03f0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0400*/ STG.E.64 [R2.64], R4 ; /* 0x0000000402007986 */ /* 0x000fe2000c101b04 */ /*0410*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0420*/ STG.E.64 [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe2000c101b04 */ /*0430*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0440*/ FSETP.GEU.AND P0, PT, |R9|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000900780b */ /* 0x040fe20003f0e200 */ /*0450*/ IMAD.MOV.U32 R7, RZ, RZ, R17 ; /* 0x000000ffff077224 */ /* 0x000fe200078e0011 */ /*0460*/ LOP3.LUT R4, R9.reuse, 0x800fffff, RZ, 0xc0, !PT ; /* 0x800fffff09047812 */ /* 0x040fe200078ec0ff */ /*0470*/ IMAD.MOV.U32 R13, RZ, RZ, 0x1ca00000 ; /* 0x1ca00000ff0d7424 */ /* 0x000fe200078e00ff */ /*0480*/ LOP3.LUT R15, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000090f7812 */ /* 0x000fe200078ec0ff */ /*0490*/ IMAD.MOV.U32 R6, RZ, RZ, R16 ; /* 0x000000ffff067224 */ /* 0x000fe200078e0010 */ /*04a0*/ LOP3.LUT R5, R4, 0x3ff00000, RZ, 0xfc, !PT ; /* 0x3ff0000004057812 */ /* 0x000fe200078efcff */ /*04b0*/ IMAD.MOV.U32 R4, RZ, RZ, R8 ; /* 0x000000ffff047224 */ /* 0x000fe200078e0008 */ /*04c0*/ FSETP.GEU.AND P2, PT, |R7|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000700780b */ /* 0x040fe20003f4e200 */ /*04d0*/ IMAD.MOV.U32 R16, RZ, RZ, 0x1 ; /* 0x00000001ff107424 */ /* 0x000fe200078e00ff */ /*04e0*/ LOP3.LUT R12, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000070c7812 */ /* 0x000fe200078ec0ff */ /*04f0*/ IMAD.MOV.U32 R10, RZ, RZ, R6 ; /* 0x000000ffff0a7224 */ /* 0x000fe200078e0006 */ /*0500*/ BSSY B1, 0xa00 ; /* 0x000004f000017945 */ /* 0x000fe20003800000 */ /*0510*/ @!P0 DMUL R4, R8, 8.98846567431157953865e+307 ; /* 0x7fe0000008048828 */ /* 0x000e220000000000 */ /*0520*/ ISETP.GE.U32.AND P1, PT, R12, R15, PT ; /* 0x0000000f0c00720c */ /* 0x000fca0003f26070 */ /*0530*/ MUFU.RCP64H R17, R5 ; /* 0x0000000500117308 */ /* 0x001e240000001800 */ /*0540*/ @!P2 LOP3.LUT R11, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff00000090ba812 */ /* 0x000fe200078ec0ff */ /*0550*/ @!P2 IMAD.MOV.U32 R20, RZ, RZ, RZ ; /* 0x000000ffff14a224 */ /* 0x000fc600078e00ff */ /*0560*/ @!P2 ISETP.GE.U32.AND P3, PT, R12, R11, PT ; /* 0x0000000b0c00a20c */ /* 0x000fe40003f66070 */ /*0570*/ SEL R11, R13.reuse, 0x63400000, !P1 ; /* 0x634000000d0b7807 */ /* 0x040fe40004800000 */ /*0580*/ @!P2 SEL R21, R13, 0x63400000, !P3 ; /* 0x634000000d15a807 */ /* 0x000fe40005800000 */ /*0590*/ LOP3.LUT R11, R11, 0x800fffff, R7.reuse, 0xf8, !PT ; /* 0x800fffff0b0b7812 */ /* 0x100fe400078ef807 */ /*05a0*/ @!P2 LOP3.LUT R21, R21, 0x80000000, R7, 0xf8, !PT ; /* 0x800000001515a812 */ /* 0x000fe200078ef807 */ /*05b0*/ DFMA R18, R16, -R4, 1 ; /* 0x3ff000001012742b */ /* 0x001e060000000804 */ /*05c0*/ @!P2 LOP3.LUT R21, R21, 0x100000, RZ, 0xfc, !PT ; /* 0x001000001515a812 */ /* 0x000fc600078efcff */ /*05d0*/ DFMA R18, R18, R18, R18 ; /* 0x000000121212722b */ /* 0x001e080000000012 */ /*05e0*/ @!P2 DFMA R10, R10, 2, -R20 ; /* 0x400000000a0aa82b */ /* 0x000fc80000000814 */ /*05f0*/ DFMA R18, R16, R18, R16 ; /* 0x000000121012722b */ /* 0x0010640000000010 */ /*0600*/ IMAD.MOV.U32 R16, RZ, RZ, R12 ; /* 0x000000ffff107224 */ /* 0x001fe400078e000c */ /*0610*/ IMAD.MOV.U32 R17, RZ, RZ, R15 ; /* 0x000000ffff117224 */ /* 0x000fe200078e000f */ /*0620*/ @!P0 LOP3.LUT R17, R5, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000005118812 */ /* 0x000fe200078ec0ff */ /*0630*/ DFMA R20, R18, -R4, 1 ; /* 0x3ff000001214742b */ /* 0x002e220000000804 */ /*0640*/ @!P2 LOP3.LUT R16, R11, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff000000b10a812 */ /* 0x000fe400078ec0ff */ /*0650*/ IADD3 R22, R17, -0x1, RZ ; /* 0xffffffff11167810 */ /* 0x000fc60007ffe0ff */ /*0660*/ DFMA R18, R18, R20, R18 ; /* 0x000000141212722b */ /* 0x0010640000000012 */ /*0670*/ IADD3 R20, R16, -0x1, RZ ; /* 0xffffffff10147810 */ /* 0x001fc80007ffe0ff */ /*0680*/ ISETP.GT.U32.AND P0, PT, R20, 0x7feffffe, PT ; /* 0x7feffffe1400780c */ /* 0x000fe20003f04070 */ /*0690*/ DMUL R14, R18, R10 ; /* 0x0000000a120e7228 */ /* 0x002e060000000000 */ /*06a0*/ ISETP.GT.U32.OR P0, PT, R22, 0x7feffffe, P0 ; /* 0x7feffffe1600780c */ /* 0x000fc60000704470 */ /*06b0*/ DFMA R20, R14, -R4, R10 ; /* 0x800000040e14722b */ /* 0x001e0c000000000a */ /*06c0*/ DFMA R14, R18, R20, R14 ; /* 0x00000014120e722b */ /* 0x001048000000000e */ /*06d0*/ @P0 BRA 0x8a0 ; /* 0x000001c000000947 */ /* 0x000fea0003800000 */ /*06e0*/ LOP3.LUT R7, R9, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000009077812 */ /* 0x003fc800078ec0ff */ /*06f0*/ ISETP.GE.U32.AND P0, PT, R12.reuse, R7, PT ; /* 0x000000070c00720c */ /* 0x040fe20003f06070 */ /*0700*/ IMAD.IADD R6, R12, 0x1, -R7 ; /* 0x000000010c067824 */ /* 0x000fc600078e0a07 */ /*0710*/ SEL R13, R13, 0x63400000, !P0 ; /* 0x634000000d0d7807 */ /* 0x000fe40004000000 */ /*0720*/ IMNMX R6, R6, -0x46a00000, !PT ; /* 0xb960000006067817 */ /* 0x000fc80007800200 */ /*0730*/ IMNMX R6, R6, 0x46a00000, PT ; /* 0x46a0000006067817 */ /* 0x000fca0003800200 */ /*0740*/ IMAD.IADD R16, R6, 0x1, -R13 ; /* 0x0000000106107824 */ /* 0x000fe400078e0a0d */ /*0750*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fc600078e00ff */ /*0760*/ IADD3 R7, R16, 0x7fe00000, RZ ; /* 0x7fe0000010077810 */ /* 0x000fcc0007ffe0ff */ /*0770*/ DMUL R12, R14, R6 ; /* 0x000000060e0c7228 */ /* 0x000e140000000000 */ /*0780*/ FSETP.GTU.AND P0, PT, |R13|, 1.469367938527859385e-39, PT ; /* 0x001000000d00780b */ /* 0x001fda0003f0c200 */ /*0790*/ @P0 BRA 0x9f0 ; /* 0x0000025000000947 */ /* 0x000fea0003800000 */ /*07a0*/ DFMA R4, R14, -R4, R10 ; /* 0x800000040e04722b */ /* 0x000e22000000000a */ /*07b0*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fd200078e00ff */ /*07c0*/ FSETP.NEU.AND P0, PT, R5.reuse, RZ, PT ; /* 0x000000ff0500720b */ /* 0x041fe40003f0d000 */ /*07d0*/ LOP3.LUT R9, R5, 0x80000000, R9, 0x48, !PT ; /* 0x8000000005097812 */ /* 0x000fc800078e4809 */ /*07e0*/ LOP3.LUT R7, R9, R7, RZ, 0xfc, !PT ; /* 0x0000000709077212 */ /* 0x000fce00078efcff */ /*07f0*/ @!P0 BRA 0x9f0 ; /* 0x000001f000008947 */ /* 0x000fea0003800000 */ /*0800*/ IMAD.MOV R5, RZ, RZ, -R16 ; /* 0x000000ffff057224 */ /* 0x000fe200078e0a10 */ /*0810*/ DMUL.RP R6, R14, R6 ; /* 0x000000060e067228 */ /* 0x000e220000008000 */ /*0820*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fcc00078e00ff */ /*0830*/ DFMA R4, R12, -R4, R14 ; /* 0x800000040c04722b */ /* 0x000e46000000000e */ /*0840*/ LOP3.LUT R9, R7, R9, RZ, 0x3c, !PT ; /* 0x0000000907097212 */ /* 0x001fc600078e3cff */ /*0850*/ IADD3 R4, -R16, -0x43300000, RZ ; /* 0xbcd0000010047810 */ /* 0x002fc80007ffe1ff */ /*0860*/ FSETP.NEU.AND P0, PT, |R5|, R4, PT ; /* 0x000000040500720b */ /* 0x000fc80003f0d200 */ /*0870*/ FSEL R12, R6, R12, !P0 ; /* 0x0000000c060c7208 */ /* 0x000fe40004000000 */ /*0880*/ FSEL R13, R9, R13, !P0 ; /* 0x0000000d090d7208 */ /* 0x000fe20004000000 */ /*0890*/ BRA 0x9f0 ; /* 0x0000015000007947 */ /* 0x000fea0003800000 */ /*08a0*/ DSETP.NAN.AND P0, PT, R6, R6, PT ; /* 0x000000060600722a */ /* 0x003e1c0003f08000 */ /*08b0*/ @P0 BRA 0x9d0 ; /* 0x0000011000000947 */ /* 0x001fea0003800000 */ /*08c0*/ DSETP.NAN.AND P0, PT, R8, R8, PT ; /* 0x000000080800722a */ /* 0x000e1c0003f08000 */ /*08d0*/ @P0 BRA 0x9a0 ; /* 0x000000c000000947 */ /* 0x001fea0003800000 */ /*08e0*/ ISETP.NE.AND P0, PT, R16, R17, PT ; /* 0x000000111000720c */ /* 0x000fe20003f05270 */ /*08f0*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0900*/ IMAD.MOV.U32 R13, RZ, RZ, -0x80000 ; /* 0xfff80000ff0d7424 */ /* 0x000fd400078e00ff */ /*0910*/ @!P0 BRA 0x9f0 ; /* 0x000000d000008947 */ /* 0x000fea0003800000 */ /*0920*/ ISETP.NE.AND P0, PT, R16, 0x7ff00000, PT ; /* 0x7ff000001000780c */ /* 0x000fe40003f05270 */ /*0930*/ LOP3.LUT R13, R7, 0x80000000, R9, 0x48, !PT ; /* 0x80000000070d7812 */ /* 0x000fe400078e4809 */ /*0940*/ ISETP.EQ.OR P0, PT, R17, RZ, !P0 ; /* 0x000000ff1100720c */ /* 0x000fda0004702670 */ /*0950*/ @P0 LOP3.LUT R4, R13, 0x7ff00000, RZ, 0xfc, !PT ; /* 0x7ff000000d040812 */ /* 0x000fe200078efcff */ /*0960*/ @!P0 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0c8224 */ /* 0x000fe400078e00ff */ /*0970*/ @P0 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0c0224 */ /* 0x000fe400078e00ff */ /*0980*/ @P0 IMAD.MOV.U32 R13, RZ, RZ, R4 ; /* 0x000000ffff0d0224 */ /* 0x000fe200078e0004 */ /*0990*/ BRA 0x9f0 ; /* 0x0000005000007947 */ /* 0x000fea0003800000 */ /*09a0*/ LOP3.LUT R13, R9, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000090d7812 */ /* 0x000fe200078efcff */ /*09b0*/ IMAD.MOV.U32 R12, RZ, RZ, R8 ; /* 0x000000ffff0c7224 */ /* 0x000fe200078e0008 */ /*09c0*/ BRA 0x9f0 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*09d0*/ LOP3.LUT R13, R7, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000070d7812 */ /* 0x000fe200078efcff */ /*09e0*/ IMAD.MOV.U32 R12, RZ, RZ, R6 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0006 */ /*09f0*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0a00*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0000 */ /*0a10*/ IMAD.MOV.U32 R5, RZ, RZ, 0x0 ; /* 0x00000000ff057424 */ /* 0x000fc800078e00ff */ /*0a20*/ RET.REL.NODEC R4 0x0 ; /* 0xfffff5d004007950 */ /* 0x000fea0003c3ffff */ /*0a30*/ BRA 0xa30; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z13StarRadKernelPdS_S_diiS_S_S_ .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .p2align 8 .type _Z13StarRadKernelPdS_S_diiS_S_S_,@function _Z13StarRadKernelPdS_S_diiS_S_S_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x4c s_load_b64 s[4:5], s[0:1], 0x20 v_bfe_u32 v2, v0, 10, 10 v_and_b32_e32 v4, 0x3ff, v0 s_waitcnt lgkmcnt(0) s_and_b32 s3, s2, 0xffff s_lshr_b32 s2, s2, 16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s15, s2, v[2:3] v_mad_u64_u32 v[2:3], null, s14, s3, v[4:5] v_cmp_gt_i32_e32 vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_gt_i32_e64 s2, s5, v2 s_and_b32 s2, s2, vcc_lo s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_6 s_add_i32 s2, s4, -1 v_cmp_ne_u32_e32 vcc_lo, 0, v0 v_cmp_ne_u32_e64 s2, s2, v0 v_mov_b32_e32 v3, 0 v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s3, vcc_lo, s2 s_and_saveexec_b32 s2, s3 s_cbranch_execz .LBB0_5 v_add_nc_u32_e32 v3, 1, v0 v_add_nc_u32_e32 v1, -1, v0 s_clause 0x1 s_load_b64 s[6:7], s[0:1], 0x0 s_load_b64 s[8:9], s[0:1], 0x28 v_mad_u64_u32 v[4:5], null, v0, s5, v[2:3] v_mad_u64_u32 v[6:7], null, v1, s5, v[2:3] v_mad_u64_u32 v[8:9], null, v3, s5, v[2:3] v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_ashrrev_i32_e32 v5, 31, v4 v_ashrrev_i32_e32 v7, 31, v6 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_lshlrev_b64 v[12:13], 3, v[0:1] v_ashrrev_i32_e32 v9, 31, v8 v_lshlrev_b64 v[4:5], 3, v[4:5] s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_lshlrev_b64 v[6:7], 3, v[6:7] v_lshlrev_b64 v[8:9], 3, v[8:9] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_u32 v4, vcc_lo, s6, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v5, vcc_lo s_delay_alu instid0(VALU_DEP_4) v_add_co_u32 v6, vcc_lo, s6, v6 v_add_co_ci_u32_e32 v7, vcc_lo, s7, v7, vcc_lo v_add_co_u32 v8, vcc_lo, s6, v8 v_add_co_ci_u32_e32 v9, vcc_lo, s7, v9, vcc_lo s_clause 0x2 global_load_b64 v[10:11], v[4:5], off global_load_b64 v[5:6], v[6:7], off global_load_b64 v[7:8], v[8:9], off v_ashrrev_i32_e32 v4, 31, v3 v_add_co_u32 v12, vcc_lo, s8, v12 v_add_co_ci_u32_e32 v13, vcc_lo, s9, v13, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 3, v[3:4] v_add_co_u32 v3, vcc_lo, s8, v3 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s9, v4, vcc_lo s_clause 0x1 global_load_b64 v[12:13], v[12:13], off global_load_b64 v[3:4], v[3:4], off s_waitcnt vmcnt(3) v_add_f64 v[5:6], v[10:11], -v[5:6] s_waitcnt vmcnt(2) v_add_f64 v[7:8], v[7:8], -v[10:11] s_waitcnt vmcnt(1) s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_f64 v[5:6], v[5:6], v[12:13] s_waitcnt vmcnt(0) v_mul_f64 v[7:8], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[5:6], v[7:8] v_cmp_lt_f64_e32 vcc_lo, 0, v[3:4] v_mov_b32_e32 v3, 0 v_mov_b32_e32 v4, 0 s_and_saveexec_b32 s3, vcc_lo s_cbranch_execz .LBB0_4 v_add_f64 v[3:4], v[7:8], v[7:8] v_add_f64 v[7:8], v[5:6], v[7:8] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[5:6], v[3:4] v_div_scale_f64 v[5:6], null, v[7:8], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f64_e32 v[9:10], v[5:6] s_waitcnt_depctr 0xfff v_fma_f64 v[11:12], -v[5:6], v[9:10], 1.0 v_fma_f64 v[9:10], v[9:10], v[11:12], v[9:10] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[11:12], -v[5:6], v[9:10], 1.0 v_fma_f64 v[9:10], v[9:10], v[11:12], v[9:10] v_div_scale_f64 v[11:12], vcc_lo, v[3:4], v[7:8], v[3:4] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[13:14], v[11:12], v[9:10] v_fma_f64 v[5:6], -v[5:6], v[13:14], v[11:12] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fmas_f64 v[5:6], v[5:6], v[9:10], v[13:14] v_div_fixup_f64 v[3:4], v[5:6], v[7:8], v[3:4] .LBB0_4: s_or_b32 exec_lo, exec_lo, s3 .LBB0_5: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) s_or_b32 exec_lo, exec_lo, s2 s_load_b64 s[0:1], s[0:1], 0x38 v_mad_u64_u32 v[5:6], null, v2, s4, v[0:1] v_ashrrev_i32_e32 v6, 31, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[5:6] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b64 v[0:1], v[3:4], off .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13StarRadKernelPdS_S_diiS_S_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 320 .amdhsa_user_sgpr_count 14 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z13StarRadKernelPdS_S_diiS_S_S_, .Lfunc_end0-_Z13StarRadKernelPdS_S_diiS_S_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 8 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 48 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 56 .size: 8 .value_kind: global_buffer - .offset: 64 .size: 4 .value_kind: hidden_block_count_x - .offset: 68 .size: 4 .value_kind: hidden_block_count_y - .offset: 72 .size: 4 .value_kind: hidden_block_count_z - .offset: 76 .size: 2 .value_kind: hidden_group_size_x - .offset: 78 .size: 2 .value_kind: hidden_group_size_y - .offset: 80 .size: 2 .value_kind: hidden_group_size_z - .offset: 82 .size: 2 .value_kind: hidden_remainder_x - .offset: 84 .size: 2 .value_kind: hidden_remainder_y - .offset: 86 .size: 2 .value_kind: hidden_remainder_z - .offset: 104 .size: 8 .value_kind: hidden_global_offset_x - .offset: 112 .size: 8 .value_kind: hidden_global_offset_y - .offset: 120 .size: 8 .value_kind: hidden_global_offset_z - .offset: 128 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 320 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z13StarRadKernelPdS_S_diiS_S_S_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13StarRadKernelPdS_S_diiS_S_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00035ec0_00000000-6_StarRadKernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ .type _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_, @function _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_: .LFB2051: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movsd %xmm0, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movq %r9, 16(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq 232(%rsp), %rax movq %rax, (%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 28(%rsp), %rax movq %rax, 160(%rsp) leaq 24(%rsp), %rax movq %rax, 168(%rsp) leaq 16(%rsp), %rax movq %rax, 176(%rsp) leaq 8(%rsp), %rax movq %rax, 184(%rsp) movq %rsp, %rax movq %rax, 192(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $1, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) leaq 72(%rsp), %rcx leaq 64(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 200(%rsp), %rax subq %fs:40, %rax jne .L8 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 72(%rsp) .cfi_def_cfa_offset 232 pushq 72(%rsp) .cfi_def_cfa_offset 240 leaq 144(%rsp), %r9 movq 108(%rsp), %rcx movl 116(%rsp), %r8d movq 96(%rsp), %rsi movl 104(%rsp), %edx leaq _Z13StarRadKernelPdS_S_diiS_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_, .-_Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .type _Z13StarRadKernelPdS_S_diiS_S_S_, @function _Z13StarRadKernelPdS_S_diiS_S_S_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z46__device_stub__Z13StarRadKernelPdS_S_diiS_S_S_PdS_S_diiS_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z13StarRadKernelPdS_S_diiS_S_S_, .-_Z13StarRadKernelPdS_S_diiS_S_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z13StarRadKernelPdS_S_diiS_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z13StarRadKernelPdS_S_diiS_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "StarRadKernel.hip" .globl _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ # -- Begin function _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .p2align 4, 0x90 .type _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_,@function _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_: # @_Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movsd %xmm0, 64(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) movq %r9, 56(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rax movq %rax, 144(%rsp) leaq 176(%rsp), %rax movq %rax, 152(%rsp) leaq 184(%rsp), %rax movq %rax, 160(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z13StarRadKernelPdS_S_diiS_S_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_, .Lfunc_end0-_Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z13StarRadKernelPdS_S_diiS_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z13StarRadKernelPdS_S_diiS_S_S_,@object # @_Z13StarRadKernelPdS_S_diiS_S_S_ .section .rodata,"a",@progbits .globl _Z13StarRadKernelPdS_S_diiS_S_S_ .p2align 3, 0x0 _Z13StarRadKernelPdS_S_diiS_S_S_: .quad _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .size _Z13StarRadKernelPdS_S_diiS_S_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z13StarRadKernelPdS_S_diiS_S_S_" .size .L__unnamed_1, 33 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z28__device_stub__StarRadKernelPdS_S_diiS_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z13StarRadKernelPdS_S_diiS_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <cuda.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device cudaMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); cudaMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, cudaMemcpyHostToDevice); //allocate the histogram data on the device cudaMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); cudaMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,cudaMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu cudaMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, cudaMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); cudaFree(d_gpu_histogram); cudaFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
code for sm_80 Function : _Z11PDH_kernel2P10hist_entryP8atomdescxd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z10PDH_kernelP10hist_entryP8atomdescxd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_CTAID.X ; /* 0x0000000000087919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R8, R8, c[0x0][0x0], R3 ; /* 0x0000000008087a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R8, c[0x0][0x170], PT ; /* 0x00005c0008007a0c */ /* 0x000fe40003f06070 */ /*0050*/ SHF.R.S32.HI R2, RZ, 0x1f, R8 ; /* 0x0000001fff027819 */ /* 0x000fc80000011408 */ /*0060*/ ISETP.GE.AND.EX P0, PT, R2, c[0x0][0x174], PT, P0 ; /* 0x00005d0002007a0c */ /* 0x000fda0003f06300 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ IADD3 R0, R8, 0x1, RZ ; /* 0x0000000108007810 */ /* 0x000fc80007ffe0ff */ /*0090*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe40003f06070 */ /*00a0*/ SHF.R.S32.HI R3, RZ, 0x1f, R0 ; /* 0x0000001fff037819 */ /* 0x000fc80000011400 */ /*00b0*/ ISETP.GE.AND.EX P0, PT, R3, c[0x0][0x174], PT, P0 ; /* 0x00005d0003007a0c */ /* 0x000fda0003f06300 */ /*00c0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00d0*/ IMAD.MOV.U32 R9, RZ, RZ, 0x18 ; /* 0x00000018ff097424 */ /* 0x000fe200078e00ff */ /*00e0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00f0*/ IMAD R5, R2, 0x18, RZ ; /* 0x0000001802057824 */ /* 0x000fe400078e02ff */ /*0100*/ IMAD.WIDE.U32 R8, R8, R9, c[0x0][0x168] ; /* 0x00005a0008087625 */ /* 0x000fc800078e0009 */ /*0110*/ IMAD.MOV.U32 R11, RZ, RZ, R3 ; /* 0x000000ffff0b7224 */ /* 0x000fe400078e0003 */ /*0120*/ IMAD.MOV.U32 R10, RZ, RZ, R0 ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e0000 */ /*0130*/ IMAD.IADD R9, R9, 0x1, R5 ; /* 0x0000000109097824 */ /* 0x000fc800078e0205 */ /*0140*/ IMAD.MOV.U32 R3, RZ, RZ, 0x18 ; /* 0x00000018ff037424 */ /* 0x001fe200078e00ff */ /*0150*/ LDG.E.64 R12, [R8.64+0x8] ; /* 0x00000804080c7981 */ /* 0x000ea2000c1e1b00 */ /*0160*/ IMAD R11, R11, 0x18, RZ ; /* 0x000000180b0b7824 */ /* 0x000fe400078e02ff */ /*0170*/ IMAD.WIDE.U32 R2, R10, R3, c[0x0][0x168] ; /* 0x00005a000a027625 */ /* 0x000fe200078e0003 */ /*0180*/ LDG.E.64 R6, [R8.64] ; /* 0x0000000408067981 */ /* 0x000ee6000c1e1b00 */ /*0190*/ IMAD.IADD R3, R3, 0x1, R11 ; /* 0x0000000103037824 */ /* 0x000fe200078e020b */ /*01a0*/ LDG.E.64 R16, [R8.64+0x10] ; /* 0x0000100408107981 */ /* 0x000f28000c1e1b00 */ /*01b0*/ LDG.E.64 R10, [R2.64+0x8] ; /* 0x00000804020a7981 */ /* 0x000ea8000c1e1b00 */ /*01c0*/ LDG.E.64 R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ee8000c1e1b00 */ /*01d0*/ LDG.E.64 R14, [R2.64+0x10] ; /* 0x00001004020e7981 */ /* 0x000f22000c1e1b00 */ /*01e0*/ YIELD ; /* 0x0000000000007946 */ /* 0x000fe20003800000 */ /*01f0*/ BSSY B0, 0x3b0 ; /* 0x000001b000007945 */ /* 0x000fe20003800000 */ /*0200*/ DADD R10, -R10, R12 ; /* 0x000000000a0a7229 */ /* 0x004e08000000010c */ /*0210*/ DADD R4, -R4, R6 ; /* 0x0000000004047229 */ /* 0x008fc80000000106 */ /*0220*/ DMUL R10, R10, R10 ; /* 0x0000000a0a0a7228 */ /* 0x001e080000000000 */ /*0230*/ DADD R14, -R14, R16 ; /* 0x000000000e0e7229 */ /* 0x010fc80000000110 */ /*0240*/ DFMA R6, R4, R4, R10 ; /* 0x000000040406722b */ /* 0x001e0c000000000a */ /*0250*/ DFMA R6, R14, R14, R6 ; /* 0x0000000e0e06722b */ /* 0x001e0c0000000006 */ /*0260*/ MUFU.RSQ64H R5, R7 ; /* 0x0000000700057308 */ /* 0x001e280000001c00 */ /*0270*/ IADD3 R4, R7, -0x3500000, RZ ; /* 0xfcb0000007047810 */ /* 0x000fe20007ffe0ff */ /*0280*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0290*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3fd80000 ; /* 0x3fd80000ff0d7424 */ /* 0x000fc600078e00ff */ /*02a0*/ DMUL R10, R4, R4 ; /* 0x00000004040a7228 */ /* 0x001e0c0000000000 */ /*02b0*/ DFMA R10, R6, -R10, 1 ; /* 0x3ff00000060a742b */ /* 0x001e0c000000080a */ /*02c0*/ DFMA R2, R10, R12, 0.5 ; /* 0x3fe000000a02742b */ /* 0x001fc8000000000c */ /*02d0*/ DMUL R10, R4, R10 ; /* 0x0000000a040a7228 */ /* 0x000e220000000000 */ /*02e0*/ ISETP.GE.U32.AND P0, PT, R4, 0x7ca00000, PT ; /* 0x7ca000000400780c */ /* 0x000fca0003f06070 */ /*02f0*/ DFMA R12, R2, R10, R4 ; /* 0x0000000a020c722b */ /* 0x001e0c0000000004 */ /*0300*/ DMUL R14, R6, R12 ; /* 0x0000000c060e7228 */ /* 0x001e080000000000 */ /*0310*/ IADD3 R11, R13, -0x100000, RZ ; /* 0xfff000000d0b7810 */ /* 0x000fe20007ffe0ff */ /*0320*/ IMAD.MOV.U32 R10, RZ, RZ, R12 ; /* 0x000000ffff0a7224 */ /* 0x000fe200078e000c */ /*0330*/ DFMA R16, R14, -R14, R6 ; /* 0x8000000e0e10722b */ /* 0x001e0c0000000006 */ /*0340*/ DFMA R2, R16, R10, R14 ; /* 0x0000000a1002722b */ /* 0x001062000000000e */ /*0350*/ @!P0 BRA 0x3a0 ; /* 0x0000004000008947 */ /* 0x000fea0003800000 */ /*0360*/ MOV R2, 0x380 ; /* 0x0000038000027802 */ /* 0x002fca0000000f00 */ /*0370*/ CALL.REL.NOINC 0xbd0 ; /* 0x0000085000007944 */ /* 0x001fea0003c00000 */ /*0380*/ IMAD.MOV.U32 R2, RZ, RZ, R4 ; /* 0x000000ffff027224 */ /* 0x002fe400078e0004 */ /*0390*/ IMAD.MOV.U32 R3, RZ, RZ, R5 ; /* 0x000000ffff037224 */ /* 0x000fe400078e0005 */ /*03a0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*03b0*/ MUFU.RCP64H R5, c[0x0][0x17c] ; /* 0x00005f0000057b08 */ /* 0x000ea20000001800 */ /*03c0*/ IMAD.MOV.U32 R10, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff0a7624 */ /* 0x001fe200078e00ff */ /*03d0*/ FSETP.GEU.AND P1, PT, |R3|, 6.5827683646048100446e-37, PT ; /* 0x036000000300780b */ /* 0x002fe20003f2e200 */ /*03e0*/ IMAD.MOV.U32 R11, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff0b7624 */ /* 0x000fe200078e00ff */ /*03f0*/ BSSY B0, 0x510 ; /* 0x0000011000007945 */ /* 0x000fe20003800000 */ /*0400*/ IMAD.MOV.U32 R4, RZ, RZ, 0x1 ; /* 0x00000001ff047424 */ /* 0x000fcc00078e00ff */ /*0410*/ DFMA R6, R4, -R10, 1 ; /* 0x3ff000000406742b */ /* 0x004e0c000000080a */ /*0420*/ DFMA R6, R6, R6, R6 ; /* 0x000000060606722b */ /* 0x001e0c0000000006 */ /*0430*/ DFMA R6, R4, R6, R4 ; /* 0x000000060406722b */ /* 0x001e0c0000000004 */ /*0440*/ DFMA R4, R6, -R10, 1 ; /* 0x3ff000000604742b */ /* 0x001e0c000000080a */ /*0450*/ DFMA R4, R6, R4, R6 ; /* 0x000000040604722b */ /* 0x001e0c0000000006 */ /*0460*/ DMUL R6, R4, R2 ; /* 0x0000000204067228 */ /* 0x001e0c0000000000 */ /*0470*/ DFMA R10, R6, -c[0x0][0x178], R2 ; /* 0x80005e00060a7a2b */ /* 0x001e0c0000000002 */ /*0480*/ DFMA R4, R4, R10, R6 ; /* 0x0000000a0404722b */ /* 0x001e140000000006 */ /*0490*/ FFMA R6, RZ, c[0x0][0x17c], R5 ; /* 0x00005f00ff067a23 */ /* 0x001fca0000000005 */ /*04a0*/ FSETP.GT.AND P0, PT, |R6|, 1.469367938527859385e-39, PT ; /* 0x001000000600780b */ /* 0x000fda0003f04200 */ /*04b0*/ @P0 BRA P1, 0x500 ; /* 0x0000004000000947 */ /* 0x000fea0000800000 */ /*04c0*/ IMAD.MOV.U32 R4, RZ, RZ, R2 ; /* 0x000000ffff047224 */ /* 0x000fe200078e0002 */ /*04d0*/ MOV R16, 0x500 ; /* 0x0000050000107802 */ /* 0x000fe20000000f00 */ /*04e0*/ IMAD.MOV.U32 R5, RZ, RZ, R3 ; /* 0x000000ffff057224 */ /* 0x000fe400078e0003 */ /*04f0*/ CALL.REL.NOINC 0x5e0 ; /* 0x000000e000007944 */ /* 0x000fea0003c00000 */ /*0500*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0510*/ F2I.F64.TRUNC R2, R4 ; /* 0x0000000400027311 */ /* 0x000e22000030d100 */ /*0520*/ IADD3 R10, R0, 0x1, RZ ; /* 0x00000001000a7810 */ /* 0x000fe20007ffe0ff */ /*0530*/ IMAD.MOV.U32 R3, RZ, RZ, 0x8 ; /* 0x00000008ff037424 */ /* 0x000fe400078e00ff */ /*0540*/ IMAD.MOV.U32 R6, RZ, RZ, 0x1 ; /* 0x00000001ff067424 */ /* 0x000fe200078e00ff */ /*0550*/ ISETP.GE.U32.AND P0, PT, R10, c[0x0][0x170], PT ; /* 0x00005c000a007a0c */ /* 0x000fe20003f06070 */ /*0560*/ IMAD.MOV.U32 R7, RZ, RZ, 0x0 ; /* 0x00000000ff077424 */ /* 0x000fe200078e00ff */ /*0570*/ SHF.R.S32.HI R11, RZ, 0x1f, R10 ; /* 0x0000001fff0b7819 */ /* 0x000fc8000001140a */ /*0580*/ ISETP.GE.AND.EX P0, PT, R11, c[0x0][0x174], PT, P0 ; /* 0x00005d000b007a0c */ /* 0x000fe20003f06300 */ /*0590*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x001fca00078e0203 */ /*05a0*/ RED.E.ADD.64.STRONG.GPU [R2.64], R6 ; /* 0x000000060200798e */ /* 0x0001e2000c10e584 */ /*05b0*/ IMAD.MOV.U32 R0, RZ, RZ, R10 ; /* 0x000000ffff007224 */ /* 0x000fcc00078e000a */ /*05c0*/ @!P0 BRA 0x140 ; /* 0xfffffb7000008947 */ /* 0x000fea000383ffff */ /*05d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*05e0*/ IMAD.MOV.U32 R18, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff127624 */ /* 0x000fe200078e00ff */ /*05f0*/ FSETP.GEU.AND P2, PT, |R5|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000500780b */ /* 0x040fe20003f4e200 */ /*0600*/ IMAD.MOV.U32 R17, RZ, RZ, 0x1ca00000 ; /* 0x1ca00000ff117424 */ /* 0x000fe200078e00ff */ /*0610*/ LOP3.LUT R20, R5, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000005147812 */ /* 0x000fe200078ec0ff */ /*0620*/ IMAD.MOV.U32 R10, RZ, RZ, 0x1 ; /* 0x00000001ff0a7424 */ /* 0x000fe200078e00ff */ /*0630*/ FSETP.GEU.AND P0, PT, |R18|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000001200780b */ /* 0x040fe20003f0e200 */ /*0640*/ BSSY B1, 0xb90 ; /* 0x0000054000017945 */ /* 0x000fe20003800000 */ /*0650*/ LOP3.LUT R6, R18.reuse, 0x800fffff, RZ, 0xc0, !PT ; /* 0x800fffff12067812 */ /* 0x040fe400078ec0ff */ /*0660*/ LOP3.LUT R19, R18, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000012137812 */ /* 0x000fe400078ec0ff */ /*0670*/ LOP3.LUT R7, R6, 0x3ff00000, RZ, 0xfc, !PT ; /* 0x3ff0000006077812 */ /* 0x000fe200078efcff */ /*0680*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff067624 */ /* 0x000fe200078e00ff */ /*0690*/ ISETP.GE.U32.AND P1, PT, R20, R19, PT ; /* 0x000000131400720c */ /* 0x000fe20003f26070 */ /*06a0*/ IMAD.MOV.U32 R22, RZ, RZ, R19 ; /* 0x000000ffff167224 */ /* 0x000fe200078e0013 */ /*06b0*/ MOV R21, R20 ; /* 0x0000001400157202 */ /* 0x000fc40000000f00 */ /*06c0*/ SEL R17, R17, 0x63400000, !P1 ; /* 0x6340000011117807 */ /* 0x000fe20004800000 */ /*06d0*/ @!P0 IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff028624 */ /* 0x000fe400078e00ff */ /*06e0*/ @!P0 IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff038624 */ /* 0x000fe200078e00ff */ /*06f0*/ @!P2 LOP3.LUT R12, R17, 0x80000000, R5, 0xf8, !PT ; /* 0x80000000110ca812 */ /* 0x000fc800078ef805 */ /*0700*/ @!P2 LOP3.LUT R13, R12, 0x100000, RZ, 0xfc, !PT ; /* 0x001000000c0da812 */ /* 0x000fe200078efcff */ /*0710*/ @!P0 DMUL R6, R2, 8.98846567431157953865e+307 ; /* 0x7fe0000002068828 */ /* 0x000e220000000000 */ /*0720*/ @!P2 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0ca224 */ /* 0x000fca00078e00ff */ /*0730*/ MUFU.RCP64H R11, R7 ; /* 0x00000007000b7308 */ /* 0x001e280000001800 */ /*0740*/ @!P0 LOP3.LUT R22, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000007168812 */ /* 0x000fc800078ec0ff */ /*0750*/ IADD3 R23, R22, -0x1, RZ ; /* 0xffffffff16177810 */ /* 0x000fe20007ffe0ff */ /*0760*/ DFMA R2, R10, -R6, 1 ; /* 0x3ff000000a02742b */ /* 0x001e0c0000000806 */ /*0770*/ DFMA R14, R2, R2, R2 ; /* 0x00000002020e722b */ /* 0x0010640000000002 */ /*0780*/ LOP3.LUT R3, R17, 0x800fffff, R5, 0xf8, !PT ; /* 0x800fffff11037812 */ /* 0x001fe200078ef805 */ /*0790*/ IMAD.MOV.U32 R2, RZ, RZ, R4 ; /* 0x000000ffff027224 */ /* 0x000fc600078e0004 */ /*07a0*/ DFMA R10, R10, R14, R10 ; /* 0x0000000e0a0a722b */ /* 0x002e08000000000a */ /*07b0*/ @!P2 DFMA R2, R2, 2, -R12 ; /* 0x400000000202a82b */ /* 0x000e48000000080c */ /*07c0*/ DFMA R12, R10, -R6, 1 ; /* 0x3ff000000a0c742b */ /* 0x001e0c0000000806 */ /*07d0*/ @!P2 LOP3.LUT R21, R3, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff000000315a812 */ /* 0x002fe200078ec0ff */ /*07e0*/ DFMA R10, R10, R12, R10 ; /* 0x0000000c0a0a722b */ /* 0x001e06000000000a */ /*07f0*/ IADD3 R14, R21, -0x1, RZ ; /* 0xffffffff150e7810 */ /* 0x000fc60007ffe0ff */ /*0800*/ DMUL R12, R10, R2 ; /* 0x000000020a0c7228 */ /* 0x001e220000000000 */ /*0810*/ ISETP.GT.U32.AND P0, PT, R14, 0x7feffffe, PT ; /* 0x7feffffe0e00780c */ /* 0x000fc80003f04070 */ /*0820*/ ISETP.GT.U32.OR P0, PT, R23, 0x7feffffe, P0 ; /* 0x7feffffe1700780c */ /* 0x000fe20000704470 */ /*0830*/ DFMA R14, R12, -R6, R2 ; /* 0x800000060c0e722b */ /* 0x001e0c0000000002 */ /*0840*/ DFMA R14, R10, R14, R12 ; /* 0x0000000e0a0e722b */ /* 0x00104c000000000c */ /*0850*/ @P0 BRA 0xa00 ; /* 0x000001a000000947 */ /* 0x000fea0003800000 */ /*0860*/ IMAD.IADD R19, R20, 0x1, -R19 ; /* 0x0000000114137824 */ /* 0x003fca00078e0a13 */ /*0870*/ IMNMX R19, R19, -0x46a00000, !PT ; /* 0xb960000013137817 */ /* 0x000fc80007800200 */ /*0880*/ IMNMX R4, R19, 0x46a00000, PT ; /* 0x46a0000013047817 */ /* 0x000fca0003800200 */ /*0890*/ IMAD.IADD R17, R4, 0x1, -R17 ; /* 0x0000000104117824 */ /* 0x000fe400078e0a11 */ /*08a0*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fc600078e00ff */ /*08b0*/ IADD3 R5, R17, 0x7fe00000, RZ ; /* 0x7fe0000011057810 */ /* 0x000fcc0007ffe0ff */ /*08c0*/ DMUL R10, R14, R4 ; /* 0x000000040e0a7228 */ /* 0x000e140000000000 */ /*08d0*/ FSETP.GTU.AND P0, PT, |R11|, 1.469367938527859385e-39, PT ; /* 0x001000000b00780b */ /* 0x001fda0003f0c200 */ /*08e0*/ @P0 BRA 0xb80 ; /* 0x0000029000000947 */ /* 0x000fea0003800000 */ /*08f0*/ DFMA R2, R14, -R6, R2 ; /* 0x800000060e02722b */ /* 0x000e220000000002 */ /*0900*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fd200078e00ff */ /*0910*/ FSETP.NEU.AND P0, PT, R3.reuse, RZ, PT ; /* 0x000000ff0300720b */ /* 0x041fe40003f0d000 */ /*0920*/ LOP3.LUT R2, R3, c[0x0][0x17c], RZ, 0x3c, !PT ; /* 0x00005f0003027a12 */ /* 0x000fc800078e3cff */ /*0930*/ LOP3.LUT R7, R2, 0x80000000, RZ, 0xc0, !PT ; /* 0x8000000002077812 */ /* 0x000fc800078ec0ff */ /*0940*/ LOP3.LUT R5, R7, R5, RZ, 0xfc, !PT ; /* 0x0000000507057212 */ /* 0x000fc600078efcff */ /*0950*/ @!P0 BRA 0xb80 ; /* 0x0000022000008947 */ /* 0x000fea0003800000 */ /*0960*/ IMAD.MOV R3, RZ, RZ, -R17 ; /* 0x000000ffff037224 */ /* 0x000fe200078e0a11 */ /*0970*/ DMUL.RP R4, R14, R4 ; /* 0x000000040e047228 */ /* 0x000e220000008000 */ /*0980*/ IMAD.MOV.U32 R2, RZ, RZ, RZ ; /* 0x000000ffff027224 */ /* 0x000fcc00078e00ff */ /*0990*/ DFMA R2, R10, -R2, R14 ; /* 0x800000020a02722b */ /* 0x000e46000000000e */ /*09a0*/ LOP3.LUT R7, R5, R7, RZ, 0x3c, !PT ; /* 0x0000000705077212 */ /* 0x001fc600078e3cff */ /*09b0*/ IADD3 R2, -R17, -0x43300000, RZ ; /* 0xbcd0000011027810 */ /* 0x002fc80007ffe1ff */ /*09c0*/ FSETP.NEU.AND P0, PT, |R3|, R2, PT ; /* 0x000000020300720b */ /* 0x000fc80003f0d200 */ /*09d0*/ FSEL R10, R4, R10, !P0 ; /* 0x0000000a040a7208 */ /* 0x000fe40004000000 */ /*09e0*/ FSEL R11, R7, R11, !P0 ; /* 0x0000000b070b7208 */ /* 0x000fe20004000000 */ /*09f0*/ BRA 0xb80 ; /* 0x0000018000007947 */ /* 0x000fea0003800000 */ /*0a00*/ DSETP.NAN.AND P0, PT, R4, R4, PT ; /* 0x000000040400722a */ /* 0x003e1c0003f08000 */ /*0a10*/ @P0 BRA 0xb60 ; /* 0x0000014000000947 */ /* 0x001fea0003800000 */ /*0a20*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff027624 */ /* 0x000fe400078e00ff */ /*0a30*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff037624 */ /* 0x000fcc00078e00ff */ /*0a40*/ DSETP.NAN.AND P0, PT, R2, c[0x0][0x178], PT ; /* 0x00005e000200762a */ /* 0x000e1c0003f08000 */ /*0a50*/ @P0 BRA 0xb30 ; /* 0x000000d000000947 */ /* 0x001fea0003800000 */ /*0a60*/ ISETP.NE.AND P0, PT, R21, R22, PT ; /* 0x000000161500720c */ /* 0x000fe20003f05270 */ /*0a70*/ IMAD.MOV.U32 R10, RZ, RZ, 0x0 ; /* 0x00000000ff0a7424 */ /* 0x000fe400078e00ff */ /*0a80*/ IMAD.MOV.U32 R11, RZ, RZ, -0x80000 ; /* 0xfff80000ff0b7424 */ /* 0x000fd400078e00ff */ /*0a90*/ @!P0 BRA 0xb80 ; /* 0x000000e000008947 */ /* 0x000fea0003800000 */ /*0aa0*/ ISETP.NE.AND P0, PT, R21, 0x7ff00000, PT ; /* 0x7ff000001500780c */ /* 0x000fe40003f05270 */ /*0ab0*/ LOP3.LUT R4, R5, c[0x0][0x17c], RZ, 0x3c, !PT ; /* 0x00005f0005047a12 */ /* 0x000fe400078e3cff */ /*0ac0*/ ISETP.EQ.OR P0, PT, R22, RZ, !P0 ; /* 0x000000ff1600720c */ /* 0x000fe40004702670 */ /*0ad0*/ LOP3.LUT R11, R4, 0x80000000, RZ, 0xc0, !PT ; /* 0x80000000040b7812 */ /* 0x000fd600078ec0ff */ /*0ae0*/ @P0 LOP3.LUT R2, R11, 0x7ff00000, RZ, 0xfc, !PT ; /* 0x7ff000000b020812 */ /* 0x000fe200078efcff */ /*0af0*/ @!P0 IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a8224 */ /* 0x000fe400078e00ff */ /*0b00*/ @P0 IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a0224 */ /* 0x000fe400078e00ff */ /*0b10*/ @P0 IMAD.MOV.U32 R11, RZ, RZ, R2 ; /* 0x000000ffff0b0224 */ /* 0x000fe200078e0002 */ /*0b20*/ BRA 0xb80 ; /* 0x0000005000007947 */ /* 0x000fea0003800000 */ /*0b30*/ LOP3.LUT R11, R18, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000120b7812 */ /* 0x000fe200078efcff */ /*0b40*/ IMAD.MOV.U32 R10, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff0a7624 */ /* 0x000fe200078e00ff */ /*0b50*/ BRA 0xb80 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*0b60*/ LOP3.LUT R11, R5, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000050b7812 */ /* 0x000fe400078efcff */ /*0b70*/ MOV R10, R4 ; /* 0x00000004000a7202 */ /* 0x000fc40000000f00 */ /*0b80*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0b90*/ IMAD.MOV.U32 R17, RZ, RZ, 0x0 ; /* 0x00000000ff117424 */ /* 0x000fe400078e00ff */ /*0ba0*/ IMAD.MOV.U32 R4, RZ, RZ, R10 ; /* 0x000000ffff047224 */ /* 0x000fc400078e000a */ /*0bb0*/ IMAD.MOV.U32 R5, RZ, RZ, R11 ; /* 0x000000ffff057224 */ /* 0x000fe200078e000b */ /*0bc0*/ RET.REL.NODEC R16 0x0 ; /* 0xfffff43010007950 */ /* 0x000fec0003c3ffff */ /*0bd0*/ ISETP.GE.U32.AND P0, PT, R4, -0x3400000, PT ; /* 0xfcc000000400780c */ /* 0x000fe20003f06070 */ /*0be0*/ BSSY B1, 0xe50 ; /* 0x0000026000017945 */ /* 0x000fe20003800000 */ /*0bf0*/ IMAD.MOV.U32 R10, RZ, RZ, R12 ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e000c */ /*0c00*/ IMAD.MOV.U32 R4, RZ, RZ, R16 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0010 */ /*0c10*/ IMAD.MOV.U32 R5, RZ, RZ, R17 ; /* 0x000000ffff057224 */ /* 0x000fce00078e0011 */ /*0c20*/ @!P0 BRA 0xcb0 ; /* 0x0000008000008947 */ /* 0x000fea0003800000 */ /*0c30*/ DFMA.RM R4, R4, R10, R14 ; /* 0x0000000a0404722b */ /* 0x000e14000000400e */ /*0c40*/ IADD3 R10, P0, R4, 0x1, RZ ; /* 0x00000001040a7810 */ /* 0x001fca0007f1e0ff */ /*0c50*/ IMAD.X R11, RZ, RZ, R5, P0 ; /* 0x000000ffff0b7224 */ /* 0x000fcc00000e0605 */ /*0c60*/ DFMA.RP R6, -R4, R10, R6 ; /* 0x0000000a0406722b */ /* 0x000e0c0000008106 */ /*0c70*/ DSETP.GT.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600722a */ /* 0x001e0c0003f04000 */ /*0c80*/ FSEL R4, R10, R4, P0 ; /* 0x000000040a047208 */ /* 0x001fe40000000000 */ /*0c90*/ FSEL R5, R11, R5, P0 ; /* 0x000000050b057208 */ /* 0x000fe20000000000 */ /*0ca0*/ BRA 0xe40 ; /* 0x0000019000007947 */ /* 0x000fea0003800000 */ /*0cb0*/ DSETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600722a */ /* 0x000e1c0003f05000 */ /*0cc0*/ @!P0 BRA 0xe30 ; /* 0x0000016000008947 */ /* 0x001fea0003800000 */ /*0cd0*/ ISETP.GE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fda0003f06270 */ /*0ce0*/ @!P0 IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff048424 */ /* 0x000fe400078e00ff */ /*0cf0*/ @!P0 IMAD.MOV.U32 R5, RZ, RZ, -0x80000 ; /* 0xfff80000ff058424 */ /* 0x000fe200078e00ff */ /*0d00*/ @!P0 BRA 0xe40 ; /* 0x0000013000008947 */ /* 0x000fea0003800000 */ /*0d10*/ ISETP.GT.AND P0, PT, R7, 0x7fefffff, PT ; /* 0x7fefffff0700780c */ /* 0x000fda0003f04270 */ /*0d20*/ @P0 BRA 0xe30 ; /* 0x0000010000000947 */ /* 0x000fea0003800000 */ /*0d30*/ DMUL R4, R6, 8.11296384146066816958e+31 ; /* 0x4690000006047828 */ /* 0x0000620000000000 */ /*0d40*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0d50*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x001fe400078e00ff */ /*0d60*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3fd80000 ; /* 0x3fd80000ff0d7424 */ /* 0x000fe200078e00ff */ /*0d70*/ MUFU.RSQ64H R7, R5 ; /* 0x0000000500077308 */ /* 0x002e260000001c00 */ /*0d80*/ DMUL R10, R6, R6 ; /* 0x00000006060a7228 */ /* 0x001e0c0000000000 */ /*0d90*/ DFMA R10, R4, -R10, 1 ; /* 0x3ff00000040a742b */ /* 0x001e0c000000080a */ /*0da0*/ DFMA R12, R10, R12, 0.5 ; /* 0x3fe000000a0c742b */ /* 0x001fc8000000000c */ /*0db0*/ DMUL R10, R6, R10 ; /* 0x0000000a060a7228 */ /* 0x000e0c0000000000 */ /*0dc0*/ DFMA R10, R12, R10, R6 ; /* 0x0000000a0c0a722b */ /* 0x001e0c0000000006 */ /*0dd0*/ DMUL R6, R4, R10 ; /* 0x0000000a04067228 */ /* 0x0010480000000000 */ /*0de0*/ IADD3 R11, R11, -0x100000, RZ ; /* 0xfff000000b0b7810 */ /* 0x001fe40007ffe0ff */ /*0df0*/ DFMA R12, R6, -R6, R4 ; /* 0x80000006060c722b */ /* 0x002e0c0000000004 */ /*0e00*/ DFMA R4, R10, R12, R6 ; /* 0x0000000c0a04722b */ /* 0x001e140000000006 */ /*0e10*/ IADD3 R5, R5, -0x3500000, RZ ; /* 0xfcb0000005057810 */ /* 0x001fe20007ffe0ff */ /*0e20*/ BRA 0xe40 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*0e30*/ DADD R4, R6, R6 ; /* 0x0000000006047229 */ /* 0x00004c0000000006 */ /*0e40*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0e50*/ IMAD.MOV.U32 R3, RZ, RZ, 0x0 ; /* 0x00000000ff037424 */ /* 0x000fc800078e00ff */ /*0e60*/ RET.REL.NODEC R2 0x0 ; /* 0xfffff19002007950 */ /* 0x000fea0003c3ffff */ /*0e70*/ BRA 0xe70; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0e80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0e90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ea0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0eb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ec0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ed0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ee0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ef0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <cuda.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device cudaMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); cudaMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, cudaMemcpyHostToDevice); //allocate the histogram data on the device cudaMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); cudaMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,cudaMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu cudaMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, cudaMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); cudaFree(d_gpu_histogram); cudaFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
.file "tmpxft_0015a26d_00000000-6_proj1-tanmay.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2065: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2065: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z12p2p_distanceii .type _Z12p2p_distanceii, @function _Z12p2p_distanceii: .LFB2057: .cfi_startproc endbr64 movq atom_list(%rip), %rax movslq %edi, %rdi leaq (%rdi,%rdi,2), %rdx leaq (%rax,%rdx,8), %rdx movslq %esi, %rsi leaq (%rsi,%rsi,2), %rcx leaq (%rax,%rcx,8), %rax movsd (%rdx), %xmm1 subsd (%rax), %xmm1 movsd 8(%rdx), %xmm2 subsd 8(%rax), %xmm2 movsd 16(%rdx), %xmm0 subsd 16(%rax), %xmm0 mulsd %xmm1, %xmm1 mulsd %xmm2, %xmm2 addsd %xmm2, %xmm1 mulsd %xmm0, %xmm0 addsd %xmm1, %xmm0 sqrtsd %xmm0, %xmm0 ret .cfi_endproc .LFE2057: .size _Z12p2p_distanceii, .-_Z12p2p_distanceii .globl _Z12PDH_baselinev .type _Z12PDH_baselinev, @function _Z12PDH_baselinev: .LFB2058: .cfi_startproc endbr64 cmpq $0, PDH_acnt(%rip) jle .L11 pushq %r14 .cfi_def_cfa_offset 16 .cfi_offset 14, -16 pushq %r13 .cfi_def_cfa_offset 24 .cfi_offset 13, -24 pushq %r12 .cfi_def_cfa_offset 32 .cfi_offset 12, -32 pushq %rbp .cfi_def_cfa_offset 40 .cfi_offset 6, -40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset 3, -48 movl $1, %r13d movl $0, %r14d .L7: movl %r14d, %r12d movl %r13d, %ebx cmpq %r13, PDH_acnt(%rip) jle .L5 movq %r13, %rbp .L6: movl %ebx, %esi movl %r12d, %edi call _Z12p2p_distanceii divsd PDH_res(%rip), %xmm0 cvttsd2sil %xmm0, %eax cltq movq histogram(%rip), %rdx addq $1, (%rdx,%rax,8) addl $1, %ebx movq PDH_acnt(%rip), %rax addq $1, %rbp cmpq %rbp, %rax jg .L6 addq $1, %r14 addq $1, %r13 cmpq %r14, %rax jg .L7 .L5: movl $0, %eax popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .L11: .cfi_restore 3 .cfi_restore 6 .cfi_restore 12 .cfi_restore 13 .cfi_restore 14 movl $0, %eax ret .cfi_endproc .LFE2058: .size _Z12PDH_baselinev, .-_Z12PDH_baselinev .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Running time for CPU version: %ld.%06ld\n" .text .globl _Z19report_running_timev .type _Z19report_running_timev, @function _Z19report_running_timev: .LFB2059: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $8, %rsp .cfi_def_cfa_offset 32 leaq Idunno(%rip), %rsi leaq endTime(%rip), %rdi call gettimeofday@PLT movq endTime(%rip), %rbp subq startTime(%rip), %rbp movq 8+endTime(%rip), %rbx subq 8+startTime(%rip), %rbx js .L17 .L15: movq %rbx, %rcx movq %rbp, %rdx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq %rbp, %xmm1 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state subq $1, %rbp addq $1000000, %rbx jmp .L15 .cfi_endproc .LFE2059: .size _Z19report_running_timev, .-_Z19report_running_timev .section .rodata.str1.8 .align 8 .LC2: .string "Running time for GPU version: %ld.%06ld\n" .text .globl _Z23report_running_time_GPUv .type _Z23report_running_time_GPUv, @function _Z23report_running_time_GPUv: .LFB2060: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $8, %rsp .cfi_def_cfa_offset 32 leaq Idunno(%rip), %rsi leaq endTime(%rip), %rdi call gettimeofday@PLT movq endTime(%rip), %rbp subq startTime(%rip), %rbp movq 8+endTime(%rip), %rbx subq 8+startTime(%rip), %rbx js .L21 .L19: movq %rbx, %rcx movq %rbp, %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq %rbp, %xmm1 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L21: .cfi_restore_state subq $1, %rbp addq $1000000, %rbx jmp .L19 .cfi_endproc .LFE2060: .size _Z23report_running_time_GPUv, .-_Z23report_running_time_GPUv .section .rodata.str1.1,"aMS",@progbits,1 .LC3: .string "\n%02d: " .LC4: .string "%15lld " .LC5: .string "\n T:%lld \n" .LC6: .string "| " .text .globl _Z16output_histogramP10hist_entry .type _Z16output_histogramP10hist_entry, @function _Z16output_histogramP10hist_entry: .LFB2061: .cfi_startproc endbr64 cmpl $0, num_buckets(%rip) jle .L30 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $8, %rsp .cfi_def_cfa_offset 64 movq %rdi, %rbp movl $0, %ebx movl $0, %r12d leaq .LC4(%rip), %r13 leaq .LC6(%rip), %r14 leaq .LC5(%rip), %r15 jmp .L27 .L34: movl %ebx, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L24 .L35: movq %r15, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L26: addq $1, %rbx cmpl %ebx, num_buckets(%rip) jle .L33 .L27: movslq %ebx, %rax imulq $1717986919, %rax, %rax sarq $33, %rax movl %ebx, %edx sarl $31, %edx subl %edx, %eax leal (%rax,%rax,4), %eax cmpl %ebx, %eax je .L34 .L24: movq 0(%rbp,%rbx,8), %rdx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rdx addq 0(%rbp,%rbx,8), %rdx movq %rdx, %r12 movl num_buckets(%rip), %eax subl $1, %eax cmpl %ebx, %eax je .L35 movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L26 .L33: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L30: .cfi_restore 3 .cfi_restore 6 .cfi_restore 12 .cfi_restore 13 .cfi_restore 14 .cfi_restore 15 ret .cfi_endproc .LFE2061: .size _Z16output_histogramP10hist_entry, .-_Z16output_histogramP10hist_entry .globl _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .type _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, @function _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movsd %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L40 .L36: movq 136(%rsp), %rax subq %fs:40, %rax jne .L41 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L40: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z10PDH_kernelP10hist_entryP8atomdescxd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L36 .L41: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, .-_Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .type _Z10PDH_kernelP10hist_entryP8atomdescxd, @function _Z10PDH_kernelP10hist_entryP8atomdescxd: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z10PDH_kernelP10hist_entryP8atomdescxd, .-_Z10PDH_kernelP10hist_entryP8atomdescxd .section .rodata.str1.1 .LC10: .string "Starting CPU...\n" .LC11: .string "Starting GPU...\n" .LC16: .string "Difference: \n" .text .globl main .type main, @function main: .LFB2062: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $40, %rsp .cfi_def_cfa_offset 64 movq %rsi, %rbx movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT cltq movq %rax, PDH_acnt(%rip) movq 16(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, PDH_res(%rip) movsd .LC7(%rip), %xmm1 divsd %xmm0, %xmm1 cvttsd2sil %xmm1, %edi addl $1, %edi movl %edi, num_buckets(%rip) movslq %edi, %rdi salq $3, %rdi call malloc@PLT movq %rax, histogram(%rip) movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rdi salq $3, %rdi call malloc@PLT movq %rax, atom_list(%rip) movl $1, %edi call srand@PLT cmpq $0, PDH_acnt(%rip) jle .L45 movl $0, %ebx movl $0, %ebp .L46: call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, (%rax,%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 8(%rax,%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 16(%rax,%rbx) addq $1, %rbp addq $24, %rbx cmpq %rbp, PDH_acnt(%rip) jg .L46 .L45: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq Idunno(%rip), %rbp movq %rbp, %rsi leaq startTime(%rip), %rbx movq %rbx, %rdi call gettimeofday@PLT call _Z12PDH_baselinev call _Z19report_running_timev movq histogram(%rip), %rdi call _Z16output_histogramP10hist_entry leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movslq num_buckets(%rip), %rdi salq $3, %rdi call malloc@PLT movq %rax, h_gpu_histogram(%rip) movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rsi salq $3, %rsi leaq d_atom_list(%rip), %rdi call cudaMalloc@PLT movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rdx salq $3, %rdx movl $1, %ecx movq atom_list(%rip), %rsi movq d_atom_list(%rip), %rdi call cudaMemcpy@PLT movslq num_buckets(%rip), %rsi salq $3, %rsi leaq d_gpu_histogram(%rip), %rdi call cudaMalloc@PLT movslq num_buckets(%rip), %rdx salq $3, %rdx movl $1, %ecx movq h_gpu_histogram(%rip), %rsi movq d_gpu_histogram(%rip), %rdi call cudaMemcpy@PLT movq %rbp, %rsi movq %rbx, %rdi call gettimeofday@PLT movl $256, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) pxor %xmm0, %xmm0 cvtsi2sdq PDH_acnt(%rip), %xmm0 mulsd .LC12(%rip), %xmm0 movapd %xmm0, %xmm3 movsd .LC17(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC13(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L47 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC15(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L47: cvttsd2siq %xmm3, %rax movl %eax, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L54 .L48: movslq num_buckets(%rip), %rdx salq $3, %rdx movl $2, %ecx movq d_gpu_histogram(%rip), %rsi movq h_gpu_histogram(%rip), %rdi call cudaMemcpy@PLT call _Z23report_running_time_GPUv movq h_gpu_histogram(%rip), %rdi call _Z16output_histogramP10hist_entry leaq .LC16(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl num_buckets(%rip), %ebp movslq %ebp, %rbx salq $3, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, diff_histogram(%rip) testl %ebp, %ebp jle .L49 movl $0, %eax .L50: movq histogram(%rip), %rdx movq (%rdx,%rax), %rdx movq h_gpu_histogram(%rip), %rcx subq (%rcx,%rax), %rdx movq diff_histogram(%rip), %rcx movq %rdx, (%rcx,%rax) addq $8, %rax cmpq %rax, %rbx jne .L50 .L49: movq diff_histogram(%rip), %rdi call _Z16output_histogramP10hist_entry movq d_gpu_histogram(%rip), %rdi call cudaFree@PLT movq d_atom_list(%rip), %rdi call cudaFree@PLT movq histogram(%rip), %rdi call free@PLT movq atom_list(%rip), %rdi call free@PLT movq h_gpu_histogram(%rip), %rdi call free@PLT movq diff_histogram(%rip), %rdi call free@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L54: .cfi_restore_state movsd PDH_res(%rip), %xmm0 movq PDH_acnt(%rip), %rdx movq d_atom_list(%rip), %rsi movq d_gpu_histogram(%rip), %rdi call _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd jmp .L48 .cfi_endproc .LFE2062: .size main, .-main .globl _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .type _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, @function _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd: .LFB2089: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movsd %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L59 .L55: movq 136(%rsp), %rax subq %fs:40, %rax jne .L60 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L59: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11PDH_kernel2P10hist_entryP8atomdescxd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L55 .L60: call __stack_chk_fail@PLT .cfi_endproc .LFE2089: .size _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, .-_Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .type _Z11PDH_kernel2P10hist_entryP8atomdescxd, @function _Z11PDH_kernel2P10hist_entryP8atomdescxd: .LFB2090: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2090: .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, .-_Z11PDH_kernel2P10hist_entryP8atomdescxd .section .rodata.str1.8 .align 8 .LC18: .string "_Z11PDH_kernel2P10hist_entryP8atomdescxd" .align 8 .LC19: .string "_Z10PDH_kernelP10hist_entryP8atomdescxd" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2092: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC18(%rip), %rdx movq %rdx, %rcx leaq _Z11PDH_kernel2P10hist_entryP8atomdescxd(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC19(%rip), %rdx movq %rdx, %rcx leaq _Z10PDH_kernelP10hist_entryP8atomdescxd(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2092: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .globl endTime .bss .align 16 .type endTime, @object .size endTime, 16 endTime: .zero 16 .globl startTime .align 16 .type startTime, @object .size startTime, 16 startTime: .zero 16 .globl Idunno .align 8 .type Idunno, @object .size Idunno, 8 Idunno: .zero 8 .globl d_atom_list .align 8 .type d_atom_list, @object .size d_atom_list, 8 d_atom_list: .zero 8 .globl atom_list .align 8 .type atom_list, @object .size atom_list, 8 atom_list: .zero 8 .globl PDH_res .align 8 .type PDH_res, @object .size PDH_res, 8 PDH_res: .zero 8 .globl num_buckets .align 4 .type num_buckets, @object .size num_buckets, 4 num_buckets: .zero 4 .globl PDH_acnt .align 8 .type PDH_acnt, @object .size PDH_acnt, 8 PDH_acnt: .zero 8 .globl diff_histogram .align 8 .type diff_histogram, @object .size diff_histogram, 8 diff_histogram: .zero 8 .globl d_gpu_histogram .align 8 .type d_gpu_histogram, @object .size d_gpu_histogram, 8 d_gpu_histogram: .zero 8 .globl h_gpu_histogram .align 8 .type h_gpu_histogram, @object .size h_gpu_histogram, 8 h_gpu_histogram: .zero 8 .globl histogram .align 8 .type histogram, @object .size histogram, 8 histogram: .zero 8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1093567616 .align 8 .LC7: .long 0 .long 1088648064 .align 8 .LC8: .long -4194304 .long 1105199103 .align 8 .LC9: .long 0 .long 1087796736 .align 8 .LC12: .long 0 .long 1064304640 .align 8 .LC13: .long 0 .long 1127219200 .align 8 .LC15: .long 0 .long 1072693248 .align 8 .LC17: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <cuda.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device cudaMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); cudaMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, cudaMemcpyHostToDevice); //allocate the histogram data on the device cudaMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); cudaMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,cudaMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu cudaMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, cudaMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); cudaFree(d_gpu_histogram); cudaFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <hip/hip_runtime.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device hipMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); hipMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, hipMemcpyHostToDevice); //allocate the histogram data on the device hipMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); hipMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,hipMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu hipMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, hipMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); hipFree(d_gpu_histogram); hipFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <hip/hip_runtime.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device hipMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); hipMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, hipMemcpyHostToDevice); //allocate the histogram data on the device hipMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); hipMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,hipMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu hipMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, hipMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); hipFree(d_gpu_histogram); hipFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10PDH_kernelP10hist_entryP8atomdescxd .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .p2align 8 .type _Z10PDH_kernelP10hist_entryP8atomdescxd,@function _Z10PDH_kernelP10hist_entryP8atomdescxd: s_clause 0x1 s_load_b32 s4, s[0:1], 0x2c s_load_b64 s[2:3], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) s_mul_i32 s15, s15, s4 s_mov_b32 s4, exec_lo v_add_nc_u32_e32 v2, s15, v0 v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i64_e64 s[2:3], v[2:3] s_cbranch_execz .LBB0_4 v_add_nc_u32_e32 v4, 1, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v5, 31, v4 v_cmp_gt_i64_e32 vcc_lo, s[2:3], v[4:5] s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_4 s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x18 v_add3_u32 v3, v0, s15, 2 s_mov_b32 s8, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_ashrrev_i32_e32 v6, 31, v3 s_waitcnt lgkmcnt(0) v_mad_i64_i32 v[0:1], null, v2, 24, s[6:7] v_add_co_u32 v2, vcc_lo, v3, -1 v_add_co_ci_u32_e32 v3, vcc_lo, -1, v6, vcc_lo .LBB0_3: v_mad_u64_u32 v[8:9], null, v4, 24, s[6:7] s_delay_alu instid0(VALU_DEP_4) v_add_co_u32 v10, vcc_lo, v0, 16 v_add_co_ci_u32_e32 v11, vcc_lo, 0, v1, vcc_lo s_clause 0x1 global_load_b64 v[12:13], v[0:1], off global_load_b64 v[14:15], v[10:11], off offset:-8 v_mov_b32_e32 v4, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[6:7], null, v5, 24, v[4:5] v_mov_b32_e32 v9, v6 s_clause 0x2 global_load_b128 v[4:7], v[8:9], off global_load_b64 v[10:11], v[10:11], off global_load_b64 v[8:9], v[8:9], off offset:16 s_waitcnt vmcnt(2) v_add_f64 v[6:7], v[14:15], -v[6:7] v_add_f64 v[4:5], v[12:13], -v[4:5] s_waitcnt vmcnt(0) v_add_f64 v[8:9], v[10:11], -v[8:9] s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[6:7], v[6:7], v[6:7] v_fma_f64 v[4:5], v[4:5], v[4:5], v[6:7] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[4:5], v[8:9], v[8:9], v[4:5] v_cmp_gt_f64_e32 vcc_lo, 0x10000000, v[4:5] v_cndmask_b32_e64 v6, 0, 1, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b32_e32 v6, 8, v6 v_ldexp_f64 v[4:5], v[4:5], v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_rsq_f64_e32 v[6:7], v[4:5] s_waitcnt_depctr 0xfff v_mul_f64 v[8:9], v[4:5], v[6:7] v_mul_f64 v[6:7], v[6:7], 0.5 v_fma_f64 v[10:11], -v[6:7], v[8:9], 0.5 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] v_fma_f64 v[6:7], v[6:7], v[10:11], v[6:7] v_fma_f64 v[10:11], -v[8:9], v[8:9], v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[8:9], v[10:11], v[6:7], v[8:9] v_fma_f64 v[10:11], -v[8:9], v[8:9], v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_fma_f64 v[6:7], v[10:11], v[6:7], v[8:9] v_cndmask_b32_e64 v8, 0, 0xffffff80, vcc_lo v_cmp_class_f64_e64 vcc_lo, v[4:5], 0x260 v_ldexp_f64 v[6:7], v[6:7], v8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v5, v7, v5 :: v_dual_cndmask_b32 v4, v6, v4 v_div_scale_f64 v[6:7], null, s[0:1], s[0:1], v[4:5] v_div_scale_f64 v[12:13], vcc_lo, v[4:5], s[0:1], v[4:5] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f64_e32 v[8:9], v[6:7] s_waitcnt_depctr 0xfff v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[10:11], v[12:13], v[8:9] v_fma_f64 v[6:7], -v[6:7], v[10:11], v[12:13] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fmas_f64 v[6:7], v[6:7], v[8:9], v[10:11] v_div_fixup_f64 v[4:5], v[6:7], s[0:1], v[4:5] v_mov_b32_e32 v6, 1 v_mov_b32_e32 v7, 0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_i32_f64_e32 v4, v[4:5] v_ashrrev_i32_e32 v5, 31, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 3, v[4:5] v_add_co_u32 v4, vcc_lo, s4, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo v_add_co_u32 v2, vcc_lo, v2, 1 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo global_atomic_add_u64 v[4:5], v[6:7], off v_mov_b32_e32 v4, v2 v_ashrrev_i32_e32 v5, 31, v2 v_cmp_le_i64_e32 vcc_lo, s[2:3], v[2:3] s_or_b32 s8, vcc_lo, s8 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s8 s_cbranch_execnz .LBB0_3 .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z10PDH_kernelP10hist_entryP8atomdescxd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10PDH_kernelP10hist_entryP8atomdescxd, .Lfunc_end0-_Z10PDH_kernelP10hist_entryP8atomdescxd .section .AMDGPU.csdata,"",@progbits .text .protected _Z11PDH_kernel2P10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .p2align 8 .type _Z11PDH_kernel2P10hist_entryP8atomdescxd,@function _Z11PDH_kernel2P10hist_entryP8atomdescxd: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11PDH_kernel2P10hist_entryP8atomdescxd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end1: .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, .Lfunc_end1-_Z11PDH_kernel2P10hist_entryP8atomdescxd .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value - .offset: 24 .size: 8 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: hidden_block_count_x - .offset: 36 .size: 4 .value_kind: hidden_block_count_y - .offset: 40 .size: 4 .value_kind: hidden_block_count_z - .offset: 44 .size: 2 .value_kind: hidden_group_size_x - .offset: 46 .size: 2 .value_kind: hidden_group_size_y - .offset: 48 .size: 2 .value_kind: hidden_group_size_z - .offset: 50 .size: 2 .value_kind: hidden_remainder_x - .offset: 52 .size: 2 .value_kind: hidden_remainder_y - .offset: 54 .size: 2 .value_kind: hidden_remainder_z - .offset: 72 .size: 8 .value_kind: hidden_global_offset_x - .offset: 80 .size: 8 .value_kind: hidden_global_offset_y - .offset: 88 .size: 8 .value_kind: hidden_global_offset_z - .offset: 96 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z10PDH_kernelP10hist_entryP8atomdescxd .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10PDH_kernelP10hist_entryP8atomdescxd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value - .offset: 24 .size: 8 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11PDH_kernel2P10hist_entryP8atomdescxd .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z11PDH_kernel2P10hist_entryP8atomdescxd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
/* ================================================================== Programmer: Yicheng Tu (ytu@cse.usf.edu) The basic SDH algorithm implementation for 3D data To compile: nvcc SDH.c -o SDH in the C4 lab machines ================================================================== */ #include <stdio.h> #include <stdlib.h> #include <math.h> #include <sys/time.h> #include <hip/hip_runtime.h> #define BOX_SIZE 23000 /* size of the data box on one dimension */ /* descriptors for single atom in the tree */ typedef struct atomdesc { double x_pos; double y_pos; double z_pos; } atom; typedef struct hist_entry{ //float min; //float max; long long d_cnt; /* need a long long type as the count might be huge */ } bucket; bucket * histogram; /* list of all buckets in the histogram */ bucket * h_gpu_histogram; bucket * d_gpu_histogram; bucket * diff_histogram; long long PDH_acnt; /* total number of data points */ int num_buckets; /* total number of buckets in the histogram */ double PDH_res; /* value of w */ atom * atom_list; /* list of all data points */ atom * d_atom_list; /* These are for an old way of tracking time */ struct timezone Idunno; struct timeval startTime, endTime; /* distance of two points in the atom_list */ double p2p_distance(int ind1, int ind2) { double x1 = atom_list[ind1].x_pos; double x2 = atom_list[ind2].x_pos; double y1 = atom_list[ind1].y_pos; double y2 = atom_list[ind2].y_pos; double z1 = atom_list[ind1].z_pos; double z2 = atom_list[ind2].z_pos; return sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); } //get cuda error // void cuda_err(cudaError_t e, char in[]) // { // if (e != cudaSuccess){ // printf("CUDA Error: %s, %s \n", in, cudaGetErrorString(e)); // exit(EXIT_FAILURE); // } // } /* brute-force SDH solution in a single CPU thread */ int PDH_baseline() { int i, j, h_pos; double dist; for(i = 0; i < PDH_acnt; i++) { for(j = i+1; j < PDH_acnt; j++) { dist = p2p_distance(i,j); h_pos = (int) (dist / PDH_res); histogram[h_pos].d_cnt++; } } return 0; } /* SDH kernel - a really crappy one */ __global__ void PDH_kernel(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { int id = blockIdx.x*blockDim.x + threadIdx.x; int j, h_pos; double dist; double x1; double x2; double y1; double y2; double z1; double z2; if(id < acnt) for(j = id+1; j < acnt; j++) { x1 = d_atom_list[id].x_pos; x2 = d_atom_list[j].x_pos; y1 = d_atom_list[id].y_pos; y2 = d_atom_list[j].y_pos; z1 = d_atom_list[id].z_pos; z2 = d_atom_list[j].z_pos; dist = sqrt((x1 - x2)*(x1-x2) + (y1 - y2)*(y1 - y2) + (z1 - z2)*(z1 - z2)); h_pos = (int) (dist / res); // __syncthreads(); // d_histogram[h_pos].d_cnt += 1; //very odd that this doesnt work but atomicAdd does... I wonder why //long story short... it's basically because having synchronized threads //at this point doesnt prevent race condiions on the increment itself // __syncthreads(); atomicAdd((unsigned long long int*)&d_histogram[h_pos].d_cnt,1); } } /* SDH kernel 2.0 - now the goal is to make a better kernel, that is at least 10x faster (at first) */ __global__ void PDH_kernel2(bucket* d_histogram, atom* d_atom_list, long long acnt, double res) { } /* set a checkpoint and show the (natural) running time in seconds */ double report_running_time() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for CPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } double report_running_time_GPU() { long sec_diff, usec_diff; gettimeofday(&endTime, &Idunno); sec_diff = endTime.tv_sec - startTime.tv_sec; usec_diff= endTime.tv_usec-startTime.tv_usec; if(usec_diff < 0) { sec_diff --; usec_diff += 1000000; } printf("Running time for GPU version: %ld.%06ld\n", sec_diff, usec_diff); return (double)(sec_diff*1.0 + usec_diff/1000000.0); } /* print the counts in all buckets of the histogram */ void output_histogram(bucket* histogram){ int i; long long total_cnt = 0; for(i=0; i< num_buckets; i++) { if(i%5 == 0) /* we print 5 buckets in a row */ printf("\n%02d: ", i); printf("%15lld ", histogram[i].d_cnt); total_cnt += histogram[i].d_cnt; /* we also want to make sure the total distance count is correct */ if(i == num_buckets - 1) printf("\n T:%lld \n", total_cnt); else printf("| "); } } int main(int argc, char **argv) { int i; PDH_acnt = atoi(argv[1]); PDH_res = atof(argv[2]); //printf("args are %d and %f\n", PDH_acnt, PDH_res); num_buckets = (int)(BOX_SIZE * 1.732 / PDH_res) + 1; histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); atom_list = (atom *)malloc(sizeof(atom)*PDH_acnt); srand(1); /* generate data following a uniform distribution */ for(i = 0; i < PDH_acnt; i++) { atom_list[i].x_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].y_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; atom_list[i].z_pos = ((double)(rand()) / RAND_MAX) * BOX_SIZE; } /* start counting time */ printf("Starting CPU...\n"); gettimeofday(&startTime, &Idunno); /* call CPU single thread version to compute the histogram */ PDH_baseline(); /* check the total running time */ report_running_time(); /* print out the histogram */ output_histogram(histogram); printf("Starting GPU...\n"); //cudaDeviceReset(); //gpu code-------------------------------------------------------------------------------- h_gpu_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); //copy the atomlist over from host to device hipMalloc((void**)&d_atom_list, sizeof(atom)*PDH_acnt); hipMemcpy(d_atom_list, atom_list, sizeof(atom)*PDH_acnt, hipMemcpyHostToDevice); //allocate the histogram data on the device hipMalloc((void**)&d_gpu_histogram, sizeof(bucket)*num_buckets); hipMemcpy(d_gpu_histogram, h_gpu_histogram, sizeof(bucket)*num_buckets,hipMemcpyHostToDevice); //start the timer gettimeofday(&startTime, &Idunno); //run the kernel PDH_kernel<<<ceil(PDH_acnt/256.0), 256>>>(d_gpu_histogram, d_atom_list, PDH_acnt, PDH_res); //copy the histogram results back from gpu over to cpu hipMemcpy(h_gpu_histogram, d_gpu_histogram, sizeof(bucket)*num_buckets, hipMemcpyDeviceToHost); //check total running time report_running_time_GPU(); //print out the resulting histogram from the GPU output_histogram(h_gpu_histogram); //difference calculation-------------------------------------------------------------------------------- printf("Difference: \n"); diff_histogram = (bucket *)malloc(sizeof(bucket)*num_buckets); int bi; for(bi = 0; bi < num_buckets; bi++) { diff_histogram[bi].d_cnt = histogram[bi].d_cnt - h_gpu_histogram[bi].d_cnt; } output_histogram(diff_histogram); hipFree(d_gpu_histogram); hipFree(d_atom_list); free(histogram); free(atom_list); free(h_gpu_histogram); free(diff_histogram); return 0; }
.text .file "proj1-tanmay.hip" .globl _Z12p2p_distanceii # -- Begin function _Z12p2p_distanceii .p2align 4, 0x90 .type _Z12p2p_distanceii,@function _Z12p2p_distanceii: # @_Z12p2p_distanceii .cfi_startproc # %bb.0: movq atom_list(%rip), %rax movslq %edi, %rcx leaq (%rcx,%rcx,2), %rcx movslq %esi, %rdx leaq (%rdx,%rdx,2), %rdx movsd (%rax,%rcx,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%rcx,8), %xmm2 # xmm2 = mem[0],zero movsd 16(%rax,%rcx,8), %xmm0 # xmm0 = mem[0],zero subsd (%rax,%rdx,8), %xmm1 mulsd %xmm1, %xmm1 subsd 8(%rax,%rdx,8), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd 16(%rax,%rdx,8), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 xorpd %xmm1, %xmm1 ucomisd %xmm1, %xmm0 jb sqrt # TAILCALL # %bb.1: # %.split sqrtsd %xmm0, %xmm0 retq .Lfunc_end0: .size _Z12p2p_distanceii, .Lfunc_end0-_Z12p2p_distanceii .cfi_endproc # -- End function .globl _Z12PDH_baselinev # -- Begin function _Z12PDH_baselinev .p2align 4, 0x90 .type _Z12PDH_baselinev,@function _Z12PDH_baselinev: # @_Z12PDH_baselinev .cfi_startproc # %bb.0: cmpq $0, PDH_acnt(%rip) jle .LBB1_10 # %bb.1: # %.lr.ph11.preheader pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 pushq %rax .cfi_def_cfa_offset 64 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $1, %ebx movl $40, %r14d xorl %r15d, %r15d xorpd %xmm3, %xmm3 jmp .LBB1_3 .p2align 4, 0x90 .LBB1_2: # %.loopexit # in Loop: Header=BB1_3 Depth=1 incq %rbx addq $24, %r14 cmpq %r15, PDH_acnt(%rip) jle .LBB1_9 .LBB1_3: # %.lr.ph11 # =>This Loop Header: Depth=1 # Child Loop BB1_5 Depth 2 movq %r15, %rax incq %r15 cmpq %r15, PDH_acnt(%rip) jle .LBB1_2 # %bb.4: # %.lr.ph.preheader # in Loop: Header=BB1_3 Depth=1 leaq (%rax,%rax,2), %r12 movq %r14, %r13 movq %rbx, %rbp jmp .LBB1_5 .p2align 4, 0x90 .LBB1_7: # %call.sqrt # in Loop: Header=BB1_5 Depth=2 callq sqrt xorpd %xmm3, %xmm3 .LBB1_8: # %.lr.ph.split # in Loop: Header=BB1_5 Depth=2 divsd PDH_res(%rip), %xmm0 cvttsd2si %xmm0, %eax movq histogram(%rip), %rcx cltq incq (%rcx,%rax,8) incq %rbp addq $24, %r13 cmpq %rbp, PDH_acnt(%rip) jle .LBB1_2 .LBB1_5: # %.lr.ph # Parent Loop BB1_3 Depth=1 # => This Inner Loop Header: Depth=2 movq atom_list(%rip), %rax movsd (%rax,%r12,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%r12,8), %xmm2 # xmm2 = mem[0],zero subsd -16(%rax,%r13), %xmm1 movsd 16(%rax,%r12,8), %xmm0 # xmm0 = mem[0],zero mulsd %xmm1, %xmm1 subsd -8(%rax,%r13), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd (%rax,%r13), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 ucomisd %xmm3, %xmm0 jb .LBB1_7 # %bb.6: # in Loop: Header=BB1_5 Depth=2 sqrtsd %xmm0, %xmm0 jmp .LBB1_8 .LBB1_9: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 .cfi_restore %rbx .cfi_restore %r12 .cfi_restore %r13 .cfi_restore %r14 .cfi_restore %r15 .cfi_restore %rbp .LBB1_10: # %._crit_edge xorl %eax, %eax retq .Lfunc_end1: .size _Z12PDH_baselinev, .Lfunc_end1-_Z12PDH_baselinev .cfi_endproc # -- End function .globl _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd # -- Begin function _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .p2align 4, 0x90 .type _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd,@function _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd: # @_Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd, .Lfunc_end2-_Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .cfi_endproc # -- End function .globl _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd # -- Begin function _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .p2align 4, 0x90 .type _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd,@function _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd: # @_Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z11PDH_kernel2P10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end3: .size _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd, .Lfunc_end3-_Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z19report_running_timev .LCPI4_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl _Z19report_running_timev .p2align 4, 0x90 .type _Z19report_running_timev,@function _Z19report_running_timev: # @_Z19report_running_timev .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rbx movq %rcx, %r14 sarq $63, %r14 addq %rax, %r14 testq %rcx, %rcx cmovnsq %rcx, %rbx movl $.L.str, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf cvtsi2sd %r14, %xmm1 cvtsi2sd %rbx, %xmm0 divsd .LCPI4_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size _Z19report_running_timev, .Lfunc_end4-_Z19report_running_timev .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z23report_running_time_GPUv .LCPI5_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl _Z23report_running_time_GPUv .p2align 4, 0x90 .type _Z23report_running_time_GPUv,@function _Z23report_running_time_GPUv: # @_Z23report_running_time_GPUv .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rbx movq %rcx, %r14 sarq $63, %r14 addq %rax, %r14 testq %rcx, %rcx cmovnsq %rcx, %rbx movl $.L.str.1, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf cvtsi2sd %r14, %xmm1 cvtsi2sd %rbx, %xmm0 divsd .LCPI5_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end5: .size _Z23report_running_time_GPUv, .Lfunc_end5-_Z23report_running_time_GPUv .cfi_endproc # -- End function .globl _Z16output_histogramP10hist_entry # -- Begin function _Z16output_histogramP10hist_entry .p2align 4, 0x90 .type _Z16output_histogramP10hist_entry,@function _Z16output_histogramP10hist_entry: # @_Z16output_histogramP10hist_entry .cfi_startproc # %bb.0: cmpl $0, num_buckets(%rip) jle .LBB6_9 # %bb.1: # %.lr.ph.preheader pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 pushq %rax .cfi_def_cfa_offset 48 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdi, %rbx movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %r14d, %r14d xorl %r15d, %r15d jmp .LBB6_2 .p2align 4, 0x90 .LBB6_5: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.4, %edi movq %r15, %rsi xorl %eax, %eax callq printf .LBB6_7: # in Loop: Header=BB6_2 Depth=1 incq %r14 movslq num_buckets(%rip), %rax cmpq %rax, %r14 jge .LBB6_8 .LBB6_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl %r14d, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %r14d jne .LBB6_4 # %bb.3: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.2, %edi movl %r14d, %esi xorl %eax, %eax callq printf .LBB6_4: # in Loop: Header=BB6_2 Depth=1 movq (%rbx,%r14,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%rbx,%r14,8), %r15 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %r14 je .LBB6_5 # %bb.6: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf jmp .LBB6_7 .LBB6_8: addq $8, %rsp .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 .cfi_restore %rbx .cfi_restore %r12 .cfi_restore %r14 .cfi_restore %r15 .LBB6_9: # %._crit_edge retq .Lfunc_end6: .size _Z16output_histogramP10hist_entry, .Lfunc_end6-_Z16output_histogramP10hist_entry .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI7_0: .quad 0x40e3738000000000 # double 39836 .LCPI7_1: .quad 0x41dfffffffc00000 # double 2147483647 .LCPI7_2: .quad 0x40d6760000000000 # double 23000 .LCPI7_3: .quad 0x3f70000000000000 # double 0.00390625 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $120, %rsp .cfi_def_cfa_offset 176 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movq %rsi, %rbx movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol cltq movq %rax, PDH_acnt(%rip) movq 16(%rbx), %rdi xorl %esi, %esi callq strtod movsd %xmm0, PDH_res(%rip) movsd .LCPI7_0(%rip), %xmm1 # xmm1 = mem[0],zero divsd %xmm0, %xmm1 cvttsd2si %xmm1, %eax leal 1(%rax), %ecx movl %ecx, num_buckets(%rip) cltq leaq 8(,%rax,8), %rdi callq malloc movq %rax, histogram(%rip) movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rdi callq malloc movq %rax, atom_list(%rip) movl $1, %edi callq srand cmpq $0, PDH_acnt(%rip) jle .LBB7_3 # %bb.1: # %.lr.ph.preheader xorl %ebx, %ebx xorl %r14d, %r14d .p2align 4, 0x90 .LBB7_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 movsd .LCPI7_1(%rip), %xmm1 # xmm1 = mem[0],zero divsd %xmm1, %xmm0 movsd .LCPI7_2(%rip), %xmm1 # xmm1 = mem[0],zero mulsd %xmm1, %xmm0 movq atom_list(%rip), %rax movsd %xmm0, (%rax,%rbx) callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 divsd .LCPI7_1(%rip), %xmm0 mulsd .LCPI7_2(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 8(%rax,%rbx) callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 divsd .LCPI7_1(%rip), %xmm0 mulsd .LCPI7_2(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 16(%rax,%rbx) incq %r14 addq $24, %rbx cmpq %r14, PDH_acnt(%rip) jg .LBB7_2 .LBB7_3: # %._crit_edge movl $.Lstr, %edi callq puts@PLT movl $startTime, %edi movl $Idunno, %esi callq gettimeofday cmpq $0, PDH_acnt(%rip) jle .LBB7_12 # %bb.4: # %.lr.ph11.i.preheader movl $1, %ebx movl $40, %r14d xorl %r15d, %r15d xorpd %xmm3, %xmm3 jmp .LBB7_6 .p2align 4, 0x90 .LBB7_5: # %.loopexit.i # in Loop: Header=BB7_6 Depth=1 incq %rbx addq $24, %r14 cmpq %r15, PDH_acnt(%rip) jle .LBB7_12 .LBB7_6: # %.lr.ph11.i # =>This Loop Header: Depth=1 # Child Loop BB7_8 Depth 2 movq %r15, %rax incq %r15 cmpq %r15, PDH_acnt(%rip) jle .LBB7_5 # %bb.7: # %.lr.ph.i.preheader # in Loop: Header=BB7_6 Depth=1 leaq (%rax,%rax,2), %r12 movq %r14, %r13 movq %rbx, %rbp jmp .LBB7_8 .p2align 4, 0x90 .LBB7_10: # %call.sqrt # in Loop: Header=BB7_8 Depth=2 callq sqrt xorpd %xmm3, %xmm3 .LBB7_11: # %.lr.ph.i.split # in Loop: Header=BB7_8 Depth=2 divsd PDH_res(%rip), %xmm0 cvttsd2si %xmm0, %eax movq histogram(%rip), %rcx cltq incq (%rcx,%rax,8) incq %rbp addq $24, %r13 cmpq %rbp, PDH_acnt(%rip) jle .LBB7_5 .LBB7_8: # %.lr.ph.i # Parent Loop BB7_6 Depth=1 # => This Inner Loop Header: Depth=2 movq atom_list(%rip), %rax movsd (%rax,%r12,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%r12,8), %xmm2 # xmm2 = mem[0],zero subsd -16(%rax,%r13), %xmm1 movsd 16(%rax,%r12,8), %xmm0 # xmm0 = mem[0],zero mulsd %xmm1, %xmm1 subsd -8(%rax,%r13), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd (%rax,%r13), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 ucomisd %xmm3, %xmm0 jb .LBB7_10 # %bb.9: # in Loop: Header=BB7_8 Depth=2 sqrtsd %xmm0, %xmm0 jmp .LBB7_11 .LBB7_12: # %_Z12PDH_baselinev.exit movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rdx movq %rcx, %rsi sarq $63, %rsi addq %rax, %rsi testq %rcx, %rcx cmovnsq %rcx, %rdx movl $.L.str, %edi xorl %eax, %eax callq printf cmpl $0, num_buckets(%rip) jle .LBB7_20 # %bb.13: # %.lr.ph.i17.preheader movq histogram(%rip), %r15 movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %ebx, %ebx xorl %r14d, %r14d jmp .LBB7_14 .p2align 4, 0x90 .LBB7_18: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_19: # in Loop: Header=BB7_14 Depth=1 incq %rbx movslq num_buckets(%rip), %rax cmpq %rax, %rbx jge .LBB7_20 .LBB7_14: # %.lr.ph.i17 # =>This Inner Loop Header: Depth=1 movl %ebx, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %ebx jne .LBB7_16 # %bb.15: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.2, %edi movl %ebx, %esi xorl %eax, %eax callq printf .LBB7_16: # in Loop: Header=BB7_14 Depth=1 movq (%r15,%rbx,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%r15,%rbx,8), %r14 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %rbx jne .LBB7_18 # %bb.17: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq printf jmp .LBB7_19 .LBB7_20: # %_Z16output_histogramP10hist_entry.exit movl $.Lstr.1, %edi callq puts@PLT movslq num_buckets(%rip), %rdi shlq $3, %rdi callq malloc movq %rax, h_gpu_histogram(%rip) movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rsi movl $d_atom_list, %edi callq hipMalloc movq d_atom_list(%rip), %rdi movq atom_list(%rip), %rsi movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rdx movl $1, %ecx callq hipMemcpy movslq num_buckets(%rip), %rsi shlq $3, %rsi movl $d_gpu_histogram, %edi callq hipMalloc movq d_gpu_histogram(%rip), %rdi movq h_gpu_histogram(%rip), %rsi movslq num_buckets(%rip), %rdx shlq $3, %rdx movl $1, %ecx callq hipMemcpy movl $startTime, %edi movl $Idunno, %esi callq gettimeofday xorps %xmm0, %xmm0 cvtsi2sdq PDH_acnt(%rip), %xmm0 mulsd .LCPI7_3(%rip), %xmm0 callq ceil@PLT cvttsd2si %xmm0, %rax movl %eax, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $256, %rdx # imm = 0x100 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB7_22 # %bb.21: movq d_gpu_histogram(%rip), %rax movq d_atom_list(%rip), %rcx movq PDH_acnt(%rip), %rdx movsd PDH_res(%rip), %xmm0 # xmm0 = mem[0],zero movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB7_22: movq h_gpu_histogram(%rip), %rdi movq d_gpu_histogram(%rip), %rsi movslq num_buckets(%rip), %rdx shlq $3, %rdx movl $2, %ecx callq hipMemcpy movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rdx movq %rcx, %rsi sarq $63, %rsi addq %rax, %rsi testq %rcx, %rcx cmovnsq %rcx, %rdx movl $.L.str.1, %edi xorl %eax, %eax callq printf cmpl $0, num_buckets(%rip) jle .LBB7_30 # %bb.23: # %.lr.ph.i23.preheader movq h_gpu_histogram(%rip), %r15 movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %ebx, %ebx xorl %r14d, %r14d jmp .LBB7_24 .p2align 4, 0x90 .LBB7_28: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_29: # in Loop: Header=BB7_24 Depth=1 incq %rbx movslq num_buckets(%rip), %rax cmpq %rax, %rbx jge .LBB7_30 .LBB7_24: # %.lr.ph.i23 # =>This Inner Loop Header: Depth=1 movl %ebx, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %ebx jne .LBB7_26 # %bb.25: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.2, %edi movl %ebx, %esi xorl %eax, %eax callq printf .LBB7_26: # in Loop: Header=BB7_24 Depth=1 movq (%r15,%rbx,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%r15,%rbx,8), %r14 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %rbx jne .LBB7_28 # %bb.27: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq printf jmp .LBB7_29 .LBB7_30: # %_Z16output_histogramP10hist_entry.exit27 movl $.Lstr.2, %edi callq puts@PLT movslq num_buckets(%rip), %r14 leaq (,%r14,8), %rdi callq malloc movq %rax, %rbx movl %r14d, %eax movq %rbx, diff_histogram(%rip) testq %r14, %r14 jle .LBB7_33 # %bb.31: # %.lr.ph39 movq histogram(%rip), %rcx movq h_gpu_histogram(%rip), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB7_32: # =>This Inner Loop Header: Depth=1 movq (%rcx,%rsi,8), %rdi subq (%rdx,%rsi,8), %rdi movq %rdi, (%rbx,%rsi,8) incq %rsi cmpq %rsi, %rax jne .LBB7_32 .LBB7_33: # %._crit_edge40 testl %eax, %eax jle .LBB7_41 # %bb.34: # %.lr.ph.i28.preheader movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %r14d, %r14d xorl %r15d, %r15d jmp .LBB7_35 .p2align 4, 0x90 .LBB7_39: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_40: # in Loop: Header=BB7_35 Depth=1 incq %r14 movslq num_buckets(%rip), %rax cmpq %rax, %r14 jge .LBB7_41 .LBB7_35: # %.lr.ph.i28 # =>This Inner Loop Header: Depth=1 movl %r14d, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %r14d jne .LBB7_37 # %bb.36: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.2, %edi movl %r14d, %esi xorl %eax, %eax callq printf .LBB7_37: # in Loop: Header=BB7_35 Depth=1 movq (%rbx,%r14,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%rbx,%r14,8), %r15 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %r14 jne .LBB7_39 # %bb.38: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.4, %edi movq %r15, %rsi xorl %eax, %eax callq printf jmp .LBB7_40 .LBB7_41: # %_Z16output_histogramP10hist_entry.exit32 movq d_gpu_histogram(%rip), %rdi callq hipFree movq d_atom_list(%rip), %rdi callq hipFree movq histogram(%rip), %rdi callq free movq atom_list(%rip), %rdi callq free movq h_gpu_histogram(%rip), %rdi callq free movq diff_histogram(%rip), %rdi callq free xorl %eax, %eax addq $120, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end7: .size main, .Lfunc_end7-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB8_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB8_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11PDH_kernel2P10hist_entryP8atomdescxd, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end8: .size __hip_module_ctor, .Lfunc_end8-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB9_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB9_2: retq .Lfunc_end9: .size __hip_module_dtor, .Lfunc_end9-__hip_module_dtor .cfi_endproc # -- End function .type histogram,@object # @histogram .bss .globl histogram .p2align 3, 0x0 histogram: .quad 0 .size histogram, 8 .type h_gpu_histogram,@object # @h_gpu_histogram .globl h_gpu_histogram .p2align 3, 0x0 h_gpu_histogram: .quad 0 .size h_gpu_histogram, 8 .type d_gpu_histogram,@object # @d_gpu_histogram .globl d_gpu_histogram .p2align 3, 0x0 d_gpu_histogram: .quad 0 .size d_gpu_histogram, 8 .type diff_histogram,@object # @diff_histogram .globl diff_histogram .p2align 3, 0x0 diff_histogram: .quad 0 .size diff_histogram, 8 .type PDH_acnt,@object # @PDH_acnt .globl PDH_acnt .p2align 3, 0x0 PDH_acnt: .quad 0 # 0x0 .size PDH_acnt, 8 .type num_buckets,@object # @num_buckets .globl num_buckets .p2align 2, 0x0 num_buckets: .long 0 # 0x0 .size num_buckets, 4 .type PDH_res,@object # @PDH_res .globl PDH_res .p2align 3, 0x0 PDH_res: .quad 0x0000000000000000 # double 0 .size PDH_res, 8 .type atom_list,@object # @atom_list .globl atom_list .p2align 3, 0x0 atom_list: .quad 0 .size atom_list, 8 .type d_atom_list,@object # @d_atom_list .globl d_atom_list .p2align 3, 0x0 d_atom_list: .quad 0 .size d_atom_list, 8 .type Idunno,@object # @Idunno .globl Idunno .p2align 2, 0x0 Idunno: .zero 8 .size Idunno, 8 .type startTime,@object # @startTime .globl startTime .p2align 3, 0x0 startTime: .zero 16 .size startTime, 16 .type endTime,@object # @endTime .globl endTime .p2align 3, 0x0 endTime: .zero 16 .size endTime, 16 .type _Z10PDH_kernelP10hist_entryP8atomdescxd,@object # @_Z10PDH_kernelP10hist_entryP8atomdescxd .section .rodata,"a",@progbits .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .p2align 3, 0x0 _Z10PDH_kernelP10hist_entryP8atomdescxd: .quad _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .size _Z10PDH_kernelP10hist_entryP8atomdescxd, 8 .type _Z11PDH_kernel2P10hist_entryP8atomdescxd,@object # @_Z11PDH_kernel2P10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .p2align 3, 0x0 _Z11PDH_kernel2P10hist_entryP8atomdescxd: .quad _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Running time for CPU version: %ld.%06ld\n" .size .L.str, 41 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Running time for GPU version: %ld.%06ld\n" .size .L.str.1, 41 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "\n%02d: " .size .L.str.2, 8 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "%15lld " .size .L.str.3, 8 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "\n T:%lld \n" .size .L.str.4, 11 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "| " .size .L.str.5, 3 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10PDH_kernelP10hist_entryP8atomdescxd" .size .L__unnamed_1, 40 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z11PDH_kernel2P10hist_entryP8atomdescxd" .size .L__unnamed_2, 41 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Starting CPU..." .size .Lstr, 16 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Starting GPU..." .size .Lstr.1, 16 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Difference: " .size .Lstr.2, 13 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .addrsig_sym _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym d_gpu_histogram .addrsig_sym d_atom_list .addrsig_sym Idunno .addrsig_sym startTime .addrsig_sym endTime .addrsig_sym _Z10PDH_kernelP10hist_entryP8atomdescxd .addrsig_sym _Z11PDH_kernel2P10hist_entryP8atomdescxd .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z11PDH_kernel2P10hist_entryP8atomdescxd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z10PDH_kernelP10hist_entryP8atomdescxd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_CTAID.X ; /* 0x0000000000087919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R8, R8, c[0x0][0x0], R3 ; /* 0x0000000008087a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R8, c[0x0][0x170], PT ; /* 0x00005c0008007a0c */ /* 0x000fe40003f06070 */ /*0050*/ SHF.R.S32.HI R2, RZ, 0x1f, R8 ; /* 0x0000001fff027819 */ /* 0x000fc80000011408 */ /*0060*/ ISETP.GE.AND.EX P0, PT, R2, c[0x0][0x174], PT, P0 ; /* 0x00005d0002007a0c */ /* 0x000fda0003f06300 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ IADD3 R0, R8, 0x1, RZ ; /* 0x0000000108007810 */ /* 0x000fc80007ffe0ff */ /*0090*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe40003f06070 */ /*00a0*/ SHF.R.S32.HI R3, RZ, 0x1f, R0 ; /* 0x0000001fff037819 */ /* 0x000fc80000011400 */ /*00b0*/ ISETP.GE.AND.EX P0, PT, R3, c[0x0][0x174], PT, P0 ; /* 0x00005d0003007a0c */ /* 0x000fda0003f06300 */ /*00c0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00d0*/ IMAD.MOV.U32 R9, RZ, RZ, 0x18 ; /* 0x00000018ff097424 */ /* 0x000fe200078e00ff */ /*00e0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00f0*/ IMAD R5, R2, 0x18, RZ ; /* 0x0000001802057824 */ /* 0x000fe400078e02ff */ /*0100*/ IMAD.WIDE.U32 R8, R8, R9, c[0x0][0x168] ; /* 0x00005a0008087625 */ /* 0x000fc800078e0009 */ /*0110*/ IMAD.MOV.U32 R11, RZ, RZ, R3 ; /* 0x000000ffff0b7224 */ /* 0x000fe400078e0003 */ /*0120*/ IMAD.MOV.U32 R10, RZ, RZ, R0 ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e0000 */ /*0130*/ IMAD.IADD R9, R9, 0x1, R5 ; /* 0x0000000109097824 */ /* 0x000fc800078e0205 */ /*0140*/ IMAD.MOV.U32 R3, RZ, RZ, 0x18 ; /* 0x00000018ff037424 */ /* 0x001fe200078e00ff */ /*0150*/ LDG.E.64 R12, [R8.64+0x8] ; /* 0x00000804080c7981 */ /* 0x000ea2000c1e1b00 */ /*0160*/ IMAD R11, R11, 0x18, RZ ; /* 0x000000180b0b7824 */ /* 0x000fe400078e02ff */ /*0170*/ IMAD.WIDE.U32 R2, R10, R3, c[0x0][0x168] ; /* 0x00005a000a027625 */ /* 0x000fe200078e0003 */ /*0180*/ LDG.E.64 R6, [R8.64] ; /* 0x0000000408067981 */ /* 0x000ee6000c1e1b00 */ /*0190*/ IMAD.IADD R3, R3, 0x1, R11 ; /* 0x0000000103037824 */ /* 0x000fe200078e020b */ /*01a0*/ LDG.E.64 R16, [R8.64+0x10] ; /* 0x0000100408107981 */ /* 0x000f28000c1e1b00 */ /*01b0*/ LDG.E.64 R10, [R2.64+0x8] ; /* 0x00000804020a7981 */ /* 0x000ea8000c1e1b00 */ /*01c0*/ LDG.E.64 R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ee8000c1e1b00 */ /*01d0*/ LDG.E.64 R14, [R2.64+0x10] ; /* 0x00001004020e7981 */ /* 0x000f22000c1e1b00 */ /*01e0*/ YIELD ; /* 0x0000000000007946 */ /* 0x000fe20003800000 */ /*01f0*/ BSSY B0, 0x3b0 ; /* 0x000001b000007945 */ /* 0x000fe20003800000 */ /*0200*/ DADD R10, -R10, R12 ; /* 0x000000000a0a7229 */ /* 0x004e08000000010c */ /*0210*/ DADD R4, -R4, R6 ; /* 0x0000000004047229 */ /* 0x008fc80000000106 */ /*0220*/ DMUL R10, R10, R10 ; /* 0x0000000a0a0a7228 */ /* 0x001e080000000000 */ /*0230*/ DADD R14, -R14, R16 ; /* 0x000000000e0e7229 */ /* 0x010fc80000000110 */ /*0240*/ DFMA R6, R4, R4, R10 ; /* 0x000000040406722b */ /* 0x001e0c000000000a */ /*0250*/ DFMA R6, R14, R14, R6 ; /* 0x0000000e0e06722b */ /* 0x001e0c0000000006 */ /*0260*/ MUFU.RSQ64H R5, R7 ; /* 0x0000000700057308 */ /* 0x001e280000001c00 */ /*0270*/ IADD3 R4, R7, -0x3500000, RZ ; /* 0xfcb0000007047810 */ /* 0x000fe20007ffe0ff */ /*0280*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0290*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3fd80000 ; /* 0x3fd80000ff0d7424 */ /* 0x000fc600078e00ff */ /*02a0*/ DMUL R10, R4, R4 ; /* 0x00000004040a7228 */ /* 0x001e0c0000000000 */ /*02b0*/ DFMA R10, R6, -R10, 1 ; /* 0x3ff00000060a742b */ /* 0x001e0c000000080a */ /*02c0*/ DFMA R2, R10, R12, 0.5 ; /* 0x3fe000000a02742b */ /* 0x001fc8000000000c */ /*02d0*/ DMUL R10, R4, R10 ; /* 0x0000000a040a7228 */ /* 0x000e220000000000 */ /*02e0*/ ISETP.GE.U32.AND P0, PT, R4, 0x7ca00000, PT ; /* 0x7ca000000400780c */ /* 0x000fca0003f06070 */ /*02f0*/ DFMA R12, R2, R10, R4 ; /* 0x0000000a020c722b */ /* 0x001e0c0000000004 */ /*0300*/ DMUL R14, R6, R12 ; /* 0x0000000c060e7228 */ /* 0x001e080000000000 */ /*0310*/ IADD3 R11, R13, -0x100000, RZ ; /* 0xfff000000d0b7810 */ /* 0x000fe20007ffe0ff */ /*0320*/ IMAD.MOV.U32 R10, RZ, RZ, R12 ; /* 0x000000ffff0a7224 */ /* 0x000fe200078e000c */ /*0330*/ DFMA R16, R14, -R14, R6 ; /* 0x8000000e0e10722b */ /* 0x001e0c0000000006 */ /*0340*/ DFMA R2, R16, R10, R14 ; /* 0x0000000a1002722b */ /* 0x001062000000000e */ /*0350*/ @!P0 BRA 0x3a0 ; /* 0x0000004000008947 */ /* 0x000fea0003800000 */ /*0360*/ MOV R2, 0x380 ; /* 0x0000038000027802 */ /* 0x002fca0000000f00 */ /*0370*/ CALL.REL.NOINC 0xbd0 ; /* 0x0000085000007944 */ /* 0x001fea0003c00000 */ /*0380*/ IMAD.MOV.U32 R2, RZ, RZ, R4 ; /* 0x000000ffff027224 */ /* 0x002fe400078e0004 */ /*0390*/ IMAD.MOV.U32 R3, RZ, RZ, R5 ; /* 0x000000ffff037224 */ /* 0x000fe400078e0005 */ /*03a0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*03b0*/ MUFU.RCP64H R5, c[0x0][0x17c] ; /* 0x00005f0000057b08 */ /* 0x000ea20000001800 */ /*03c0*/ IMAD.MOV.U32 R10, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff0a7624 */ /* 0x001fe200078e00ff */ /*03d0*/ FSETP.GEU.AND P1, PT, |R3|, 6.5827683646048100446e-37, PT ; /* 0x036000000300780b */ /* 0x002fe20003f2e200 */ /*03e0*/ IMAD.MOV.U32 R11, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff0b7624 */ /* 0x000fe200078e00ff */ /*03f0*/ BSSY B0, 0x510 ; /* 0x0000011000007945 */ /* 0x000fe20003800000 */ /*0400*/ IMAD.MOV.U32 R4, RZ, RZ, 0x1 ; /* 0x00000001ff047424 */ /* 0x000fcc00078e00ff */ /*0410*/ DFMA R6, R4, -R10, 1 ; /* 0x3ff000000406742b */ /* 0x004e0c000000080a */ /*0420*/ DFMA R6, R6, R6, R6 ; /* 0x000000060606722b */ /* 0x001e0c0000000006 */ /*0430*/ DFMA R6, R4, R6, R4 ; /* 0x000000060406722b */ /* 0x001e0c0000000004 */ /*0440*/ DFMA R4, R6, -R10, 1 ; /* 0x3ff000000604742b */ /* 0x001e0c000000080a */ /*0450*/ DFMA R4, R6, R4, R6 ; /* 0x000000040604722b */ /* 0x001e0c0000000006 */ /*0460*/ DMUL R6, R4, R2 ; /* 0x0000000204067228 */ /* 0x001e0c0000000000 */ /*0470*/ DFMA R10, R6, -c[0x0][0x178], R2 ; /* 0x80005e00060a7a2b */ /* 0x001e0c0000000002 */ /*0480*/ DFMA R4, R4, R10, R6 ; /* 0x0000000a0404722b */ /* 0x001e140000000006 */ /*0490*/ FFMA R6, RZ, c[0x0][0x17c], R5 ; /* 0x00005f00ff067a23 */ /* 0x001fca0000000005 */ /*04a0*/ FSETP.GT.AND P0, PT, |R6|, 1.469367938527859385e-39, PT ; /* 0x001000000600780b */ /* 0x000fda0003f04200 */ /*04b0*/ @P0 BRA P1, 0x500 ; /* 0x0000004000000947 */ /* 0x000fea0000800000 */ /*04c0*/ IMAD.MOV.U32 R4, RZ, RZ, R2 ; /* 0x000000ffff047224 */ /* 0x000fe200078e0002 */ /*04d0*/ MOV R16, 0x500 ; /* 0x0000050000107802 */ /* 0x000fe20000000f00 */ /*04e0*/ IMAD.MOV.U32 R5, RZ, RZ, R3 ; /* 0x000000ffff057224 */ /* 0x000fe400078e0003 */ /*04f0*/ CALL.REL.NOINC 0x5e0 ; /* 0x000000e000007944 */ /* 0x000fea0003c00000 */ /*0500*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0510*/ F2I.F64.TRUNC R2, R4 ; /* 0x0000000400027311 */ /* 0x000e22000030d100 */ /*0520*/ IADD3 R10, R0, 0x1, RZ ; /* 0x00000001000a7810 */ /* 0x000fe20007ffe0ff */ /*0530*/ IMAD.MOV.U32 R3, RZ, RZ, 0x8 ; /* 0x00000008ff037424 */ /* 0x000fe400078e00ff */ /*0540*/ IMAD.MOV.U32 R6, RZ, RZ, 0x1 ; /* 0x00000001ff067424 */ /* 0x000fe200078e00ff */ /*0550*/ ISETP.GE.U32.AND P0, PT, R10, c[0x0][0x170], PT ; /* 0x00005c000a007a0c */ /* 0x000fe20003f06070 */ /*0560*/ IMAD.MOV.U32 R7, RZ, RZ, 0x0 ; /* 0x00000000ff077424 */ /* 0x000fe200078e00ff */ /*0570*/ SHF.R.S32.HI R11, RZ, 0x1f, R10 ; /* 0x0000001fff0b7819 */ /* 0x000fc8000001140a */ /*0580*/ ISETP.GE.AND.EX P0, PT, R11, c[0x0][0x174], PT, P0 ; /* 0x00005d000b007a0c */ /* 0x000fe20003f06300 */ /*0590*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x001fca00078e0203 */ /*05a0*/ RED.E.ADD.64.STRONG.GPU [R2.64], R6 ; /* 0x000000060200798e */ /* 0x0001e2000c10e584 */ /*05b0*/ IMAD.MOV.U32 R0, RZ, RZ, R10 ; /* 0x000000ffff007224 */ /* 0x000fcc00078e000a */ /*05c0*/ @!P0 BRA 0x140 ; /* 0xfffffb7000008947 */ /* 0x000fea000383ffff */ /*05d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*05e0*/ IMAD.MOV.U32 R18, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff127624 */ /* 0x000fe200078e00ff */ /*05f0*/ FSETP.GEU.AND P2, PT, |R5|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000000500780b */ /* 0x040fe20003f4e200 */ /*0600*/ IMAD.MOV.U32 R17, RZ, RZ, 0x1ca00000 ; /* 0x1ca00000ff117424 */ /* 0x000fe200078e00ff */ /*0610*/ LOP3.LUT R20, R5, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000005147812 */ /* 0x000fe200078ec0ff */ /*0620*/ IMAD.MOV.U32 R10, RZ, RZ, 0x1 ; /* 0x00000001ff0a7424 */ /* 0x000fe200078e00ff */ /*0630*/ FSETP.GEU.AND P0, PT, |R18|.reuse, 1.469367938527859385e-39, PT ; /* 0x001000001200780b */ /* 0x040fe20003f0e200 */ /*0640*/ BSSY B1, 0xb90 ; /* 0x0000054000017945 */ /* 0x000fe20003800000 */ /*0650*/ LOP3.LUT R6, R18.reuse, 0x800fffff, RZ, 0xc0, !PT ; /* 0x800fffff12067812 */ /* 0x040fe400078ec0ff */ /*0660*/ LOP3.LUT R19, R18, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000012137812 */ /* 0x000fe400078ec0ff */ /*0670*/ LOP3.LUT R7, R6, 0x3ff00000, RZ, 0xfc, !PT ; /* 0x3ff0000006077812 */ /* 0x000fe200078efcff */ /*0680*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff067624 */ /* 0x000fe200078e00ff */ /*0690*/ ISETP.GE.U32.AND P1, PT, R20, R19, PT ; /* 0x000000131400720c */ /* 0x000fe20003f26070 */ /*06a0*/ IMAD.MOV.U32 R22, RZ, RZ, R19 ; /* 0x000000ffff167224 */ /* 0x000fe200078e0013 */ /*06b0*/ MOV R21, R20 ; /* 0x0000001400157202 */ /* 0x000fc40000000f00 */ /*06c0*/ SEL R17, R17, 0x63400000, !P1 ; /* 0x6340000011117807 */ /* 0x000fe20004800000 */ /*06d0*/ @!P0 IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff028624 */ /* 0x000fe400078e00ff */ /*06e0*/ @!P0 IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff038624 */ /* 0x000fe200078e00ff */ /*06f0*/ @!P2 LOP3.LUT R12, R17, 0x80000000, R5, 0xf8, !PT ; /* 0x80000000110ca812 */ /* 0x000fc800078ef805 */ /*0700*/ @!P2 LOP3.LUT R13, R12, 0x100000, RZ, 0xfc, !PT ; /* 0x001000000c0da812 */ /* 0x000fe200078efcff */ /*0710*/ @!P0 DMUL R6, R2, 8.98846567431157953865e+307 ; /* 0x7fe0000002068828 */ /* 0x000e220000000000 */ /*0720*/ @!P2 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0ca224 */ /* 0x000fca00078e00ff */ /*0730*/ MUFU.RCP64H R11, R7 ; /* 0x00000007000b7308 */ /* 0x001e280000001800 */ /*0740*/ @!P0 LOP3.LUT R22, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000007168812 */ /* 0x000fc800078ec0ff */ /*0750*/ IADD3 R23, R22, -0x1, RZ ; /* 0xffffffff16177810 */ /* 0x000fe20007ffe0ff */ /*0760*/ DFMA R2, R10, -R6, 1 ; /* 0x3ff000000a02742b */ /* 0x001e0c0000000806 */ /*0770*/ DFMA R14, R2, R2, R2 ; /* 0x00000002020e722b */ /* 0x0010640000000002 */ /*0780*/ LOP3.LUT R3, R17, 0x800fffff, R5, 0xf8, !PT ; /* 0x800fffff11037812 */ /* 0x001fe200078ef805 */ /*0790*/ IMAD.MOV.U32 R2, RZ, RZ, R4 ; /* 0x000000ffff027224 */ /* 0x000fc600078e0004 */ /*07a0*/ DFMA R10, R10, R14, R10 ; /* 0x0000000e0a0a722b */ /* 0x002e08000000000a */ /*07b0*/ @!P2 DFMA R2, R2, 2, -R12 ; /* 0x400000000202a82b */ /* 0x000e48000000080c */ /*07c0*/ DFMA R12, R10, -R6, 1 ; /* 0x3ff000000a0c742b */ /* 0x001e0c0000000806 */ /*07d0*/ @!P2 LOP3.LUT R21, R3, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff000000315a812 */ /* 0x002fe200078ec0ff */ /*07e0*/ DFMA R10, R10, R12, R10 ; /* 0x0000000c0a0a722b */ /* 0x001e06000000000a */ /*07f0*/ IADD3 R14, R21, -0x1, RZ ; /* 0xffffffff150e7810 */ /* 0x000fc60007ffe0ff */ /*0800*/ DMUL R12, R10, R2 ; /* 0x000000020a0c7228 */ /* 0x001e220000000000 */ /*0810*/ ISETP.GT.U32.AND P0, PT, R14, 0x7feffffe, PT ; /* 0x7feffffe0e00780c */ /* 0x000fc80003f04070 */ /*0820*/ ISETP.GT.U32.OR P0, PT, R23, 0x7feffffe, P0 ; /* 0x7feffffe1700780c */ /* 0x000fe20000704470 */ /*0830*/ DFMA R14, R12, -R6, R2 ; /* 0x800000060c0e722b */ /* 0x001e0c0000000002 */ /*0840*/ DFMA R14, R10, R14, R12 ; /* 0x0000000e0a0e722b */ /* 0x00104c000000000c */ /*0850*/ @P0 BRA 0xa00 ; /* 0x000001a000000947 */ /* 0x000fea0003800000 */ /*0860*/ IMAD.IADD R19, R20, 0x1, -R19 ; /* 0x0000000114137824 */ /* 0x003fca00078e0a13 */ /*0870*/ IMNMX R19, R19, -0x46a00000, !PT ; /* 0xb960000013137817 */ /* 0x000fc80007800200 */ /*0880*/ IMNMX R4, R19, 0x46a00000, PT ; /* 0x46a0000013047817 */ /* 0x000fca0003800200 */ /*0890*/ IMAD.IADD R17, R4, 0x1, -R17 ; /* 0x0000000104117824 */ /* 0x000fe400078e0a11 */ /*08a0*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fc600078e00ff */ /*08b0*/ IADD3 R5, R17, 0x7fe00000, RZ ; /* 0x7fe0000011057810 */ /* 0x000fcc0007ffe0ff */ /*08c0*/ DMUL R10, R14, R4 ; /* 0x000000040e0a7228 */ /* 0x000e140000000000 */ /*08d0*/ FSETP.GTU.AND P0, PT, |R11|, 1.469367938527859385e-39, PT ; /* 0x001000000b00780b */ /* 0x001fda0003f0c200 */ /*08e0*/ @P0 BRA 0xb80 ; /* 0x0000029000000947 */ /* 0x000fea0003800000 */ /*08f0*/ DFMA R2, R14, -R6, R2 ; /* 0x800000060e02722b */ /* 0x000e220000000002 */ /*0900*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fd200078e00ff */ /*0910*/ FSETP.NEU.AND P0, PT, R3.reuse, RZ, PT ; /* 0x000000ff0300720b */ /* 0x041fe40003f0d000 */ /*0920*/ LOP3.LUT R2, R3, c[0x0][0x17c], RZ, 0x3c, !PT ; /* 0x00005f0003027a12 */ /* 0x000fc800078e3cff */ /*0930*/ LOP3.LUT R7, R2, 0x80000000, RZ, 0xc0, !PT ; /* 0x8000000002077812 */ /* 0x000fc800078ec0ff */ /*0940*/ LOP3.LUT R5, R7, R5, RZ, 0xfc, !PT ; /* 0x0000000507057212 */ /* 0x000fc600078efcff */ /*0950*/ @!P0 BRA 0xb80 ; /* 0x0000022000008947 */ /* 0x000fea0003800000 */ /*0960*/ IMAD.MOV R3, RZ, RZ, -R17 ; /* 0x000000ffff037224 */ /* 0x000fe200078e0a11 */ /*0970*/ DMUL.RP R4, R14, R4 ; /* 0x000000040e047228 */ /* 0x000e220000008000 */ /*0980*/ IMAD.MOV.U32 R2, RZ, RZ, RZ ; /* 0x000000ffff027224 */ /* 0x000fcc00078e00ff */ /*0990*/ DFMA R2, R10, -R2, R14 ; /* 0x800000020a02722b */ /* 0x000e46000000000e */ /*09a0*/ LOP3.LUT R7, R5, R7, RZ, 0x3c, !PT ; /* 0x0000000705077212 */ /* 0x001fc600078e3cff */ /*09b0*/ IADD3 R2, -R17, -0x43300000, RZ ; /* 0xbcd0000011027810 */ /* 0x002fc80007ffe1ff */ /*09c0*/ FSETP.NEU.AND P0, PT, |R3|, R2, PT ; /* 0x000000020300720b */ /* 0x000fc80003f0d200 */ /*09d0*/ FSEL R10, R4, R10, !P0 ; /* 0x0000000a040a7208 */ /* 0x000fe40004000000 */ /*09e0*/ FSEL R11, R7, R11, !P0 ; /* 0x0000000b070b7208 */ /* 0x000fe20004000000 */ /*09f0*/ BRA 0xb80 ; /* 0x0000018000007947 */ /* 0x000fea0003800000 */ /*0a00*/ DSETP.NAN.AND P0, PT, R4, R4, PT ; /* 0x000000040400722a */ /* 0x003e1c0003f08000 */ /*0a10*/ @P0 BRA 0xb60 ; /* 0x0000014000000947 */ /* 0x001fea0003800000 */ /*0a20*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff027624 */ /* 0x000fe400078e00ff */ /*0a30*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff037624 */ /* 0x000fcc00078e00ff */ /*0a40*/ DSETP.NAN.AND P0, PT, R2, c[0x0][0x178], PT ; /* 0x00005e000200762a */ /* 0x000e1c0003f08000 */ /*0a50*/ @P0 BRA 0xb30 ; /* 0x000000d000000947 */ /* 0x001fea0003800000 */ /*0a60*/ ISETP.NE.AND P0, PT, R21, R22, PT ; /* 0x000000161500720c */ /* 0x000fe20003f05270 */ /*0a70*/ IMAD.MOV.U32 R10, RZ, RZ, 0x0 ; /* 0x00000000ff0a7424 */ /* 0x000fe400078e00ff */ /*0a80*/ IMAD.MOV.U32 R11, RZ, RZ, -0x80000 ; /* 0xfff80000ff0b7424 */ /* 0x000fd400078e00ff */ /*0a90*/ @!P0 BRA 0xb80 ; /* 0x000000e000008947 */ /* 0x000fea0003800000 */ /*0aa0*/ ISETP.NE.AND P0, PT, R21, 0x7ff00000, PT ; /* 0x7ff000001500780c */ /* 0x000fe40003f05270 */ /*0ab0*/ LOP3.LUT R4, R5, c[0x0][0x17c], RZ, 0x3c, !PT ; /* 0x00005f0005047a12 */ /* 0x000fe400078e3cff */ /*0ac0*/ ISETP.EQ.OR P0, PT, R22, RZ, !P0 ; /* 0x000000ff1600720c */ /* 0x000fe40004702670 */ /*0ad0*/ LOP3.LUT R11, R4, 0x80000000, RZ, 0xc0, !PT ; /* 0x80000000040b7812 */ /* 0x000fd600078ec0ff */ /*0ae0*/ @P0 LOP3.LUT R2, R11, 0x7ff00000, RZ, 0xfc, !PT ; /* 0x7ff000000b020812 */ /* 0x000fe200078efcff */ /*0af0*/ @!P0 IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a8224 */ /* 0x000fe400078e00ff */ /*0b00*/ @P0 IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a0224 */ /* 0x000fe400078e00ff */ /*0b10*/ @P0 IMAD.MOV.U32 R11, RZ, RZ, R2 ; /* 0x000000ffff0b0224 */ /* 0x000fe200078e0002 */ /*0b20*/ BRA 0xb80 ; /* 0x0000005000007947 */ /* 0x000fea0003800000 */ /*0b30*/ LOP3.LUT R11, R18, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000120b7812 */ /* 0x000fe200078efcff */ /*0b40*/ IMAD.MOV.U32 R10, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff0a7624 */ /* 0x000fe200078e00ff */ /*0b50*/ BRA 0xb80 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*0b60*/ LOP3.LUT R11, R5, 0x80000, RZ, 0xfc, !PT ; /* 0x00080000050b7812 */ /* 0x000fe400078efcff */ /*0b70*/ MOV R10, R4 ; /* 0x00000004000a7202 */ /* 0x000fc40000000f00 */ /*0b80*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0b90*/ IMAD.MOV.U32 R17, RZ, RZ, 0x0 ; /* 0x00000000ff117424 */ /* 0x000fe400078e00ff */ /*0ba0*/ IMAD.MOV.U32 R4, RZ, RZ, R10 ; /* 0x000000ffff047224 */ /* 0x000fc400078e000a */ /*0bb0*/ IMAD.MOV.U32 R5, RZ, RZ, R11 ; /* 0x000000ffff057224 */ /* 0x000fe200078e000b */ /*0bc0*/ RET.REL.NODEC R16 0x0 ; /* 0xfffff43010007950 */ /* 0x000fec0003c3ffff */ /*0bd0*/ ISETP.GE.U32.AND P0, PT, R4, -0x3400000, PT ; /* 0xfcc000000400780c */ /* 0x000fe20003f06070 */ /*0be0*/ BSSY B1, 0xe50 ; /* 0x0000026000017945 */ /* 0x000fe20003800000 */ /*0bf0*/ IMAD.MOV.U32 R10, RZ, RZ, R12 ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e000c */ /*0c00*/ IMAD.MOV.U32 R4, RZ, RZ, R16 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0010 */ /*0c10*/ IMAD.MOV.U32 R5, RZ, RZ, R17 ; /* 0x000000ffff057224 */ /* 0x000fce00078e0011 */ /*0c20*/ @!P0 BRA 0xcb0 ; /* 0x0000008000008947 */ /* 0x000fea0003800000 */ /*0c30*/ DFMA.RM R4, R4, R10, R14 ; /* 0x0000000a0404722b */ /* 0x000e14000000400e */ /*0c40*/ IADD3 R10, P0, R4, 0x1, RZ ; /* 0x00000001040a7810 */ /* 0x001fca0007f1e0ff */ /*0c50*/ IMAD.X R11, RZ, RZ, R5, P0 ; /* 0x000000ffff0b7224 */ /* 0x000fcc00000e0605 */ /*0c60*/ DFMA.RP R6, -R4, R10, R6 ; /* 0x0000000a0406722b */ /* 0x000e0c0000008106 */ /*0c70*/ DSETP.GT.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600722a */ /* 0x001e0c0003f04000 */ /*0c80*/ FSEL R4, R10, R4, P0 ; /* 0x000000040a047208 */ /* 0x001fe40000000000 */ /*0c90*/ FSEL R5, R11, R5, P0 ; /* 0x000000050b057208 */ /* 0x000fe20000000000 */ /*0ca0*/ BRA 0xe40 ; /* 0x0000019000007947 */ /* 0x000fea0003800000 */ /*0cb0*/ DSETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600722a */ /* 0x000e1c0003f05000 */ /*0cc0*/ @!P0 BRA 0xe30 ; /* 0x0000016000008947 */ /* 0x001fea0003800000 */ /*0cd0*/ ISETP.GE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fda0003f06270 */ /*0ce0*/ @!P0 IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff048424 */ /* 0x000fe400078e00ff */ /*0cf0*/ @!P0 IMAD.MOV.U32 R5, RZ, RZ, -0x80000 ; /* 0xfff80000ff058424 */ /* 0x000fe200078e00ff */ /*0d00*/ @!P0 BRA 0xe40 ; /* 0x0000013000008947 */ /* 0x000fea0003800000 */ /*0d10*/ ISETP.GT.AND P0, PT, R7, 0x7fefffff, PT ; /* 0x7fefffff0700780c */ /* 0x000fda0003f04270 */ /*0d20*/ @P0 BRA 0xe30 ; /* 0x0000010000000947 */ /* 0x000fea0003800000 */ /*0d30*/ DMUL R4, R6, 8.11296384146066816958e+31 ; /* 0x4690000006047828 */ /* 0x0000620000000000 */ /*0d40*/ IMAD.MOV.U32 R12, RZ, RZ, 0x0 ; /* 0x00000000ff0c7424 */ /* 0x000fe400078e00ff */ /*0d50*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x001fe400078e00ff */ /*0d60*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3fd80000 ; /* 0x3fd80000ff0d7424 */ /* 0x000fe200078e00ff */ /*0d70*/ MUFU.RSQ64H R7, R5 ; /* 0x0000000500077308 */ /* 0x002e260000001c00 */ /*0d80*/ DMUL R10, R6, R6 ; /* 0x00000006060a7228 */ /* 0x001e0c0000000000 */ /*0d90*/ DFMA R10, R4, -R10, 1 ; /* 0x3ff00000040a742b */ /* 0x001e0c000000080a */ /*0da0*/ DFMA R12, R10, R12, 0.5 ; /* 0x3fe000000a0c742b */ /* 0x001fc8000000000c */ /*0db0*/ DMUL R10, R6, R10 ; /* 0x0000000a060a7228 */ /* 0x000e0c0000000000 */ /*0dc0*/ DFMA R10, R12, R10, R6 ; /* 0x0000000a0c0a722b */ /* 0x001e0c0000000006 */ /*0dd0*/ DMUL R6, R4, R10 ; /* 0x0000000a04067228 */ /* 0x0010480000000000 */ /*0de0*/ IADD3 R11, R11, -0x100000, RZ ; /* 0xfff000000b0b7810 */ /* 0x001fe40007ffe0ff */ /*0df0*/ DFMA R12, R6, -R6, R4 ; /* 0x80000006060c722b */ /* 0x002e0c0000000004 */ /*0e00*/ DFMA R4, R10, R12, R6 ; /* 0x0000000c0a04722b */ /* 0x001e140000000006 */ /*0e10*/ IADD3 R5, R5, -0x3500000, RZ ; /* 0xfcb0000005057810 */ /* 0x001fe20007ffe0ff */ /*0e20*/ BRA 0xe40 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*0e30*/ DADD R4, R6, R6 ; /* 0x0000000006047229 */ /* 0x00004c0000000006 */ /*0e40*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0e50*/ IMAD.MOV.U32 R3, RZ, RZ, 0x0 ; /* 0x00000000ff037424 */ /* 0x000fc800078e00ff */ /*0e60*/ RET.REL.NODEC R2 0x0 ; /* 0xfffff19002007950 */ /* 0x000fea0003c3ffff */ /*0e70*/ BRA 0xe70; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0e80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0e90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ea0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0eb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ec0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ed0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ee0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ef0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10PDH_kernelP10hist_entryP8atomdescxd .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .p2align 8 .type _Z10PDH_kernelP10hist_entryP8atomdescxd,@function _Z10PDH_kernelP10hist_entryP8atomdescxd: s_clause 0x1 s_load_b32 s4, s[0:1], 0x2c s_load_b64 s[2:3], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) s_mul_i32 s15, s15, s4 s_mov_b32 s4, exec_lo v_add_nc_u32_e32 v2, s15, v0 v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i64_e64 s[2:3], v[2:3] s_cbranch_execz .LBB0_4 v_add_nc_u32_e32 v4, 1, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v5, 31, v4 v_cmp_gt_i64_e32 vcc_lo, s[2:3], v[4:5] s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_4 s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x18 v_add3_u32 v3, v0, s15, 2 s_mov_b32 s8, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_ashrrev_i32_e32 v6, 31, v3 s_waitcnt lgkmcnt(0) v_mad_i64_i32 v[0:1], null, v2, 24, s[6:7] v_add_co_u32 v2, vcc_lo, v3, -1 v_add_co_ci_u32_e32 v3, vcc_lo, -1, v6, vcc_lo .LBB0_3: v_mad_u64_u32 v[8:9], null, v4, 24, s[6:7] s_delay_alu instid0(VALU_DEP_4) v_add_co_u32 v10, vcc_lo, v0, 16 v_add_co_ci_u32_e32 v11, vcc_lo, 0, v1, vcc_lo s_clause 0x1 global_load_b64 v[12:13], v[0:1], off global_load_b64 v[14:15], v[10:11], off offset:-8 v_mov_b32_e32 v4, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[6:7], null, v5, 24, v[4:5] v_mov_b32_e32 v9, v6 s_clause 0x2 global_load_b128 v[4:7], v[8:9], off global_load_b64 v[10:11], v[10:11], off global_load_b64 v[8:9], v[8:9], off offset:16 s_waitcnt vmcnt(2) v_add_f64 v[6:7], v[14:15], -v[6:7] v_add_f64 v[4:5], v[12:13], -v[4:5] s_waitcnt vmcnt(0) v_add_f64 v[8:9], v[10:11], -v[8:9] s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[6:7], v[6:7], v[6:7] v_fma_f64 v[4:5], v[4:5], v[4:5], v[6:7] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[4:5], v[8:9], v[8:9], v[4:5] v_cmp_gt_f64_e32 vcc_lo, 0x10000000, v[4:5] v_cndmask_b32_e64 v6, 0, 1, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b32_e32 v6, 8, v6 v_ldexp_f64 v[4:5], v[4:5], v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_rsq_f64_e32 v[6:7], v[4:5] s_waitcnt_depctr 0xfff v_mul_f64 v[8:9], v[4:5], v[6:7] v_mul_f64 v[6:7], v[6:7], 0.5 v_fma_f64 v[10:11], -v[6:7], v[8:9], 0.5 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] v_fma_f64 v[6:7], v[6:7], v[10:11], v[6:7] v_fma_f64 v[10:11], -v[8:9], v[8:9], v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[8:9], v[10:11], v[6:7], v[8:9] v_fma_f64 v[10:11], -v[8:9], v[8:9], v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_fma_f64 v[6:7], v[10:11], v[6:7], v[8:9] v_cndmask_b32_e64 v8, 0, 0xffffff80, vcc_lo v_cmp_class_f64_e64 vcc_lo, v[4:5], 0x260 v_ldexp_f64 v[6:7], v[6:7], v8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v5, v7, v5 :: v_dual_cndmask_b32 v4, v6, v4 v_div_scale_f64 v[6:7], null, s[0:1], s[0:1], v[4:5] v_div_scale_f64 v[12:13], vcc_lo, v[4:5], s[0:1], v[4:5] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f64_e32 v[8:9], v[6:7] s_waitcnt_depctr 0xfff v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[10:11], v[12:13], v[8:9] v_fma_f64 v[6:7], -v[6:7], v[10:11], v[12:13] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fmas_f64 v[6:7], v[6:7], v[8:9], v[10:11] v_div_fixup_f64 v[4:5], v[6:7], s[0:1], v[4:5] v_mov_b32_e32 v6, 1 v_mov_b32_e32 v7, 0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_i32_f64_e32 v4, v[4:5] v_ashrrev_i32_e32 v5, 31, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 3, v[4:5] v_add_co_u32 v4, vcc_lo, s4, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo v_add_co_u32 v2, vcc_lo, v2, 1 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo global_atomic_add_u64 v[4:5], v[6:7], off v_mov_b32_e32 v4, v2 v_ashrrev_i32_e32 v5, 31, v2 v_cmp_le_i64_e32 vcc_lo, s[2:3], v[2:3] s_or_b32 s8, vcc_lo, s8 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s8 s_cbranch_execnz .LBB0_3 .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z10PDH_kernelP10hist_entryP8atomdescxd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10PDH_kernelP10hist_entryP8atomdescxd, .Lfunc_end0-_Z10PDH_kernelP10hist_entryP8atomdescxd .section .AMDGPU.csdata,"",@progbits .text .protected _Z11PDH_kernel2P10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .p2align 8 .type _Z11PDH_kernel2P10hist_entryP8atomdescxd,@function _Z11PDH_kernel2P10hist_entryP8atomdescxd: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11PDH_kernel2P10hist_entryP8atomdescxd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end1: .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, .Lfunc_end1-_Z11PDH_kernel2P10hist_entryP8atomdescxd .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value - .offset: 24 .size: 8 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: hidden_block_count_x - .offset: 36 .size: 4 .value_kind: hidden_block_count_y - .offset: 40 .size: 4 .value_kind: hidden_block_count_z - .offset: 44 .size: 2 .value_kind: hidden_group_size_x - .offset: 46 .size: 2 .value_kind: hidden_group_size_y - .offset: 48 .size: 2 .value_kind: hidden_group_size_z - .offset: 50 .size: 2 .value_kind: hidden_remainder_x - .offset: 52 .size: 2 .value_kind: hidden_remainder_y - .offset: 54 .size: 2 .value_kind: hidden_remainder_z - .offset: 72 .size: 8 .value_kind: hidden_global_offset_x - .offset: 80 .size: 8 .value_kind: hidden_global_offset_y - .offset: 88 .size: 8 .value_kind: hidden_global_offset_z - .offset: 96 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z10PDH_kernelP10hist_entryP8atomdescxd .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10PDH_kernelP10hist_entryP8atomdescxd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value - .offset: 24 .size: 8 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11PDH_kernel2P10hist_entryP8atomdescxd .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z11PDH_kernel2P10hist_entryP8atomdescxd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0015a26d_00000000-6_proj1-tanmay.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2065: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2065: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z12p2p_distanceii .type _Z12p2p_distanceii, @function _Z12p2p_distanceii: .LFB2057: .cfi_startproc endbr64 movq atom_list(%rip), %rax movslq %edi, %rdi leaq (%rdi,%rdi,2), %rdx leaq (%rax,%rdx,8), %rdx movslq %esi, %rsi leaq (%rsi,%rsi,2), %rcx leaq (%rax,%rcx,8), %rax movsd (%rdx), %xmm1 subsd (%rax), %xmm1 movsd 8(%rdx), %xmm2 subsd 8(%rax), %xmm2 movsd 16(%rdx), %xmm0 subsd 16(%rax), %xmm0 mulsd %xmm1, %xmm1 mulsd %xmm2, %xmm2 addsd %xmm2, %xmm1 mulsd %xmm0, %xmm0 addsd %xmm1, %xmm0 sqrtsd %xmm0, %xmm0 ret .cfi_endproc .LFE2057: .size _Z12p2p_distanceii, .-_Z12p2p_distanceii .globl _Z12PDH_baselinev .type _Z12PDH_baselinev, @function _Z12PDH_baselinev: .LFB2058: .cfi_startproc endbr64 cmpq $0, PDH_acnt(%rip) jle .L11 pushq %r14 .cfi_def_cfa_offset 16 .cfi_offset 14, -16 pushq %r13 .cfi_def_cfa_offset 24 .cfi_offset 13, -24 pushq %r12 .cfi_def_cfa_offset 32 .cfi_offset 12, -32 pushq %rbp .cfi_def_cfa_offset 40 .cfi_offset 6, -40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset 3, -48 movl $1, %r13d movl $0, %r14d .L7: movl %r14d, %r12d movl %r13d, %ebx cmpq %r13, PDH_acnt(%rip) jle .L5 movq %r13, %rbp .L6: movl %ebx, %esi movl %r12d, %edi call _Z12p2p_distanceii divsd PDH_res(%rip), %xmm0 cvttsd2sil %xmm0, %eax cltq movq histogram(%rip), %rdx addq $1, (%rdx,%rax,8) addl $1, %ebx movq PDH_acnt(%rip), %rax addq $1, %rbp cmpq %rbp, %rax jg .L6 addq $1, %r14 addq $1, %r13 cmpq %r14, %rax jg .L7 .L5: movl $0, %eax popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .L11: .cfi_restore 3 .cfi_restore 6 .cfi_restore 12 .cfi_restore 13 .cfi_restore 14 movl $0, %eax ret .cfi_endproc .LFE2058: .size _Z12PDH_baselinev, .-_Z12PDH_baselinev .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Running time for CPU version: %ld.%06ld\n" .text .globl _Z19report_running_timev .type _Z19report_running_timev, @function _Z19report_running_timev: .LFB2059: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $8, %rsp .cfi_def_cfa_offset 32 leaq Idunno(%rip), %rsi leaq endTime(%rip), %rdi call gettimeofday@PLT movq endTime(%rip), %rbp subq startTime(%rip), %rbp movq 8+endTime(%rip), %rbx subq 8+startTime(%rip), %rbx js .L17 .L15: movq %rbx, %rcx movq %rbp, %rdx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq %rbp, %xmm1 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state subq $1, %rbp addq $1000000, %rbx jmp .L15 .cfi_endproc .LFE2059: .size _Z19report_running_timev, .-_Z19report_running_timev .section .rodata.str1.8 .align 8 .LC2: .string "Running time for GPU version: %ld.%06ld\n" .text .globl _Z23report_running_time_GPUv .type _Z23report_running_time_GPUv, @function _Z23report_running_time_GPUv: .LFB2060: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $8, %rsp .cfi_def_cfa_offset 32 leaq Idunno(%rip), %rsi leaq endTime(%rip), %rdi call gettimeofday@PLT movq endTime(%rip), %rbp subq startTime(%rip), %rbp movq 8+endTime(%rip), %rbx subq 8+startTime(%rip), %rbx js .L21 .L19: movq %rbx, %rcx movq %rbp, %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq %rbp, %xmm1 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L21: .cfi_restore_state subq $1, %rbp addq $1000000, %rbx jmp .L19 .cfi_endproc .LFE2060: .size _Z23report_running_time_GPUv, .-_Z23report_running_time_GPUv .section .rodata.str1.1,"aMS",@progbits,1 .LC3: .string "\n%02d: " .LC4: .string "%15lld " .LC5: .string "\n T:%lld \n" .LC6: .string "| " .text .globl _Z16output_histogramP10hist_entry .type _Z16output_histogramP10hist_entry, @function _Z16output_histogramP10hist_entry: .LFB2061: .cfi_startproc endbr64 cmpl $0, num_buckets(%rip) jle .L30 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $8, %rsp .cfi_def_cfa_offset 64 movq %rdi, %rbp movl $0, %ebx movl $0, %r12d leaq .LC4(%rip), %r13 leaq .LC6(%rip), %r14 leaq .LC5(%rip), %r15 jmp .L27 .L34: movl %ebx, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L24 .L35: movq %r15, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L26: addq $1, %rbx cmpl %ebx, num_buckets(%rip) jle .L33 .L27: movslq %ebx, %rax imulq $1717986919, %rax, %rax sarq $33, %rax movl %ebx, %edx sarl $31, %edx subl %edx, %eax leal (%rax,%rax,4), %eax cmpl %ebx, %eax je .L34 .L24: movq 0(%rbp,%rbx,8), %rdx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rdx addq 0(%rbp,%rbx,8), %rdx movq %rdx, %r12 movl num_buckets(%rip), %eax subl $1, %eax cmpl %ebx, %eax je .L35 movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L26 .L33: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L30: .cfi_restore 3 .cfi_restore 6 .cfi_restore 12 .cfi_restore 13 .cfi_restore 14 .cfi_restore 15 ret .cfi_endproc .LFE2061: .size _Z16output_histogramP10hist_entry, .-_Z16output_histogramP10hist_entry .globl _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .type _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, @function _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movsd %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L40 .L36: movq 136(%rsp), %rax subq %fs:40, %rax jne .L41 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L40: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z10PDH_kernelP10hist_entryP8atomdescxd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L36 .L41: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, .-_Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .type _Z10PDH_kernelP10hist_entryP8atomdescxd, @function _Z10PDH_kernelP10hist_entryP8atomdescxd: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z10PDH_kernelP10hist_entryP8atomdescxd, .-_Z10PDH_kernelP10hist_entryP8atomdescxd .section .rodata.str1.1 .LC10: .string "Starting CPU...\n" .LC11: .string "Starting GPU...\n" .LC16: .string "Difference: \n" .text .globl main .type main, @function main: .LFB2062: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $40, %rsp .cfi_def_cfa_offset 64 movq %rsi, %rbx movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT cltq movq %rax, PDH_acnt(%rip) movq 16(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, PDH_res(%rip) movsd .LC7(%rip), %xmm1 divsd %xmm0, %xmm1 cvttsd2sil %xmm1, %edi addl $1, %edi movl %edi, num_buckets(%rip) movslq %edi, %rdi salq $3, %rdi call malloc@PLT movq %rax, histogram(%rip) movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rdi salq $3, %rdi call malloc@PLT movq %rax, atom_list(%rip) movl $1, %edi call srand@PLT cmpq $0, PDH_acnt(%rip) jle .L45 movl $0, %ebx movl $0, %ebp .L46: call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, (%rax,%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 8(%rax,%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2sdl %eax, %xmm0 divsd .LC8(%rip), %xmm0 mulsd .LC9(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 16(%rax,%rbx) addq $1, %rbp addq $24, %rbx cmpq %rbp, PDH_acnt(%rip) jg .L46 .L45: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq Idunno(%rip), %rbp movq %rbp, %rsi leaq startTime(%rip), %rbx movq %rbx, %rdi call gettimeofday@PLT call _Z12PDH_baselinev call _Z19report_running_timev movq histogram(%rip), %rdi call _Z16output_histogramP10hist_entry leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movslq num_buckets(%rip), %rdi salq $3, %rdi call malloc@PLT movq %rax, h_gpu_histogram(%rip) movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rsi salq $3, %rsi leaq d_atom_list(%rip), %rdi call cudaMalloc@PLT movq PDH_acnt(%rip), %rax leaq (%rax,%rax,2), %rdx salq $3, %rdx movl $1, %ecx movq atom_list(%rip), %rsi movq d_atom_list(%rip), %rdi call cudaMemcpy@PLT movslq num_buckets(%rip), %rsi salq $3, %rsi leaq d_gpu_histogram(%rip), %rdi call cudaMalloc@PLT movslq num_buckets(%rip), %rdx salq $3, %rdx movl $1, %ecx movq h_gpu_histogram(%rip), %rsi movq d_gpu_histogram(%rip), %rdi call cudaMemcpy@PLT movq %rbp, %rsi movq %rbx, %rdi call gettimeofday@PLT movl $256, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) pxor %xmm0, %xmm0 cvtsi2sdq PDH_acnt(%rip), %xmm0 mulsd .LC12(%rip), %xmm0 movapd %xmm0, %xmm3 movsd .LC17(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC13(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L47 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC15(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L47: cvttsd2siq %xmm3, %rax movl %eax, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L54 .L48: movslq num_buckets(%rip), %rdx salq $3, %rdx movl $2, %ecx movq d_gpu_histogram(%rip), %rsi movq h_gpu_histogram(%rip), %rdi call cudaMemcpy@PLT call _Z23report_running_time_GPUv movq h_gpu_histogram(%rip), %rdi call _Z16output_histogramP10hist_entry leaq .LC16(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl num_buckets(%rip), %ebp movslq %ebp, %rbx salq $3, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, diff_histogram(%rip) testl %ebp, %ebp jle .L49 movl $0, %eax .L50: movq histogram(%rip), %rdx movq (%rdx,%rax), %rdx movq h_gpu_histogram(%rip), %rcx subq (%rcx,%rax), %rdx movq diff_histogram(%rip), %rcx movq %rdx, (%rcx,%rax) addq $8, %rax cmpq %rax, %rbx jne .L50 .L49: movq diff_histogram(%rip), %rdi call _Z16output_histogramP10hist_entry movq d_gpu_histogram(%rip), %rdi call cudaFree@PLT movq d_atom_list(%rip), %rdi call cudaFree@PLT movq histogram(%rip), %rdi call free@PLT movq atom_list(%rip), %rdi call free@PLT movq h_gpu_histogram(%rip), %rdi call free@PLT movq diff_histogram(%rip), %rdi call free@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L54: .cfi_restore_state movsd PDH_res(%rip), %xmm0 movq PDH_acnt(%rip), %rdx movq d_atom_list(%rip), %rsi movq d_gpu_histogram(%rip), %rdi call _Z53__device_stub__Z10PDH_kernelP10hist_entryP8atomdescxdP10hist_entryP8atomdescxd jmp .L48 .cfi_endproc .LFE2062: .size main, .-main .globl _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .type _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, @function _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd: .LFB2089: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movsd %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L59 .L55: movq 136(%rsp), %rax subq %fs:40, %rax jne .L60 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L59: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11PDH_kernel2P10hist_entryP8atomdescxd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L55 .L60: call __stack_chk_fail@PLT .cfi_endproc .LFE2089: .size _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd, .-_Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .type _Z11PDH_kernel2P10hist_entryP8atomdescxd, @function _Z11PDH_kernel2P10hist_entryP8atomdescxd: .LFB2090: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z54__device_stub__Z11PDH_kernel2P10hist_entryP8atomdescxdP10hist_entryP8atomdescxd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2090: .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, .-_Z11PDH_kernel2P10hist_entryP8atomdescxd .section .rodata.str1.8 .align 8 .LC18: .string "_Z11PDH_kernel2P10hist_entryP8atomdescxd" .align 8 .LC19: .string "_Z10PDH_kernelP10hist_entryP8atomdescxd" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2092: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC18(%rip), %rdx movq %rdx, %rcx leaq _Z11PDH_kernel2P10hist_entryP8atomdescxd(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC19(%rip), %rdx movq %rdx, %rcx leaq _Z10PDH_kernelP10hist_entryP8atomdescxd(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2092: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .globl endTime .bss .align 16 .type endTime, @object .size endTime, 16 endTime: .zero 16 .globl startTime .align 16 .type startTime, @object .size startTime, 16 startTime: .zero 16 .globl Idunno .align 8 .type Idunno, @object .size Idunno, 8 Idunno: .zero 8 .globl d_atom_list .align 8 .type d_atom_list, @object .size d_atom_list, 8 d_atom_list: .zero 8 .globl atom_list .align 8 .type atom_list, @object .size atom_list, 8 atom_list: .zero 8 .globl PDH_res .align 8 .type PDH_res, @object .size PDH_res, 8 PDH_res: .zero 8 .globl num_buckets .align 4 .type num_buckets, @object .size num_buckets, 4 num_buckets: .zero 4 .globl PDH_acnt .align 8 .type PDH_acnt, @object .size PDH_acnt, 8 PDH_acnt: .zero 8 .globl diff_histogram .align 8 .type diff_histogram, @object .size diff_histogram, 8 diff_histogram: .zero 8 .globl d_gpu_histogram .align 8 .type d_gpu_histogram, @object .size d_gpu_histogram, 8 d_gpu_histogram: .zero 8 .globl h_gpu_histogram .align 8 .type h_gpu_histogram, @object .size h_gpu_histogram, 8 h_gpu_histogram: .zero 8 .globl histogram .align 8 .type histogram, @object .size histogram, 8 histogram: .zero 8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1093567616 .align 8 .LC7: .long 0 .long 1088648064 .align 8 .LC8: .long -4194304 .long 1105199103 .align 8 .LC9: .long 0 .long 1087796736 .align 8 .LC12: .long 0 .long 1064304640 .align 8 .LC13: .long 0 .long 1127219200 .align 8 .LC15: .long 0 .long 1072693248 .align 8 .LC17: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "proj1-tanmay.hip" .globl _Z12p2p_distanceii # -- Begin function _Z12p2p_distanceii .p2align 4, 0x90 .type _Z12p2p_distanceii,@function _Z12p2p_distanceii: # @_Z12p2p_distanceii .cfi_startproc # %bb.0: movq atom_list(%rip), %rax movslq %edi, %rcx leaq (%rcx,%rcx,2), %rcx movslq %esi, %rdx leaq (%rdx,%rdx,2), %rdx movsd (%rax,%rcx,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%rcx,8), %xmm2 # xmm2 = mem[0],zero movsd 16(%rax,%rcx,8), %xmm0 # xmm0 = mem[0],zero subsd (%rax,%rdx,8), %xmm1 mulsd %xmm1, %xmm1 subsd 8(%rax,%rdx,8), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd 16(%rax,%rdx,8), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 xorpd %xmm1, %xmm1 ucomisd %xmm1, %xmm0 jb sqrt # TAILCALL # %bb.1: # %.split sqrtsd %xmm0, %xmm0 retq .Lfunc_end0: .size _Z12p2p_distanceii, .Lfunc_end0-_Z12p2p_distanceii .cfi_endproc # -- End function .globl _Z12PDH_baselinev # -- Begin function _Z12PDH_baselinev .p2align 4, 0x90 .type _Z12PDH_baselinev,@function _Z12PDH_baselinev: # @_Z12PDH_baselinev .cfi_startproc # %bb.0: cmpq $0, PDH_acnt(%rip) jle .LBB1_10 # %bb.1: # %.lr.ph11.preheader pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 pushq %rax .cfi_def_cfa_offset 64 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $1, %ebx movl $40, %r14d xorl %r15d, %r15d xorpd %xmm3, %xmm3 jmp .LBB1_3 .p2align 4, 0x90 .LBB1_2: # %.loopexit # in Loop: Header=BB1_3 Depth=1 incq %rbx addq $24, %r14 cmpq %r15, PDH_acnt(%rip) jle .LBB1_9 .LBB1_3: # %.lr.ph11 # =>This Loop Header: Depth=1 # Child Loop BB1_5 Depth 2 movq %r15, %rax incq %r15 cmpq %r15, PDH_acnt(%rip) jle .LBB1_2 # %bb.4: # %.lr.ph.preheader # in Loop: Header=BB1_3 Depth=1 leaq (%rax,%rax,2), %r12 movq %r14, %r13 movq %rbx, %rbp jmp .LBB1_5 .p2align 4, 0x90 .LBB1_7: # %call.sqrt # in Loop: Header=BB1_5 Depth=2 callq sqrt xorpd %xmm3, %xmm3 .LBB1_8: # %.lr.ph.split # in Loop: Header=BB1_5 Depth=2 divsd PDH_res(%rip), %xmm0 cvttsd2si %xmm0, %eax movq histogram(%rip), %rcx cltq incq (%rcx,%rax,8) incq %rbp addq $24, %r13 cmpq %rbp, PDH_acnt(%rip) jle .LBB1_2 .LBB1_5: # %.lr.ph # Parent Loop BB1_3 Depth=1 # => This Inner Loop Header: Depth=2 movq atom_list(%rip), %rax movsd (%rax,%r12,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%r12,8), %xmm2 # xmm2 = mem[0],zero subsd -16(%rax,%r13), %xmm1 movsd 16(%rax,%r12,8), %xmm0 # xmm0 = mem[0],zero mulsd %xmm1, %xmm1 subsd -8(%rax,%r13), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd (%rax,%r13), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 ucomisd %xmm3, %xmm0 jb .LBB1_7 # %bb.6: # in Loop: Header=BB1_5 Depth=2 sqrtsd %xmm0, %xmm0 jmp .LBB1_8 .LBB1_9: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 .cfi_restore %rbx .cfi_restore %r12 .cfi_restore %r13 .cfi_restore %r14 .cfi_restore %r15 .cfi_restore %rbp .LBB1_10: # %._crit_edge xorl %eax, %eax retq .Lfunc_end1: .size _Z12PDH_baselinev, .Lfunc_end1-_Z12PDH_baselinev .cfi_endproc # -- End function .globl _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd # -- Begin function _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .p2align 4, 0x90 .type _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd,@function _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd: # @_Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd, .Lfunc_end2-_Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .cfi_endproc # -- End function .globl _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd # -- Begin function _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .p2align 4, 0x90 .type _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd,@function _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd: # @_Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z11PDH_kernel2P10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end3: .size _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd, .Lfunc_end3-_Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z19report_running_timev .LCPI4_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl _Z19report_running_timev .p2align 4, 0x90 .type _Z19report_running_timev,@function _Z19report_running_timev: # @_Z19report_running_timev .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rbx movq %rcx, %r14 sarq $63, %r14 addq %rax, %r14 testq %rcx, %rcx cmovnsq %rcx, %rbx movl $.L.str, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf cvtsi2sd %r14, %xmm1 cvtsi2sd %rbx, %xmm0 divsd .LCPI4_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size _Z19report_running_timev, .Lfunc_end4-_Z19report_running_timev .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z23report_running_time_GPUv .LCPI5_0: .quad 0x412e848000000000 # double 1.0E+6 .text .globl _Z23report_running_time_GPUv .p2align 4, 0x90 .type _Z23report_running_time_GPUv,@function _Z23report_running_time_GPUv: # @_Z23report_running_time_GPUv .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rbx movq %rcx, %r14 sarq $63, %r14 addq %rax, %r14 testq %rcx, %rcx cmovnsq %rcx, %rbx movl $.L.str.1, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf cvtsi2sd %r14, %xmm1 cvtsi2sd %rbx, %xmm0 divsd .LCPI5_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end5: .size _Z23report_running_time_GPUv, .Lfunc_end5-_Z23report_running_time_GPUv .cfi_endproc # -- End function .globl _Z16output_histogramP10hist_entry # -- Begin function _Z16output_histogramP10hist_entry .p2align 4, 0x90 .type _Z16output_histogramP10hist_entry,@function _Z16output_histogramP10hist_entry: # @_Z16output_histogramP10hist_entry .cfi_startproc # %bb.0: cmpl $0, num_buckets(%rip) jle .LBB6_9 # %bb.1: # %.lr.ph.preheader pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 pushq %rax .cfi_def_cfa_offset 48 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdi, %rbx movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %r14d, %r14d xorl %r15d, %r15d jmp .LBB6_2 .p2align 4, 0x90 .LBB6_5: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.4, %edi movq %r15, %rsi xorl %eax, %eax callq printf .LBB6_7: # in Loop: Header=BB6_2 Depth=1 incq %r14 movslq num_buckets(%rip), %rax cmpq %rax, %r14 jge .LBB6_8 .LBB6_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl %r14d, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %r14d jne .LBB6_4 # %bb.3: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.2, %edi movl %r14d, %esi xorl %eax, %eax callq printf .LBB6_4: # in Loop: Header=BB6_2 Depth=1 movq (%rbx,%r14,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%rbx,%r14,8), %r15 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %r14 je .LBB6_5 # %bb.6: # in Loop: Header=BB6_2 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf jmp .LBB6_7 .LBB6_8: addq $8, %rsp .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 .cfi_restore %rbx .cfi_restore %r12 .cfi_restore %r14 .cfi_restore %r15 .LBB6_9: # %._crit_edge retq .Lfunc_end6: .size _Z16output_histogramP10hist_entry, .Lfunc_end6-_Z16output_histogramP10hist_entry .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI7_0: .quad 0x40e3738000000000 # double 39836 .LCPI7_1: .quad 0x41dfffffffc00000 # double 2147483647 .LCPI7_2: .quad 0x40d6760000000000 # double 23000 .LCPI7_3: .quad 0x3f70000000000000 # double 0.00390625 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $120, %rsp .cfi_def_cfa_offset 176 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movq %rsi, %rbx movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol cltq movq %rax, PDH_acnt(%rip) movq 16(%rbx), %rdi xorl %esi, %esi callq strtod movsd %xmm0, PDH_res(%rip) movsd .LCPI7_0(%rip), %xmm1 # xmm1 = mem[0],zero divsd %xmm0, %xmm1 cvttsd2si %xmm1, %eax leal 1(%rax), %ecx movl %ecx, num_buckets(%rip) cltq leaq 8(,%rax,8), %rdi callq malloc movq %rax, histogram(%rip) movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rdi callq malloc movq %rax, atom_list(%rip) movl $1, %edi callq srand cmpq $0, PDH_acnt(%rip) jle .LBB7_3 # %bb.1: # %.lr.ph.preheader xorl %ebx, %ebx xorl %r14d, %r14d .p2align 4, 0x90 .LBB7_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 movsd .LCPI7_1(%rip), %xmm1 # xmm1 = mem[0],zero divsd %xmm1, %xmm0 movsd .LCPI7_2(%rip), %xmm1 # xmm1 = mem[0],zero mulsd %xmm1, %xmm0 movq atom_list(%rip), %rax movsd %xmm0, (%rax,%rbx) callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 divsd .LCPI7_1(%rip), %xmm0 mulsd .LCPI7_2(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 8(%rax,%rbx) callq rand xorps %xmm0, %xmm0 cvtsi2sd %eax, %xmm0 divsd .LCPI7_1(%rip), %xmm0 mulsd .LCPI7_2(%rip), %xmm0 movq atom_list(%rip), %rax movsd %xmm0, 16(%rax,%rbx) incq %r14 addq $24, %rbx cmpq %r14, PDH_acnt(%rip) jg .LBB7_2 .LBB7_3: # %._crit_edge movl $.Lstr, %edi callq puts@PLT movl $startTime, %edi movl $Idunno, %esi callq gettimeofday cmpq $0, PDH_acnt(%rip) jle .LBB7_12 # %bb.4: # %.lr.ph11.i.preheader movl $1, %ebx movl $40, %r14d xorl %r15d, %r15d xorpd %xmm3, %xmm3 jmp .LBB7_6 .p2align 4, 0x90 .LBB7_5: # %.loopexit.i # in Loop: Header=BB7_6 Depth=1 incq %rbx addq $24, %r14 cmpq %r15, PDH_acnt(%rip) jle .LBB7_12 .LBB7_6: # %.lr.ph11.i # =>This Loop Header: Depth=1 # Child Loop BB7_8 Depth 2 movq %r15, %rax incq %r15 cmpq %r15, PDH_acnt(%rip) jle .LBB7_5 # %bb.7: # %.lr.ph.i.preheader # in Loop: Header=BB7_6 Depth=1 leaq (%rax,%rax,2), %r12 movq %r14, %r13 movq %rbx, %rbp jmp .LBB7_8 .p2align 4, 0x90 .LBB7_10: # %call.sqrt # in Loop: Header=BB7_8 Depth=2 callq sqrt xorpd %xmm3, %xmm3 .LBB7_11: # %.lr.ph.i.split # in Loop: Header=BB7_8 Depth=2 divsd PDH_res(%rip), %xmm0 cvttsd2si %xmm0, %eax movq histogram(%rip), %rcx cltq incq (%rcx,%rax,8) incq %rbp addq $24, %r13 cmpq %rbp, PDH_acnt(%rip) jle .LBB7_5 .LBB7_8: # %.lr.ph.i # Parent Loop BB7_6 Depth=1 # => This Inner Loop Header: Depth=2 movq atom_list(%rip), %rax movsd (%rax,%r12,8), %xmm1 # xmm1 = mem[0],zero movsd 8(%rax,%r12,8), %xmm2 # xmm2 = mem[0],zero subsd -16(%rax,%r13), %xmm1 movsd 16(%rax,%r12,8), %xmm0 # xmm0 = mem[0],zero mulsd %xmm1, %xmm1 subsd -8(%rax,%r13), %xmm2 mulsd %xmm2, %xmm2 addsd %xmm1, %xmm2 subsd (%rax,%r13), %xmm0 mulsd %xmm0, %xmm0 addsd %xmm2, %xmm0 ucomisd %xmm3, %xmm0 jb .LBB7_10 # %bb.9: # in Loop: Header=BB7_8 Depth=2 sqrtsd %xmm0, %xmm0 jmp .LBB7_11 .LBB7_12: # %_Z12PDH_baselinev.exit movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rdx movq %rcx, %rsi sarq $63, %rsi addq %rax, %rsi testq %rcx, %rcx cmovnsq %rcx, %rdx movl $.L.str, %edi xorl %eax, %eax callq printf cmpl $0, num_buckets(%rip) jle .LBB7_20 # %bb.13: # %.lr.ph.i17.preheader movq histogram(%rip), %r15 movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %ebx, %ebx xorl %r14d, %r14d jmp .LBB7_14 .p2align 4, 0x90 .LBB7_18: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_19: # in Loop: Header=BB7_14 Depth=1 incq %rbx movslq num_buckets(%rip), %rax cmpq %rax, %rbx jge .LBB7_20 .LBB7_14: # %.lr.ph.i17 # =>This Inner Loop Header: Depth=1 movl %ebx, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %ebx jne .LBB7_16 # %bb.15: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.2, %edi movl %ebx, %esi xorl %eax, %eax callq printf .LBB7_16: # in Loop: Header=BB7_14 Depth=1 movq (%r15,%rbx,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%r15,%rbx,8), %r14 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %rbx jne .LBB7_18 # %bb.17: # in Loop: Header=BB7_14 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq printf jmp .LBB7_19 .LBB7_20: # %_Z16output_histogramP10hist_entry.exit movl $.Lstr.1, %edi callq puts@PLT movslq num_buckets(%rip), %rdi shlq $3, %rdi callq malloc movq %rax, h_gpu_histogram(%rip) movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rsi movl $d_atom_list, %edi callq hipMalloc movq d_atom_list(%rip), %rdi movq atom_list(%rip), %rsi movq PDH_acnt(%rip), %rax shlq $3, %rax leaq (%rax,%rax,2), %rdx movl $1, %ecx callq hipMemcpy movslq num_buckets(%rip), %rsi shlq $3, %rsi movl $d_gpu_histogram, %edi callq hipMalloc movq d_gpu_histogram(%rip), %rdi movq h_gpu_histogram(%rip), %rsi movslq num_buckets(%rip), %rdx shlq $3, %rdx movl $1, %ecx callq hipMemcpy movl $startTime, %edi movl $Idunno, %esi callq gettimeofday xorps %xmm0, %xmm0 cvtsi2sdq PDH_acnt(%rip), %xmm0 mulsd .LCPI7_3(%rip), %xmm0 callq ceil@PLT cvttsd2si %xmm0, %rax movl %eax, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $256, %rdx # imm = 0x100 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB7_22 # %bb.21: movq d_gpu_histogram(%rip), %rax movq d_atom_list(%rip), %rcx movq PDH_acnt(%rip), %rdx movsd PDH_res(%rip), %xmm0 # xmm0 = mem[0],zero movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq %rdx, 56(%rsp) movsd %xmm0, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB7_22: movq h_gpu_histogram(%rip), %rdi movq d_gpu_histogram(%rip), %rsi movslq num_buckets(%rip), %rdx shlq $3, %rdx movl $2, %ecx callq hipMemcpy movl $endTime, %edi movl $Idunno, %esi callq gettimeofday movq endTime(%rip), %rax subq startTime(%rip), %rax movq endTime+8(%rip), %rcx subq startTime+8(%rip), %rcx leaq 1000000(%rcx), %rdx movq %rcx, %rsi sarq $63, %rsi addq %rax, %rsi testq %rcx, %rcx cmovnsq %rcx, %rdx movl $.L.str.1, %edi xorl %eax, %eax callq printf cmpl $0, num_buckets(%rip) jle .LBB7_30 # %bb.23: # %.lr.ph.i23.preheader movq h_gpu_histogram(%rip), %r15 movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %ebx, %ebx xorl %r14d, %r14d jmp .LBB7_24 .p2align 4, 0x90 .LBB7_28: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_29: # in Loop: Header=BB7_24 Depth=1 incq %rbx movslq num_buckets(%rip), %rax cmpq %rax, %rbx jge .LBB7_30 .LBB7_24: # %.lr.ph.i23 # =>This Inner Loop Header: Depth=1 movl %ebx, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %ebx jne .LBB7_26 # %bb.25: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.2, %edi movl %ebx, %esi xorl %eax, %eax callq printf .LBB7_26: # in Loop: Header=BB7_24 Depth=1 movq (%r15,%rbx,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%r15,%rbx,8), %r14 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %rbx jne .LBB7_28 # %bb.27: # in Loop: Header=BB7_24 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq printf jmp .LBB7_29 .LBB7_30: # %_Z16output_histogramP10hist_entry.exit27 movl $.Lstr.2, %edi callq puts@PLT movslq num_buckets(%rip), %r14 leaq (,%r14,8), %rdi callq malloc movq %rax, %rbx movl %r14d, %eax movq %rbx, diff_histogram(%rip) testq %r14, %r14 jle .LBB7_33 # %bb.31: # %.lr.ph39 movq histogram(%rip), %rcx movq h_gpu_histogram(%rip), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB7_32: # =>This Inner Loop Header: Depth=1 movq (%rcx,%rsi,8), %rdi subq (%rdx,%rsi,8), %rdi movq %rdi, (%rbx,%rsi,8) incq %rsi cmpq %rsi, %rax jne .LBB7_32 .LBB7_33: # %._crit_edge40 testl %eax, %eax jle .LBB7_41 # %bb.34: # %.lr.ph.i28.preheader movl $3435973837, %r12d # imm = 0xCCCCCCCD xorl %r14d, %r14d xorl %r15d, %r15d jmp .LBB7_35 .p2align 4, 0x90 .LBB7_39: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.5, %edi xorl %eax, %eax callq printf .LBB7_40: # in Loop: Header=BB7_35 Depth=1 incq %r14 movslq num_buckets(%rip), %rax cmpq %rax, %r14 jge .LBB7_41 .LBB7_35: # %.lr.ph.i28 # =>This Inner Loop Header: Depth=1 movl %r14d, %eax imulq %r12, %rax shrq $34, %rax leal (%rax,%rax,4), %eax cmpl %eax, %r14d jne .LBB7_37 # %bb.36: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.2, %edi movl %r14d, %esi xorl %eax, %eax callq printf .LBB7_37: # in Loop: Header=BB7_35 Depth=1 movq (%rbx,%r14,8), %rsi movl $.L.str.3, %edi xorl %eax, %eax callq printf addq (%rbx,%r14,8), %r15 movl num_buckets(%rip), %eax decl %eax cmpq %rax, %r14 jne .LBB7_39 # %bb.38: # in Loop: Header=BB7_35 Depth=1 movl $.L.str.4, %edi movq %r15, %rsi xorl %eax, %eax callq printf jmp .LBB7_40 .LBB7_41: # %_Z16output_histogramP10hist_entry.exit32 movq d_gpu_histogram(%rip), %rdi callq hipFree movq d_atom_list(%rip), %rdi callq hipFree movq histogram(%rip), %rdi callq free movq atom_list(%rip), %rdi callq free movq h_gpu_histogram(%rip), %rdi callq free movq diff_histogram(%rip), %rdi callq free xorl %eax, %eax addq $120, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end7: .size main, .Lfunc_end7-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB8_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB8_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z10PDH_kernelP10hist_entryP8atomdescxd, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11PDH_kernel2P10hist_entryP8atomdescxd, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end8: .size __hip_module_ctor, .Lfunc_end8-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB9_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB9_2: retq .Lfunc_end9: .size __hip_module_dtor, .Lfunc_end9-__hip_module_dtor .cfi_endproc # -- End function .type histogram,@object # @histogram .bss .globl histogram .p2align 3, 0x0 histogram: .quad 0 .size histogram, 8 .type h_gpu_histogram,@object # @h_gpu_histogram .globl h_gpu_histogram .p2align 3, 0x0 h_gpu_histogram: .quad 0 .size h_gpu_histogram, 8 .type d_gpu_histogram,@object # @d_gpu_histogram .globl d_gpu_histogram .p2align 3, 0x0 d_gpu_histogram: .quad 0 .size d_gpu_histogram, 8 .type diff_histogram,@object # @diff_histogram .globl diff_histogram .p2align 3, 0x0 diff_histogram: .quad 0 .size diff_histogram, 8 .type PDH_acnt,@object # @PDH_acnt .globl PDH_acnt .p2align 3, 0x0 PDH_acnt: .quad 0 # 0x0 .size PDH_acnt, 8 .type num_buckets,@object # @num_buckets .globl num_buckets .p2align 2, 0x0 num_buckets: .long 0 # 0x0 .size num_buckets, 4 .type PDH_res,@object # @PDH_res .globl PDH_res .p2align 3, 0x0 PDH_res: .quad 0x0000000000000000 # double 0 .size PDH_res, 8 .type atom_list,@object # @atom_list .globl atom_list .p2align 3, 0x0 atom_list: .quad 0 .size atom_list, 8 .type d_atom_list,@object # @d_atom_list .globl d_atom_list .p2align 3, 0x0 d_atom_list: .quad 0 .size d_atom_list, 8 .type Idunno,@object # @Idunno .globl Idunno .p2align 2, 0x0 Idunno: .zero 8 .size Idunno, 8 .type startTime,@object # @startTime .globl startTime .p2align 3, 0x0 startTime: .zero 16 .size startTime, 16 .type endTime,@object # @endTime .globl endTime .p2align 3, 0x0 endTime: .zero 16 .size endTime, 16 .type _Z10PDH_kernelP10hist_entryP8atomdescxd,@object # @_Z10PDH_kernelP10hist_entryP8atomdescxd .section .rodata,"a",@progbits .globl _Z10PDH_kernelP10hist_entryP8atomdescxd .p2align 3, 0x0 _Z10PDH_kernelP10hist_entryP8atomdescxd: .quad _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .size _Z10PDH_kernelP10hist_entryP8atomdescxd, 8 .type _Z11PDH_kernel2P10hist_entryP8atomdescxd,@object # @_Z11PDH_kernel2P10hist_entryP8atomdescxd .globl _Z11PDH_kernel2P10hist_entryP8atomdescxd .p2align 3, 0x0 _Z11PDH_kernel2P10hist_entryP8atomdescxd: .quad _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .size _Z11PDH_kernel2P10hist_entryP8atomdescxd, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Running time for CPU version: %ld.%06ld\n" .size .L.str, 41 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Running time for GPU version: %ld.%06ld\n" .size .L.str.1, 41 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "\n%02d: " .size .L.str.2, 8 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "%15lld " .size .L.str.3, 8 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "\n T:%lld \n" .size .L.str.4, 11 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "| " .size .L.str.5, 3 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10PDH_kernelP10hist_entryP8atomdescxd" .size .L__unnamed_1, 40 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z11PDH_kernel2P10hist_entryP8atomdescxd" .size .L__unnamed_2, 41 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Starting CPU..." .size .Lstr, 16 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Starting GPU..." .size .Lstr.1, 16 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Difference: " .size .Lstr.2, 13 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z25__device_stub__PDH_kernelP10hist_entryP8atomdescxd .addrsig_sym _Z26__device_stub__PDH_kernel2P10hist_entryP8atomdescxd .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym d_gpu_histogram .addrsig_sym d_atom_list .addrsig_sym Idunno .addrsig_sym startTime .addrsig_sym endTime .addrsig_sym _Z10PDH_kernelP10hist_entryP8atomdescxd .addrsig_sym _Z11PDH_kernel2P10hist_entryP8atomdescxd .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
// Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
// Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
.file "tmpxft_0001e20d_00000000-6_upscale.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
// Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
#include <hip/hip_runtime.h> // Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> // Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> // Author: Noah Van Der Weide // 3/30/2020 // upscale the image by doubling height and width // fill in empty areas according to neighboring pixels and difference thresholds // THREE APPROACHES // FIRST APPROACH: // each thread will process one pixel // SECOND APPROACH: // each thread will process one original pixel and surrounding pixels // THIRD APPROACH: // each thread will process one original pixel and pixels to the right and below // Two filling approaches: // First: // Tackle everything at once. // Second: // Stretch out original image and fill in adjacent pixels with original pixel value, // Then go through and SAXPY if original pixel differences aren't too great. // dimension of image: upper left = (0,0), bottom right = (width-1, height-1) // *img_original is the original image // *img_new width = *img_original width * 3 - 2 // *img_new width = *img_original height * 3 - 2 // 8 bits per color (0 - 255) // upscale function is called independently for each color. // this allows it to be faster for black and white images as it only needs to be called once. // Can therefore also be applied to images which use a different color map than RGB (JPEG, for example). /* #include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <opencv2/opencv.hpp> #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <vector> #include <stdlib.h> #include <string.h> #include "upscale.cuh" #define THREADS_PER_BLOCK 64 //__global__ void upscale_CUDA(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold) { __global__ void upscale_CUDA(unsigned char * dst, unsigned char * src, int src_width, int src_height, int src_channels, unsigned char threshold) { // not using shared memory right now // there is 48 KB of shared memory available. // images are typically more than that, so I'll have to think about how it could be implemented //extern __shared__ unsigned char pic[]; //int pixel = blockIdx.x * blockdim.x + threadIdx.x; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; // not relevant to code function, but shows how a thread could access a pixel in every channel. // pixel values are from 0 to 255. //for (int k = 0; k < channels; k++){ // img[idx + k]; //} int dst_width = src_width * 3 - 2; //int dst_height = src_height * 3 - 2; //long int dst_elements = dst_width * dst_height * src_channels; //long int src_elements = src_width * src_height * src_channels; int src_stride = src_width * src_channels; int dst_stride = dst_width * src_channels; // if invalid location do nothing. //if (i >= dst_width || j >= dst_height) // is that width or width-1? if (i >= src_width || j >= src_height) return; // all channels for a pixel are grouped together. To access an adjacent pixel, you must add by the number of channels. for (int k = 0; k < src_channels; k++) { int dst_index = (j * 21 + i * 3) + k; // this is strictly for my predefined dst width and height (*3 -2) int src_index = (j * src_width + i) + k; // transfer known src values to dst // to access different channels, the number of elements of the src/dst image must be added to the respective array index. dst[dst_index] = src[src_index]; // vertical comparison acts on src image and applies values to dst image int y_diff = src[src_index + src_stride] - src[src_index]; if (y_diff < threshold) { // apply third-average // linear fill int step = y_diff / 3; dst[dst_index + dst_stride] = src[src_index] + step; dst[dst_index + 2 * dst_stride] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride] = src[src_index]; dst[dst_index + 2 * dst_stride] = src[src_index + src_stride]; } __syncthreads(); // horizontal // I know this is painfully inefficient. int x_diff_0 = src[src_index] - src[src_index + src_channels]; int x_diff_1 = dst[dst_index + dst_stride] - dst[dst_index + dst_stride + src_channels]; int x_diff_2 = dst[dst_index + 2 * dst_stride] - dst[dst_index + 2 * dst_stride + src_channels]; int step = 0; if (x_diff_0 < threshold) { // apply third-average // linear fill step = x_diff_0 / 3; dst[dst_index + 1] = src[src_index] + step; dst[dst_index + 2] = src[src_index] + step * 2; } else { // nearest neighbor dst[dst_index + src_channels] = src[src_index]; dst[dst_index + 2*src_channels] = src[src_index + src_channels]; } if (x_diff_1 < threshold) { // apply third-average // linear fill step = x_diff_1 / 3; dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride] + step; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + dst_stride + src_channels] = dst[dst_index + dst_stride]; dst[dst_index + dst_stride + 2*src_channels] = dst[dst_index + dst_stride + 3]; } if (x_diff_2 < threshold) { // apply third-average // linear fill step = x_diff_2 / 3; dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride] + step; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride] + step * 2; } else { // nearest neighbor dst[dst_index + 2 * dst_stride + src_channels] = dst[dst_index + 2 * dst_stride]; dst[dst_index + 2 * dst_stride + 2*src_channels] = dst[dst_index + 2 * dst_stride + 3]; } __syncthreads(); } __syncthreads(); } __global__ void upscale(unsigned char threshold) { // CUDA timing parameters cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); float ms; // file handling using opencv // IMREAD_COLOR loads image in BGR 8-bit format // IMREAD_UNCHANGED includes alpha channel // IMREAD_GRAYSCALE loads as intensity 0-1 // load image string image_path = samples::findFile("peppers.png"); Mat src = imread(image_path, IMREAD_COLOR); // check if image loaded properly if (src.empty()) { cout << "Could not read image: " << image_path << endl; return 1; } // ------------------------------------------ // properties of the source and upscaled image // ------------------------------------------ // input dimensions int src_height = src.rows; int src_width = src.cols; // channels (e.g. Red, Green, Blue) int channels = src.channels(); //int type = src.type; // CV_8UC3? // output dimensions int dst_height = src_height * 3 - 2; int dst_width = src_width * 3 - 2; // number of elements (if a picture has 3 channels, this is 3 * pixels) int dst_elements = dst_width * dst_height * channels; int src_elements = src_width * src_height * channels; // number of bytes each image will take int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); // create new image with same datatype as input Mat dst(dst_height, dst_width, CV_8UC3, Scalar(0, 0, 0)); //Mat dst(dst_height, dst_width, type); // image data for upscale function unsigned char* src_img = src.data; unsigned char* dst_img = dst.data; cout << "Loaded " << image_path << " -- " << src_height << ", " << src_width << " -- Channels: " << channels << endl; // initialize device variables unsigned char* dev_src, * dev_dst; // number of blocks to call in kernel. Max threads per block is usually 1024 int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); // used for shared memory if eventually implemented //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst_img, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src_img, src_size, cudaMemcpyHostToDevice); // start timer for performance evaluation cudaEventRecord(start); // call upscale function //upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_dst, dev_src, src_elements, src_width, src_height, threshold); // <<<blocks, threads per block, shared mem>>> dim3 grid(src_width, src_height); upscale_CUDA < < <grid, 1 > > >(dev_dst, dev_src, src_width, src_height, channels, threshold); cudaDeviceSynchronize(); // end timer cudaEventRecord(stop); cudaEventSynchronize(stop); cudaEventElapsedTime(&ms, start, stop); // copy data back from GPU to CPU cudaMemcpy(dst_img, dev_dst, dst_size, cudaMemcpyDeviceToHost); cudaMemcpy(src_img, dev_src, dst_size, cudaMemcpyDeviceToHost); // might not need this // free GPU cudaFree(dev_dst); cudaFree(dev_src); // create output image. I might not need another Mat -- just use 'dst' instead of 'output' //Mat output = Mat(dst_height, dst_width, type, dst); imshow("source", src); imshow("output", dst); imwrite("upscaled_image.png", dst); waitKey(0); std::cout << "\ntime (ms) = " << ms << std::endl; } /* void upscale(unsigned char* dst, unsigned char* src, int src_height, int src_width, int src_channels, int threshold){ unsigned char* dev_src, * dev_dst; int dst_width = src_width * 3 - 2; int dst_height = src_height * 3 - 2; int dst_elements = dst_width * dst_height * src_channels; int src_elements = src_width * src_height * src_channels; int dst_size = dst_elements * sizeof(unsigned char); int src_size = src_elements * sizeof(unsigned char); int blocks = (src_elements + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK; // allocate memory in GPU cudaMalloc((void**)&dev_dst, dst_size); cudaMalloc((void**)&dev_src, src_size); //cudaMallocManaged(&dst, dst_size); //cudaMallocManaged(&src, src_size); // copy data from CPU to GPU cudaMemcpy(dev_dst, dst, dst_size, cudaMemcpyHostToDevice); cudaMemcpy(dev_src, src, src_size, cudaMemcpyHostToDevice); upscale_CUDA<<<blocks, THREADS_PER_BLOCK>>> (dev_src, src_height, src_width, src_channels, threshold); // <<<blocks, threads per block, shared mem>>> cudaDeviceSynchronize(); // copy data back from GPU to CPU cudaMemcpy(dst, dev_dst, dst_size, cudaMemcpyDeviceToHost); // free GPU cudaFree(dev_dst); cudaFree(dev_src); } */
.text .file "upscale.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0001e20d_00000000-6_upscale.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "upscale.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #include <stdlib.h> #define N 1024 #define N_THR 512 void fill_ints(int* a, int size){ for(int i =0; i<size; i++) a[i]=i; } __global__ void dotVecs(int *x, int *y, int *r){ __shared__ int s_tmp[N_THR]; int index = threadIdx.x + blockIdx.x * blockDim.x; int temp = x[index] * y[index]; s_tmp[threadIdx.x] = temp; // store the multiplication to the shared memory __syncthreads(); // Thread 0 performs the reduction if(threadIdx.x == 0){ int sum = 0; for(int i = 0 ; i < N_THR ; i++) sum += s_tmp[i]; *r += sum; } } int main(void){ int *a, *b, *c; // host pointers int *d_a, *d_b, *d_c; // device pointers int size = N * sizeof(int); // Alloc space for device copies of a, b, c cudaMalloc((void **)&d_a, size); cudaMalloc((void **)&d_b, size); cudaMalloc((void **)&d_c, sizeof(int)); a = (int *)malloc(size); fill_ints(a, N); // Alloc space host, random initialization b = (int *)malloc(size); fill_ints(b, N); c = (int *)malloc(sizeof(int)); // Copy data from host to device memory // cudaMemcpyHostToDevice is a flag determining copying from host to dev. cudaMemcpy(d_a, a, size, cudaMemcpyHostToDevice); cudaMemcpy(d_b, b, size, cudaMemcpyHostToDevice); cudaMemset(d_c,0,sizeof(int)); // Launch kernel to add two vector with N threads and 1 block // Kernel calls are asynchronous dotVecs<<<2,N_THR>>>(d_a, d_b, d_c); // Copy results from device to host // cudaMemcpy blocks CPU until Kernels finish execution cudaMemcpy(c, d_c, sizeof(int), cudaMemcpyDeviceToHost); printf("%d\n",*c); // needs cudaFree to deallocate device pointers cudaFree(d_a); cudaFree(d_b); cudaFree(d_c); free(a); free(b); free(c); return 0; }
.file "tmpxft_001b0530_00000000-6_ex3_3.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z9fill_intsPii .type _Z9fill_intsPii, @function _Z9fill_intsPii: .LFB2057: .cfi_startproc endbr64 testl %esi, %esi jle .L3 movslq %esi, %rsi movl $0, %eax .L5: movl %eax, (%rdi,%rax,4) addq $1, %rax cmpq %rsi, %rax jne .L5 .L3: ret .cfi_endproc .LFE2057: .size _Z9fill_intsPii, .-_Z9fill_intsPii .globl _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ .type _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_, @function _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_: .LFB2083: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 120(%rsp), %rax subq %fs:40, %rax jne .L12 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7dotVecsPiS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_, .-_Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ .globl _Z7dotVecsPiS_S_ .type _Z7dotVecsPiS_S_, @function _Z7dotVecsPiS_S_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7dotVecsPiS_S_, .-_Z7dotVecsPiS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r12 .cfi_def_cfa_offset 16 .cfi_offset 12, -16 pushq %rbp .cfi_def_cfa_offset 24 .cfi_offset 6, -24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset 3, -32 subq $64, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $4, %esi call cudaMalloc@PLT movl $4096, %edi call malloc@PLT movq %rax, %r12 movl $1024, %esi movq %rax, %rdi call _Z9fill_intsPii movl $4096, %edi call malloc@PLT movq %rax, %rbp movl $1024, %esi movq %rax, %rdi call _Z9fill_intsPii movl $4, %edi call malloc@PLT movq %rax, %rbx movl $1, %ecx movl $4096, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4096, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $4, %edx movl $0, %esi movq 24(%rsp), %rdi call cudaMemset@PLT movl $512, 44(%rsp) movl $1, 48(%rsp) movl $2, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L19 .L16: movl $2, %ecx movl $4, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl (%rbx), %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L20 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ jmp .L16 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z7dotVecsPiS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z7dotVecsPiS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <stdlib.h> #define N 1024 #define N_THR 512 void fill_ints(int* a, int size){ for(int i =0; i<size; i++) a[i]=i; } __global__ void dotVecs(int *x, int *y, int *r){ __shared__ int s_tmp[N_THR]; int index = threadIdx.x + blockIdx.x * blockDim.x; int temp = x[index] * y[index]; s_tmp[threadIdx.x] = temp; // store the multiplication to the shared memory __syncthreads(); // Thread 0 performs the reduction if(threadIdx.x == 0){ int sum = 0; for(int i = 0 ; i < N_THR ; i++) sum += s_tmp[i]; *r += sum; } } int main(void){ int *a, *b, *c; // host pointers int *d_a, *d_b, *d_c; // device pointers int size = N * sizeof(int); // Alloc space for device copies of a, b, c cudaMalloc((void **)&d_a, size); cudaMalloc((void **)&d_b, size); cudaMalloc((void **)&d_c, sizeof(int)); a = (int *)malloc(size); fill_ints(a, N); // Alloc space host, random initialization b = (int *)malloc(size); fill_ints(b, N); c = (int *)malloc(sizeof(int)); // Copy data from host to device memory // cudaMemcpyHostToDevice is a flag determining copying from host to dev. cudaMemcpy(d_a, a, size, cudaMemcpyHostToDevice); cudaMemcpy(d_b, b, size, cudaMemcpyHostToDevice); cudaMemset(d_c,0,sizeof(int)); // Launch kernel to add two vector with N threads and 1 block // Kernel calls are asynchronous dotVecs<<<2,N_THR>>>(d_a, d_b, d_c); // Copy results from device to host // cudaMemcpy blocks CPU until Kernels finish execution cudaMemcpy(c, d_c, sizeof(int), cudaMemcpyDeviceToHost); printf("%d\n",*c); // needs cudaFree to deallocate device pointers cudaFree(d_a); cudaFree(d_b); cudaFree(d_c); free(a); free(b); free(c); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define N 1024 #define N_THR 512 void fill_ints(int* a, int size){ for(int i =0; i<size; i++) a[i]=i; } __global__ void dotVecs(int *x, int *y, int *r){ __shared__ int s_tmp[N_THR]; int index = threadIdx.x + blockIdx.x * blockDim.x; int temp = x[index] * y[index]; s_tmp[threadIdx.x] = temp; // store the multiplication to the shared memory __syncthreads(); // Thread 0 performs the reduction if(threadIdx.x == 0){ int sum = 0; for(int i = 0 ; i < N_THR ; i++) sum += s_tmp[i]; *r += sum; } } int main(void){ int *a, *b, *c; // host pointers int *d_a, *d_b, *d_c; // device pointers int size = N * sizeof(int); // Alloc space for device copies of a, b, c hipMalloc((void **)&d_a, size); hipMalloc((void **)&d_b, size); hipMalloc((void **)&d_c, sizeof(int)); a = (int *)malloc(size); fill_ints(a, N); // Alloc space host, random initialization b = (int *)malloc(size); fill_ints(b, N); c = (int *)malloc(sizeof(int)); // Copy data from host to device memory // cudaMemcpyHostToDevice is a flag determining copying from host to dev. hipMemcpy(d_a, a, size, hipMemcpyHostToDevice); hipMemcpy(d_b, b, size, hipMemcpyHostToDevice); hipMemset(d_c,0,sizeof(int)); // Launch kernel to add two vector with N threads and 1 block // Kernel calls are asynchronous dotVecs<<<2,N_THR>>>(d_a, d_b, d_c); // Copy results from device to host // cudaMemcpy blocks CPU until Kernels finish execution hipMemcpy(c, d_c, sizeof(int), hipMemcpyDeviceToHost); printf("%d\n",*c); // needs cudaFree to deallocate device pointers hipFree(d_a); hipFree(d_b); hipFree(d_c); free(a); free(b); free(c); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define N 1024 #define N_THR 512 void fill_ints(int* a, int size){ for(int i =0; i<size; i++) a[i]=i; } __global__ void dotVecs(int *x, int *y, int *r){ __shared__ int s_tmp[N_THR]; int index = threadIdx.x + blockIdx.x * blockDim.x; int temp = x[index] * y[index]; s_tmp[threadIdx.x] = temp; // store the multiplication to the shared memory __syncthreads(); // Thread 0 performs the reduction if(threadIdx.x == 0){ int sum = 0; for(int i = 0 ; i < N_THR ; i++) sum += s_tmp[i]; *r += sum; } } int main(void){ int *a, *b, *c; // host pointers int *d_a, *d_b, *d_c; // device pointers int size = N * sizeof(int); // Alloc space for device copies of a, b, c hipMalloc((void **)&d_a, size); hipMalloc((void **)&d_b, size); hipMalloc((void **)&d_c, sizeof(int)); a = (int *)malloc(size); fill_ints(a, N); // Alloc space host, random initialization b = (int *)malloc(size); fill_ints(b, N); c = (int *)malloc(sizeof(int)); // Copy data from host to device memory // cudaMemcpyHostToDevice is a flag determining copying from host to dev. hipMemcpy(d_a, a, size, hipMemcpyHostToDevice); hipMemcpy(d_b, b, size, hipMemcpyHostToDevice); hipMemset(d_c,0,sizeof(int)); // Launch kernel to add two vector with N threads and 1 block // Kernel calls are asynchronous dotVecs<<<2,N_THR>>>(d_a, d_b, d_c); // Copy results from device to host // cudaMemcpy blocks CPU until Kernels finish execution hipMemcpy(c, d_c, sizeof(int), hipMemcpyDeviceToHost); printf("%d\n",*c); // needs cudaFree to deallocate device pointers hipFree(d_a); hipFree(d_b); hipFree(d_c); free(a); free(b); free(c); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z7dotVecsPiS_S_ .globl _Z7dotVecsPiS_S_ .p2align 8 .type _Z7dotVecsPiS_S_,@function _Z7dotVecsPiS_S_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b128 s[4:7], s[0:1], 0x0 s_mov_b32 s3, exec_lo s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, 0 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[1:2], 2, v[1:2] v_add_co_u32 v3, vcc_lo, s4, v1 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s5, v2, vcc_lo v_add_co_u32 v1, vcc_lo, s6, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v2, vcc_lo global_load_b32 v3, v[3:4], off global_load_b32 v1, v[1:2], off v_lshlrev_b32_e32 v2, 2, v0 s_waitcnt vmcnt(0) v_mul_lo_u32 v1, v1, v3 ds_store_b32 v2, v1 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv v_cmpx_eq_u32_e32 0, v0 s_cbranch_execz .LBB0_4 v_mov_b32_e32 v0, 0 .LBB0_2: v_mov_b32_e32 v1, s2 s_add_i32 s2, s2, 4 s_delay_alu instid0(SALU_CYCLE_1) s_cmpk_eq_i32 s2, 0x800 ds_load_b32 v1, v1 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v0, v1, v0 s_cbranch_scc0 .LBB0_2 s_load_b64 s[0:1], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_load_b32 s2, s[0:1], 0x0 s_waitcnt lgkmcnt(0) v_dual_mov_b32 v1, 0 :: v_dual_add_nc_u32 v0, s2, v0 global_store_b32 v1, v0, s[0:1] .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7dotVecsPiS_S_ .amdhsa_group_segment_fixed_size 2048 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 280 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 5 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z7dotVecsPiS_S_, .Lfunc_end0-_Z7dotVecsPiS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 2048 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z7dotVecsPiS_S_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z7dotVecsPiS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 5 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define N 1024 #define N_THR 512 void fill_ints(int* a, int size){ for(int i =0; i<size; i++) a[i]=i; } __global__ void dotVecs(int *x, int *y, int *r){ __shared__ int s_tmp[N_THR]; int index = threadIdx.x + blockIdx.x * blockDim.x; int temp = x[index] * y[index]; s_tmp[threadIdx.x] = temp; // store the multiplication to the shared memory __syncthreads(); // Thread 0 performs the reduction if(threadIdx.x == 0){ int sum = 0; for(int i = 0 ; i < N_THR ; i++) sum += s_tmp[i]; *r += sum; } } int main(void){ int *a, *b, *c; // host pointers int *d_a, *d_b, *d_c; // device pointers int size = N * sizeof(int); // Alloc space for device copies of a, b, c hipMalloc((void **)&d_a, size); hipMalloc((void **)&d_b, size); hipMalloc((void **)&d_c, sizeof(int)); a = (int *)malloc(size); fill_ints(a, N); // Alloc space host, random initialization b = (int *)malloc(size); fill_ints(b, N); c = (int *)malloc(sizeof(int)); // Copy data from host to device memory // cudaMemcpyHostToDevice is a flag determining copying from host to dev. hipMemcpy(d_a, a, size, hipMemcpyHostToDevice); hipMemcpy(d_b, b, size, hipMemcpyHostToDevice); hipMemset(d_c,0,sizeof(int)); // Launch kernel to add two vector with N threads and 1 block // Kernel calls are asynchronous dotVecs<<<2,N_THR>>>(d_a, d_b, d_c); // Copy results from device to host // cudaMemcpy blocks CPU until Kernels finish execution hipMemcpy(c, d_c, sizeof(int), hipMemcpyDeviceToHost); printf("%d\n",*c); // needs cudaFree to deallocate device pointers hipFree(d_a); hipFree(d_b); hipFree(d_c); free(a); free(b); free(c); return 0; }
.text .file "ex3_3.hip" .globl _Z9fill_intsPii # -- Begin function _Z9fill_intsPii .p2align 4, 0x90 .type _Z9fill_intsPii,@function _Z9fill_intsPii: # @_Z9fill_intsPii .cfi_startproc # %bb.0: testl %esi, %esi jle .LBB0_3 # %bb.1: # %.lr.ph.preheader movl %esi, %eax xorl %ecx, %ecx .p2align 4, 0x90 .LBB0_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl %ecx, (%rdi,%rcx,4) incq %rcx cmpq %rcx, %rax jne .LBB0_2 .LBB0_3: # %._crit_edge retq .Lfunc_end0: .size _Z9fill_intsPii, .Lfunc_end0-_Z9fill_intsPii .cfi_endproc # -- End function .globl _Z22__device_stub__dotVecsPiS_S_ # -- Begin function _Z22__device_stub__dotVecsPiS_S_ .p2align 4, 0x90 .type _Z22__device_stub__dotVecsPiS_S_,@function _Z22__device_stub__dotVecsPiS_S_: # @_Z22__device_stub__dotVecsPiS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7dotVecsPiS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z22__device_stub__dotVecsPiS_S_, .Lfunc_end1-_Z22__device_stub__dotVecsPiS_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $128, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 leaq 16(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 8(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc movq %rsp, %rdi movl $4, %esi callq hipMalloc movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %rbx xorl %eax, %eax .p2align 4, 0x90 .LBB2_1: # %.lr.ph.i # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB2_1 # %bb.2: # %_Z9fill_intsPii.exit movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r14 xorl %eax, %eax .p2align 4, 0x90 .LBB2_3: # %.lr.ph.i16 # =>This Inner Loop Header: Depth=1 movl %eax, (%r14,%rax,4) incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB2_3 # %bb.4: # %_Z9fill_intsPii.exit20 movl $4, %edi callq malloc movq %rax, %r15 movq 16(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq (%rsp), %rdi movl $4, %edx xorl %esi, %esi callq hipMemset movabsq $4294967298, %rdi # imm = 0x100000002 leaq 510(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_6 # %bb.5: movq 16(%rsp), %rax movq 8(%rsp), %rcx movq (%rsp), %rdx movq %rax, 88(%rsp) movq %rcx, 80(%rsp) movq %rdx, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z7dotVecsPiS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_6: movq (%rsp), %rsi movl $4, %edx movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl (%r15), %esi movl $.L.str, %edi xorl %eax, %eax callq printf movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax addq $128, %rsp .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7dotVecsPiS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z7dotVecsPiS_S_,@object # @_Z7dotVecsPiS_S_ .section .rodata,"a",@progbits .globl _Z7dotVecsPiS_S_ .p2align 3, 0x0 _Z7dotVecsPiS_S_: .quad _Z22__device_stub__dotVecsPiS_S_ .size _Z7dotVecsPiS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d\n" .size .L.str, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7dotVecsPiS_S_" .size .L__unnamed_1, 17 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__dotVecsPiS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7dotVecsPiS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_001b0530_00000000-6_ex3_3.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z9fill_intsPii .type _Z9fill_intsPii, @function _Z9fill_intsPii: .LFB2057: .cfi_startproc endbr64 testl %esi, %esi jle .L3 movslq %esi, %rsi movl $0, %eax .L5: movl %eax, (%rdi,%rax,4) addq $1, %rax cmpq %rsi, %rax jne .L5 .L3: ret .cfi_endproc .LFE2057: .size _Z9fill_intsPii, .-_Z9fill_intsPii .globl _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ .type _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_, @function _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_: .LFB2083: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 120(%rsp), %rax subq %fs:40, %rax jne .L12 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7dotVecsPiS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_, .-_Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ .globl _Z7dotVecsPiS_S_ .type _Z7dotVecsPiS_S_, @function _Z7dotVecsPiS_S_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7dotVecsPiS_S_, .-_Z7dotVecsPiS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r12 .cfi_def_cfa_offset 16 .cfi_offset 12, -16 pushq %rbp .cfi_def_cfa_offset 24 .cfi_offset 6, -24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset 3, -32 subq $64, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $4096, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $4, %esi call cudaMalloc@PLT movl $4096, %edi call malloc@PLT movq %rax, %r12 movl $1024, %esi movq %rax, %rdi call _Z9fill_intsPii movl $4096, %edi call malloc@PLT movq %rax, %rbp movl $1024, %esi movq %rax, %rdi call _Z9fill_intsPii movl $4, %edi call malloc@PLT movq %rax, %rbx movl $1, %ecx movl $4096, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4096, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $4, %edx movl $0, %esi movq 24(%rsp), %rdi call cudaMemset@PLT movl $512, 44(%rsp) movl $1, 48(%rsp) movl $2, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L19 .L16: movl $2, %ecx movl $4, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl (%rbx), %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L20 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z30__device_stub__Z7dotVecsPiS_S_PiS_S_ jmp .L16 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z7dotVecsPiS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z7dotVecsPiS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "ex3_3.hip" .globl _Z9fill_intsPii # -- Begin function _Z9fill_intsPii .p2align 4, 0x90 .type _Z9fill_intsPii,@function _Z9fill_intsPii: # @_Z9fill_intsPii .cfi_startproc # %bb.0: testl %esi, %esi jle .LBB0_3 # %bb.1: # %.lr.ph.preheader movl %esi, %eax xorl %ecx, %ecx .p2align 4, 0x90 .LBB0_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl %ecx, (%rdi,%rcx,4) incq %rcx cmpq %rcx, %rax jne .LBB0_2 .LBB0_3: # %._crit_edge retq .Lfunc_end0: .size _Z9fill_intsPii, .Lfunc_end0-_Z9fill_intsPii .cfi_endproc # -- End function .globl _Z22__device_stub__dotVecsPiS_S_ # -- Begin function _Z22__device_stub__dotVecsPiS_S_ .p2align 4, 0x90 .type _Z22__device_stub__dotVecsPiS_S_,@function _Z22__device_stub__dotVecsPiS_S_: # @_Z22__device_stub__dotVecsPiS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7dotVecsPiS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z22__device_stub__dotVecsPiS_S_, .Lfunc_end1-_Z22__device_stub__dotVecsPiS_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $128, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 leaq 16(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc leaq 8(%rsp), %rdi movl $4096, %esi # imm = 0x1000 callq hipMalloc movq %rsp, %rdi movl $4, %esi callq hipMalloc movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %rbx xorl %eax, %eax .p2align 4, 0x90 .LBB2_1: # %.lr.ph.i # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB2_1 # %bb.2: # %_Z9fill_intsPii.exit movl $4096, %edi # imm = 0x1000 callq malloc movq %rax, %r14 xorl %eax, %eax .p2align 4, 0x90 .LBB2_3: # %.lr.ph.i16 # =>This Inner Loop Header: Depth=1 movl %eax, (%r14,%rax,4) incq %rax cmpq $1024, %rax # imm = 0x400 jne .LBB2_3 # %bb.4: # %_Z9fill_intsPii.exit20 movl $4, %edi callq malloc movq %rax, %r15 movq 16(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movl $4096, %edx # imm = 0x1000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq (%rsp), %rdi movl $4, %edx xorl %esi, %esi callq hipMemset movabsq $4294967298, %rdi # imm = 0x100000002 leaq 510(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_6 # %bb.5: movq 16(%rsp), %rax movq 8(%rsp), %rcx movq (%rsp), %rdx movq %rax, 88(%rsp) movq %rcx, 80(%rsp) movq %rdx, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z7dotVecsPiS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_6: movq (%rsp), %rsi movl $4, %edx movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl (%r15), %esi movl $.L.str, %edi xorl %eax, %eax callq printf movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax addq $128, %rsp .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7dotVecsPiS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z7dotVecsPiS_S_,@object # @_Z7dotVecsPiS_S_ .section .rodata,"a",@progbits .globl _Z7dotVecsPiS_S_ .p2align 3, 0x0 _Z7dotVecsPiS_S_: .quad _Z22__device_stub__dotVecsPiS_S_ .size _Z7dotVecsPiS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d\n" .size .L.str, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7dotVecsPiS_S_" .size .L__unnamed_1, 17 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__dotVecsPiS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7dotVecsPiS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> __global__ void helloFromGPU() { printf("Hello World from GPU (Thread #%d)!\n", threadIdx.x); } int main(void){ printf("Hello World from CPU!\n"); helloFromGPU<<<1,10>>>(); cudaDeviceSynchronize(); printf("Goodbye World from CPU!\n"); cudaDeviceReset(); return 0; }
code for sm_80 Function : _Z12helloFromGPUv .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fc800078e00ff */ /*0010*/ S2R R8, SR_TID.X ; /* 0x0000000000087919 */ /* 0x000e220000002100 */ /*0020*/ IADD3 R1, R1, -0x8, RZ ; /* 0xfffffff801017810 */ /* 0x000fe20007ffe0ff */ /*0030*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff047624 */ /* 0x000fe200078e00ff */ /*0040*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe20000000f00 */ /*0050*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff057624 */ /* 0x000fe200078e00ff */ /*0060*/ IADD3 R6, P0, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fe40007f1e0ff */ /*0070*/ LDC.64 R2, c[0x4][R0] ; /* 0x0100000000027b82 */ /* 0x0002a60000000a00 */ /*0080*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P0 ; /* 0x00000900ff077624 */ /* 0x000fe200000e06ff */ /*0090*/ STL [R1], R8 ; /* 0x0000000801007387 */ /* 0x0013e80000100800 */ /*00a0*/ LEPC R8 ; /* 0x000000000008734e */ /* 0x002fc60000000000 */ /*00b0*/ MOV R11, 0x120 ; /* 0x00000120000b7802 */ /* 0x000fe40000000f00 */ /*00c0*/ MOV R20, 0xa0 ; /* 0x000000a000147802 */ /* 0x000fc40000000f00 */ /*00d0*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fe40000000f00 */ /*00e0*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe40000000f00 */ /*00f0*/ IADD3 R20, P0, P1, -R20, R11, R8 ; /* 0x0000000b14147210 */ /* 0x000fc8000791e108 */ /*0100*/ IADD3.X R21, ~R0, R21, R9, P0, P1 ; /* 0x0000001500157210 */ /* 0x000fc800007e2509 */ /*0110*/ CALL.ABS.NOINC R2 ; /* 0x0000000002007343 */ /* 0x004fea0003c00000 */ /*0120*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0130*/ BRA 0x130; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> __global__ void helloFromGPU() { printf("Hello World from GPU (Thread #%d)!\n", threadIdx.x); } int main(void){ printf("Hello World from CPU!\n"); helloFromGPU<<<1,10>>>(); cudaDeviceSynchronize(); printf("Goodbye World from CPU!\n"); cudaDeviceReset(); return 0; }
.file "tmpxft_00013cc2_00000000-6_hello_world_gpu_02.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z12helloFromGPUvv .type _Z31__device_stub__Z12helloFromGPUvv, @function _Z31__device_stub__Z12helloFromGPUvv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z12helloFromGPUv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z31__device_stub__Z12helloFromGPUvv, .-_Z31__device_stub__Z12helloFromGPUvv .globl _Z12helloFromGPUv .type _Z12helloFromGPUv, @function _Z12helloFromGPUv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z12helloFromGPUvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z12helloFromGPUv, .-_Z12helloFromGPUv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Hello World from CPU!\n" .LC1: .string "Goodbye World from CPU!\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $10, 20(%rsp) movl $1, 24(%rsp) movl $1, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call cudaDeviceReset@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z31__device_stub__Z12helloFromGPUvv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC2: .string "_Z12helloFromGPUv" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _Z12helloFromGPUv(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> __global__ void helloFromGPU() { printf("Hello World from GPU (Thread #%d)!\n", threadIdx.x); } int main(void){ printf("Hello World from CPU!\n"); helloFromGPU<<<1,10>>>(); cudaDeviceSynchronize(); printf("Goodbye World from CPU!\n"); cudaDeviceReset(); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void helloFromGPU() { printf("Hello World from GPU (Thread #%d)!\n", threadIdx.x); } int main(void){ printf("Hello World from CPU!\n"); helloFromGPU<<<1,10>>>(); hipDeviceSynchronize(); printf("Goodbye World from CPU!\n"); hipDeviceReset(); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void helloFromGPU() { printf("Hello World from GPU (Thread #%d)!\n", threadIdx.x); } int main(void){ printf("Hello World from CPU!\n"); helloFromGPU<<<1,10>>>(); hipDeviceSynchronize(); printf("Goodbye World from CPU!\n"); hipDeviceReset(); return 0; }
.text .file "hello_world_gpu_02.hip" .globl _Z27__device_stub__helloFromGPUv # -- Begin function _Z27__device_stub__helloFromGPUv .p2align 4, 0x90 .type _Z27__device_stub__helloFromGPUv,@function _Z27__device_stub__helloFromGPUv: # @_Z27__device_stub__helloFromGPUv .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z12helloFromGPUv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $72, %rsp .cfi_adjust_cfa_offset -72 retq .Lfunc_end0: .size _Z27__device_stub__helloFromGPUv, .Lfunc_end0-_Z27__device_stub__helloFromGPUv .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movl $.Lstr, %edi callq puts@PLT movabsq $4294967297, %rdi # imm = 0x100000001 leaq 9(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z12helloFromGPUv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize movl $.Lstr.1, %edi callq puts@PLT callq hipDeviceReset xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z12helloFromGPUv, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z12helloFromGPUv,@object # @_Z12helloFromGPUv .section .rodata,"a",@progbits .globl _Z12helloFromGPUv .p2align 3, 0x0 _Z12helloFromGPUv: .quad _Z27__device_stub__helloFromGPUv .size _Z12helloFromGPUv, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z12helloFromGPUv" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Hello World from CPU!" .size .Lstr, 22 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Goodbye World from CPU!" .size .Lstr.1, 24 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z27__device_stub__helloFromGPUv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z12helloFromGPUv .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00013cc2_00000000-6_hello_world_gpu_02.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z12helloFromGPUvv .type _Z31__device_stub__Z12helloFromGPUvv, @function _Z31__device_stub__Z12helloFromGPUvv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z12helloFromGPUv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z31__device_stub__Z12helloFromGPUvv, .-_Z31__device_stub__Z12helloFromGPUvv .globl _Z12helloFromGPUv .type _Z12helloFromGPUv, @function _Z12helloFromGPUv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z12helloFromGPUvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z12helloFromGPUv, .-_Z12helloFromGPUv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Hello World from CPU!\n" .LC1: .string "Goodbye World from CPU!\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $10, 20(%rsp) movl $1, 24(%rsp) movl $1, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call cudaDeviceReset@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z31__device_stub__Z12helloFromGPUvv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC2: .string "_Z12helloFromGPUv" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _Z12helloFromGPUv(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello_world_gpu_02.hip" .globl _Z27__device_stub__helloFromGPUv # -- Begin function _Z27__device_stub__helloFromGPUv .p2align 4, 0x90 .type _Z27__device_stub__helloFromGPUv,@function _Z27__device_stub__helloFromGPUv: # @_Z27__device_stub__helloFromGPUv .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z12helloFromGPUv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $72, %rsp .cfi_adjust_cfa_offset -72 retq .Lfunc_end0: .size _Z27__device_stub__helloFromGPUv, .Lfunc_end0-_Z27__device_stub__helloFromGPUv .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movl $.Lstr, %edi callq puts@PLT movabsq $4294967297, %rdi # imm = 0x100000001 leaq 9(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z12helloFromGPUv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize movl $.Lstr.1, %edi callq puts@PLT callq hipDeviceReset xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z12helloFromGPUv, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z12helloFromGPUv,@object # @_Z12helloFromGPUv .section .rodata,"a",@progbits .globl _Z12helloFromGPUv .p2align 3, 0x0 _Z12helloFromGPUv: .quad _Z27__device_stub__helloFromGPUv .size _Z12helloFromGPUv, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z12helloFromGPUv" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Hello World from CPU!" .size .Lstr, 22 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Goodbye World from CPU!" .size .Lstr.1, 24 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z27__device_stub__helloFromGPUv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z12helloFromGPUv .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #include <stdlib.h> __global__ void update_h(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j>0 && k>0 ) hx[idx] -= 0.5*( ez[idx] - ez[idx-nz] - ey[idx] + ey[idx-1] ); if( i>0 && k>0 ) hy[idx] -= 0.5*( ex[idx] - ex[idx-1] - ez[idx] + ez[idx-nyz] ); if( i>0 && j>0 ) hz[idx] -= 0.5*( ey[idx] - ey[idx-nyz] - ex[idx] + ex[idx-nz] ); } __global__ void update_e(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez, float *cex, float *cey, float *cez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j<ny-1 && k<nz-1 ) ex[idx] += cex[idx]*( hz[idx+nz] - hz[idx] - hy[idx+1] + hy[idx] ); if( i<nx-1 && k<nz-1 ) ey[idx] += cey[idx]*( hx[idx+1] - hx[idx] - hz[idx+nyz] + hz[idx] ); if( i<nx-1 && j<ny-1 ) ez[idx] += cez[idx]*( hy[idx+nyz] - hy[idx] - hx[idx+nz] + hx[idx] ); } __global__ void update_src(int nx, int ny, int nz, int nyz, float tn, float *f) { int idx = threadIdx.x; int ijk = (nx/2)*nyz + (ny/2)*nz + idx; if( idx < nz ) f[ijk] += sin(0.1*tn); } __global__ void init_zero(int n, int idx0, float *f) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; if( idx < n ) f[idx] = 0; } int main() { int i, n, nx, ny, nz, tn, tmax; nx = 320; ny = 480; nz = 480; n = nx*ny*nz; tmax = 100000; printf("Simple FDTD simulation\n", nx, ny, nz); printf("Array size : %dx%dx%d\n", nx, ny, nz); printf("Total used memory : %1.2f GB\n", n*4*9./(1024*1024*1024)); printf("Iteration : %d step\n", tmax); // memory allocate float *f, *cf; f = (float *) calloc (n, sizeof(float)); cf = (float *) calloc (n, sizeof(float)); for( i=0; i<n; i++ ) cf[i] = 0.5; float *hx_gpu, *hy_gpu, *hz_gpu; float *ex_gpu, *ey_gpu, *ez_gpu; float *cex_gpu, *cey_gpu, *cez_gpu; cudaMalloc ( (void**) &hx_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &hy_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &hz_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ex_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ey_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ez_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cex_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cey_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cez_gpu, n*sizeof(float) ); cudaMemcpy ( cex_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); cudaMemcpy ( cey_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); cudaMemcpy ( cez_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); int ng = 6; // number of grid int tpb = 256; // threads per block int bpg = n/tpb/ng; // blocks per grid for( i=0; i<ng; i++) { init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hx_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hy_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hz_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ex_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ey_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ez_gpu); } // main loop for( tn=0; tn<tmax; tn++ ) { for( i=0; i<ng; i++) update_h <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu); for( i=0; i<ng; i++) update_e <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu, cex_gpu, cey_gpu, cez_gpu); update_src <<<dim3(1),dim3(512)>>> (nx, ny, nz, ny*nz, tn, ez_gpu); } cudaMemcpy( f, ez_gpu, n*sizeof(float), cudaMemcpyDeviceToHost ); printf("Complete.\n"); return 0; }
.file "tmpxft_001a0ad3_00000000-6_012-row_major.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_ .type _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_, @function _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_: .LFB2082: .cfi_startproc endbr64 subq $248, %rsp .cfi_def_cfa_offset 256 movl %edi, 76(%rsp) movl %esi, 72(%rsp) movl %edx, 68(%rsp) movl %ecx, 64(%rsp) movl %r8d, 60(%rsp) movq %r9, 48(%rsp) movq 256(%rsp), %rax movq %rax, 40(%rsp) movq 264(%rsp), %rax movq %rax, 32(%rsp) movq 272(%rsp), %rax movq %rax, 24(%rsp) movq 280(%rsp), %rax movq %rax, 16(%rsp) movq 288(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 232(%rsp) xorl %eax, %eax leaq 76(%rsp), %rax movq %rax, 144(%rsp) leaq 72(%rsp), %rax movq %rax, 152(%rsp) leaq 68(%rsp), %rax movq %rax, 160(%rsp) leaq 64(%rsp), %rax movq %rax, 168(%rsp) leaq 60(%rsp), %rax movq %rax, 176(%rsp) leaq 48(%rsp), %rax movq %rax, 184(%rsp) leaq 40(%rsp), %rax movq %rax, 192(%rsp) leaq 32(%rsp), %rax movq %rax, 200(%rsp) leaq 24(%rsp), %rax movq %rax, 208(%rsp) leaq 16(%rsp), %rax movq %rax, 216(%rsp) leaq 8(%rsp), %rax movq %rax, 224(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $1, 104(%rsp) movl $1, 108(%rsp) movl $1, 112(%rsp) movl $1, 116(%rsp) leaq 88(%rsp), %rcx leaq 80(%rsp), %rdx leaq 108(%rsp), %rsi leaq 96(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 232(%rsp), %rax subq %fs:40, %rax jne .L8 addq $248, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 88(%rsp) .cfi_def_cfa_offset 264 pushq 88(%rsp) .cfi_def_cfa_offset 272 leaq 160(%rsp), %r9 movq 124(%rsp), %rcx movl 132(%rsp), %r8d movq 112(%rsp), %rsi movl 120(%rsp), %edx leaq _Z8update_hiiiiiPfS_S_S_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 256 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_, .-_Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_ .globl _Z8update_hiiiiiPfS_S_S_S_S_ .type _Z8update_hiiiiiPfS_S_S_S_S_, @function _Z8update_hiiiiiPfS_S_S_S_S_: .LFB2083: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 56(%rsp) .cfi_def_cfa_offset 32 pushq 56(%rsp) .cfi_def_cfa_offset 40 pushq 56(%rsp) .cfi_def_cfa_offset 48 pushq 56(%rsp) .cfi_def_cfa_offset 56 pushq 56(%rsp) .cfi_def_cfa_offset 64 call _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_ addq $56, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z8update_hiiiiiPfS_S_S_S_S_, .-_Z8update_hiiiiiPfS_S_S_S_S_ .globl _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_ .type _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_, @function _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_: .LFB2084: .cfi_startproc endbr64 subq $296, %rsp .cfi_def_cfa_offset 304 movl %edi, 92(%rsp) movl %esi, 88(%rsp) movl %edx, 84(%rsp) movl %ecx, 80(%rsp) movl %r8d, 76(%rsp) movq %r9, 64(%rsp) movq 304(%rsp), %rax movq %rax, 56(%rsp) movq 312(%rsp), %rax movq %rax, 48(%rsp) movq 320(%rsp), %rax movq %rax, 40(%rsp) movq 328(%rsp), %rax movq %rax, 32(%rsp) movq 336(%rsp), %rax movq %rax, 24(%rsp) movq 344(%rsp), %rax movq %rax, 16(%rsp) movq 352(%rsp), %rax movq %rax, 8(%rsp) movq 360(%rsp), %rax movq %rax, (%rsp) movq %fs:40, %rax movq %rax, 280(%rsp) xorl %eax, %eax leaq 92(%rsp), %rax movq %rax, 160(%rsp) leaq 88(%rsp), %rax movq %rax, 168(%rsp) leaq 84(%rsp), %rax movq %rax, 176(%rsp) leaq 80(%rsp), %rax movq %rax, 184(%rsp) leaq 76(%rsp), %rax movq %rax, 192(%rsp) leaq 64(%rsp), %rax movq %rax, 200(%rsp) leaq 56(%rsp), %rax movq %rax, 208(%rsp) leaq 48(%rsp), %rax movq %rax, 216(%rsp) leaq 40(%rsp), %rax movq %rax, 224(%rsp) leaq 32(%rsp), %rax movq %rax, 232(%rsp) leaq 24(%rsp), %rax movq %rax, 240(%rsp) leaq 16(%rsp), %rax movq %rax, 248(%rsp) leaq 8(%rsp), %rax movq %rax, 256(%rsp) movq %rsp, %rax movq %rax, 264(%rsp) movl $1, 112(%rsp) movl $1, 116(%rsp) movl $1, 120(%rsp) movl $1, 124(%rsp) movl $1, 128(%rsp) movl $1, 132(%rsp) leaq 104(%rsp), %rcx leaq 96(%rsp), %rdx leaq 124(%rsp), %rsi leaq 112(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 280(%rsp), %rax subq %fs:40, %rax jne .L16 addq $296, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 104(%rsp) .cfi_def_cfa_offset 312 pushq 104(%rsp) .cfi_def_cfa_offset 320 leaq 176(%rsp), %r9 movq 140(%rsp), %rcx movl 148(%rsp), %r8d movq 128(%rsp), %rsi movl 136(%rsp), %edx leaq _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 304 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_, .-_Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_ .globl _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_ .type _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, @function _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 72(%rsp) .cfi_def_cfa_offset 24 pushq 72(%rsp) .cfi_def_cfa_offset 32 pushq 72(%rsp) .cfi_def_cfa_offset 40 pushq 72(%rsp) .cfi_def_cfa_offset 48 pushq 72(%rsp) .cfi_def_cfa_offset 56 pushq 72(%rsp) .cfi_def_cfa_offset 64 pushq 72(%rsp) .cfi_def_cfa_offset 72 pushq 72(%rsp) .cfi_def_cfa_offset 80 call _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_ addq $72, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, .-_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_ .globl _Z35__device_stub__Z10update_srciiiifPfiiiifPf .type _Z35__device_stub__Z10update_srciiiifPfiiiifPf, @function _Z35__device_stub__Z10update_srciiiifPfiiiifPf: .LFB2086: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movl %edi, 28(%rsp) movl %esi, 24(%rsp) movl %edx, 20(%rsp) movl %ecx, 16(%rsp) movss %xmm0, 12(%rsp) movq %r8, (%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 24(%rsp), %rax movq %rax, 104(%rsp) leaq 20(%rsp), %rax movq %rax, 112(%rsp) leaq 16(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L23 .L19: movq 152(%rsp), %rax subq %fs:40, %rax jne .L24 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 184 pushq 40(%rsp) .cfi_def_cfa_offset 192 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z10update_srciiiifPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L19 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2086: .size _Z35__device_stub__Z10update_srciiiifPfiiiifPf, .-_Z35__device_stub__Z10update_srciiiifPfiiiifPf .globl _Z10update_srciiiifPf .type _Z10update_srciiiifPf, @function _Z10update_srciiiifPf: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10update_srciiiifPfiiiifPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _Z10update_srciiiifPf, .-_Z10update_srciiiifPf .globl _Z30__device_stub__Z9init_zeroiiPfiiPf .type _Z30__device_stub__Z9init_zeroiiPfiiPf, @function _Z30__device_stub__Z9init_zeroiiPfiiPf: .LFB2088: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movq %rdx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L31 .L27: movq 104(%rsp), %rax subq %fs:40, %rax jne .L32 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L31: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z9init_zeroiiPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L27 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2088: .size _Z30__device_stub__Z9init_zeroiiPfiiPf, .-_Z30__device_stub__Z9init_zeroiiPfiiPf .globl _Z9init_zeroiiPf .type _Z9init_zeroiiPf, @function _Z9init_zeroiiPf: .LFB2089: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z9init_zeroiiPfiiPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2089: .size _Z9init_zeroiiPf, .-_Z9init_zeroiiPf .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Simple FDTD simulation\n" .LC1: .string "Array size : %dx%dx%d\n" .LC3: .string "Total used memory : %1.2f GB\n" .LC4: .string "Iteration : %d step\n" .LC6: .string "Complete.\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r12 .cfi_def_cfa_offset 16 .cfi_offset 12, -16 pushq %rbp .cfi_def_cfa_offset 24 .cfi_offset 6, -24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset 3, -32 subq $112, %rsp .cfi_def_cfa_offset 144 movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax movl $480, %r8d movl $480, %ecx movl $320, %edx leaq .LC0(%rip), %rsi movl $2, %edi call __printf_chk@PLT movl $480, %r8d movl $480, %ecx movl $320, %edx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movsd .LC2(%rip), %xmm0 leaq .LC3(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl $100000, %edx leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $4, %esi movl $73728000, %edi call calloc@PLT movq %rax, %r12 movl $4, %esi movl $73728000, %edi call calloc@PLT movq %rax, %rbx leaq 294912000(%rax), %rdx movss .LC5(%rip), %xmm0 .L36: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L36 leaq 8(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 40(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 48(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 56(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 64(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT leaq 72(%rsp), %rdi movl $294912000, %esi call cudaMalloc@PLT movl $1, %ecx movl $294912000, %edx movq %rbx, %rsi movq 56(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $294912000, %edx movq %rbx, %rsi movq 64(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $294912000, %edx movq %rbx, %rsi movq 72(%rsp), %rdi call cudaMemcpy@PLT movl $0, %ebx jmp .L43 .L58: movq 8(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L37 .L59: movq 16(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L38 .L60: movq 24(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L39 .L61: movq 32(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L40 .L62: movq 40(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L41 .L42: addl $12288000, %ebx cmpl $73728000, %ebx je .L57 .L43: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L58 .L37: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L59 .L38: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L60 .L39: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L61 .L40: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L62 .L41: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax jne .L42 movq 48(%rsp), %rdx movl %ebx, %esi movl $73728000, %edi call _Z30__device_stub__Z9init_zeroiiPfiiPf jmp .L42 .L57: movl $0, %ebp jmp .L44 .L45: addl $12288000, %ebx cmpl $73728000, %ebx je .L63 .L46: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax jne .L45 subq $8, %rsp .cfi_def_cfa_offset 152 pushq 56(%rsp) .cfi_def_cfa_offset 160 pushq 56(%rsp) .cfi_def_cfa_offset 168 pushq 56(%rsp) .cfi_def_cfa_offset 176 pushq 56(%rsp) .cfi_def_cfa_offset 184 pushq 56(%rsp) .cfi_def_cfa_offset 192 movq 56(%rsp), %r9 movl %ebx, %r8d movl $230400, %ecx movl $480, %edx movl $480, %esi movl $320, %edi call _Z42__device_stub__Z8update_hiiiiiPfS_S_S_S_S_iiiiiPfS_S_S_S_S_ addq $48, %rsp .cfi_def_cfa_offset 144 jmp .L45 .L63: movl $0, %ebx jmp .L48 .L47: addl $12288000, %ebx cmpl $73728000, %ebx je .L64 .L48: movl $256, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $48000, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax jne .L47 pushq 72(%rsp) .cfi_def_cfa_offset 152 pushq 72(%rsp) .cfi_def_cfa_offset 160 pushq 72(%rsp) .cfi_def_cfa_offset 168 pushq 72(%rsp) .cfi_def_cfa_offset 176 pushq 72(%rsp) .cfi_def_cfa_offset 184 pushq 72(%rsp) .cfi_def_cfa_offset 192 pushq 72(%rsp) .cfi_def_cfa_offset 200 pushq 72(%rsp) .cfi_def_cfa_offset 208 movq 72(%rsp), %r9 movl %ebx, %r8d movl $230400, %ecx movl $480, %edx movl $480, %esi movl $320, %edi call _Z48__device_stub__Z8update_eiiiiiPfS_S_S_S_S_S_S_S_iiiiiPfS_S_S_S_S_S_S_S_ addq $64, %rsp .cfi_def_cfa_offset 144 jmp .L47 .L64: movl $512, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L65 .L49: addl $1, %ebp cmpl $100000, %ebp je .L50 .L44: movl $0, %ebx jmp .L46 .L65: pxor %xmm0, %xmm0 cvtsi2ssl %ebp, %xmm0 movq 48(%rsp), %r8 movl $230400, %ecx movl $480, %edx movl $480, %esi movl $320, %edi call _Z35__device_stub__Z10update_srciiiifPfiiiifPf jmp .L49 .L50: movl $2, %ecx movl $294912000, %edx movq 48(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT leaq .LC6(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 104(%rsp), %rax subq %fs:40, %rax jne .L66 movl $0, %eax addq $112, %rsp .cfi_remember_state .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .L66: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC7: .string "_Z9init_zeroiiPf" .LC8: .string "_Z10update_srciiiifPf" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC9: .string "_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_" .section .rodata.str1.1 .LC10: .string "_Z8update_hiiiiiPfS_S_S_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2091: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z9init_zeroiiPf(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _Z10update_srciiiifPf(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _Z8update_hiiiiiPfS_S_S_S_S_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2091: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC2: .long 0 .long 1073989248 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC5: .long 1056964608 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <stdlib.h> __global__ void update_h(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j>0 && k>0 ) hx[idx] -= 0.5*( ez[idx] - ez[idx-nz] - ey[idx] + ey[idx-1] ); if( i>0 && k>0 ) hy[idx] -= 0.5*( ex[idx] - ex[idx-1] - ez[idx] + ez[idx-nyz] ); if( i>0 && j>0 ) hz[idx] -= 0.5*( ey[idx] - ey[idx-nyz] - ex[idx] + ex[idx-nz] ); } __global__ void update_e(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez, float *cex, float *cey, float *cez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j<ny-1 && k<nz-1 ) ex[idx] += cex[idx]*( hz[idx+nz] - hz[idx] - hy[idx+1] + hy[idx] ); if( i<nx-1 && k<nz-1 ) ey[idx] += cey[idx]*( hx[idx+1] - hx[idx] - hz[idx+nyz] + hz[idx] ); if( i<nx-1 && j<ny-1 ) ez[idx] += cez[idx]*( hy[idx+nyz] - hy[idx] - hx[idx+nz] + hx[idx] ); } __global__ void update_src(int nx, int ny, int nz, int nyz, float tn, float *f) { int idx = threadIdx.x; int ijk = (nx/2)*nyz + (ny/2)*nz + idx; if( idx < nz ) f[ijk] += sin(0.1*tn); } __global__ void init_zero(int n, int idx0, float *f) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; if( idx < n ) f[idx] = 0; } int main() { int i, n, nx, ny, nz, tn, tmax; nx = 320; ny = 480; nz = 480; n = nx*ny*nz; tmax = 100000; printf("Simple FDTD simulation\n", nx, ny, nz); printf("Array size : %dx%dx%d\n", nx, ny, nz); printf("Total used memory : %1.2f GB\n", n*4*9./(1024*1024*1024)); printf("Iteration : %d step\n", tmax); // memory allocate float *f, *cf; f = (float *) calloc (n, sizeof(float)); cf = (float *) calloc (n, sizeof(float)); for( i=0; i<n; i++ ) cf[i] = 0.5; float *hx_gpu, *hy_gpu, *hz_gpu; float *ex_gpu, *ey_gpu, *ez_gpu; float *cex_gpu, *cey_gpu, *cez_gpu; cudaMalloc ( (void**) &hx_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &hy_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &hz_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ex_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ey_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &ez_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cex_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cey_gpu, n*sizeof(float) ); cudaMalloc ( (void**) &cez_gpu, n*sizeof(float) ); cudaMemcpy ( cex_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); cudaMemcpy ( cey_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); cudaMemcpy ( cez_gpu, cf, n*sizeof(float), cudaMemcpyHostToDevice ); int ng = 6; // number of grid int tpb = 256; // threads per block int bpg = n/tpb/ng; // blocks per grid for( i=0; i<ng; i++) { init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hx_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hy_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hz_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ex_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ey_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ez_gpu); } // main loop for( tn=0; tn<tmax; tn++ ) { for( i=0; i<ng; i++) update_h <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu); for( i=0; i<ng; i++) update_e <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu, cex_gpu, cey_gpu, cez_gpu); update_src <<<dim3(1),dim3(512)>>> (nx, ny, nz, ny*nz, tn, ez_gpu); } cudaMemcpy( f, ez_gpu, n*sizeof(float), cudaMemcpyDeviceToHost ); printf("Complete.\n"); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void update_h(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j>0 && k>0 ) hx[idx] -= 0.5*( ez[idx] - ez[idx-nz] - ey[idx] + ey[idx-1] ); if( i>0 && k>0 ) hy[idx] -= 0.5*( ex[idx] - ex[idx-1] - ez[idx] + ez[idx-nyz] ); if( i>0 && j>0 ) hz[idx] -= 0.5*( ey[idx] - ey[idx-nyz] - ex[idx] + ex[idx-nz] ); } __global__ void update_e(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez, float *cex, float *cey, float *cez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j<ny-1 && k<nz-1 ) ex[idx] += cex[idx]*( hz[idx+nz] - hz[idx] - hy[idx+1] + hy[idx] ); if( i<nx-1 && k<nz-1 ) ey[idx] += cey[idx]*( hx[idx+1] - hx[idx] - hz[idx+nyz] + hz[idx] ); if( i<nx-1 && j<ny-1 ) ez[idx] += cez[idx]*( hy[idx+nyz] - hy[idx] - hx[idx+nz] + hx[idx] ); } __global__ void update_src(int nx, int ny, int nz, int nyz, float tn, float *f) { int idx = threadIdx.x; int ijk = (nx/2)*nyz + (ny/2)*nz + idx; if( idx < nz ) f[ijk] += sin(0.1*tn); } __global__ void init_zero(int n, int idx0, float *f) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; if( idx < n ) f[idx] = 0; } int main() { int i, n, nx, ny, nz, tn, tmax; nx = 320; ny = 480; nz = 480; n = nx*ny*nz; tmax = 100000; printf("Simple FDTD simulation\n", nx, ny, nz); printf("Array size : %dx%dx%d\n", nx, ny, nz); printf("Total used memory : %1.2f GB\n", n*4*9./(1024*1024*1024)); printf("Iteration : %d step\n", tmax); // memory allocate float *f, *cf; f = (float *) calloc (n, sizeof(float)); cf = (float *) calloc (n, sizeof(float)); for( i=0; i<n; i++ ) cf[i] = 0.5; float *hx_gpu, *hy_gpu, *hz_gpu; float *ex_gpu, *ey_gpu, *ez_gpu; float *cex_gpu, *cey_gpu, *cez_gpu; hipMalloc ( (void**) &hx_gpu, n*sizeof(float) ); hipMalloc ( (void**) &hy_gpu, n*sizeof(float) ); hipMalloc ( (void**) &hz_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ex_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ey_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ez_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cex_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cey_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cez_gpu, n*sizeof(float) ); hipMemcpy ( cex_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); hipMemcpy ( cey_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); hipMemcpy ( cez_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); int ng = 6; // number of grid int tpb = 256; // threads per block int bpg = n/tpb/ng; // blocks per grid for( i=0; i<ng; i++) { init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hx_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hy_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hz_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ex_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ey_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ez_gpu); } // main loop for( tn=0; tn<tmax; tn++ ) { for( i=0; i<ng; i++) update_h <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu); for( i=0; i<ng; i++) update_e <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu, cex_gpu, cey_gpu, cez_gpu); update_src <<<dim3(1),dim3(512)>>> (nx, ny, nz, ny*nz, tn, ez_gpu); } hipMemcpy( f, ez_gpu, n*sizeof(float), hipMemcpyDeviceToHost ); printf("Complete.\n"); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void update_h(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j>0 && k>0 ) hx[idx] -= 0.5*( ez[idx] - ez[idx-nz] - ey[idx] + ey[idx-1] ); if( i>0 && k>0 ) hy[idx] -= 0.5*( ex[idx] - ex[idx-1] - ez[idx] + ez[idx-nyz] ); if( i>0 && j>0 ) hz[idx] -= 0.5*( ey[idx] - ey[idx-nyz] - ex[idx] + ex[idx-nz] ); } __global__ void update_e(int nx, int ny, int nz, int nyz, int idx0, float *hx, float *hy, float *hz, float *ex, float *ey, float *ez, float *cex, float *cey, float *cez) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; int i = idx/(nyz); int j = (idx - i*nyz)/nz; int k = idx%nz; if( j<ny-1 && k<nz-1 ) ex[idx] += cex[idx]*( hz[idx+nz] - hz[idx] - hy[idx+1] + hy[idx] ); if( i<nx-1 && k<nz-1 ) ey[idx] += cey[idx]*( hx[idx+1] - hx[idx] - hz[idx+nyz] + hz[idx] ); if( i<nx-1 && j<ny-1 ) ez[idx] += cez[idx]*( hy[idx+nyz] - hy[idx] - hx[idx+nz] + hx[idx] ); } __global__ void update_src(int nx, int ny, int nz, int nyz, float tn, float *f) { int idx = threadIdx.x; int ijk = (nx/2)*nyz + (ny/2)*nz + idx; if( idx < nz ) f[ijk] += sin(0.1*tn); } __global__ void init_zero(int n, int idx0, float *f) { int tx = threadIdx.x; int idx = blockDim.x*blockIdx.x + tx + idx0; if( idx < n ) f[idx] = 0; } int main() { int i, n, nx, ny, nz, tn, tmax; nx = 320; ny = 480; nz = 480; n = nx*ny*nz; tmax = 100000; printf("Simple FDTD simulation\n", nx, ny, nz); printf("Array size : %dx%dx%d\n", nx, ny, nz); printf("Total used memory : %1.2f GB\n", n*4*9./(1024*1024*1024)); printf("Iteration : %d step\n", tmax); // memory allocate float *f, *cf; f = (float *) calloc (n, sizeof(float)); cf = (float *) calloc (n, sizeof(float)); for( i=0; i<n; i++ ) cf[i] = 0.5; float *hx_gpu, *hy_gpu, *hz_gpu; float *ex_gpu, *ey_gpu, *ez_gpu; float *cex_gpu, *cey_gpu, *cez_gpu; hipMalloc ( (void**) &hx_gpu, n*sizeof(float) ); hipMalloc ( (void**) &hy_gpu, n*sizeof(float) ); hipMalloc ( (void**) &hz_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ex_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ey_gpu, n*sizeof(float) ); hipMalloc ( (void**) &ez_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cex_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cey_gpu, n*sizeof(float) ); hipMalloc ( (void**) &cez_gpu, n*sizeof(float) ); hipMemcpy ( cex_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); hipMemcpy ( cey_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); hipMemcpy ( cez_gpu, cf, n*sizeof(float), hipMemcpyHostToDevice ); int ng = 6; // number of grid int tpb = 256; // threads per block int bpg = n/tpb/ng; // blocks per grid for( i=0; i<ng; i++) { init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hx_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hy_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, hz_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ex_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ey_gpu); init_zero <<<dim3(bpg),dim3(tpb)>>> (n, i*bpg*tpb, ez_gpu); } // main loop for( tn=0; tn<tmax; tn++ ) { for( i=0; i<ng; i++) update_h <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu); for( i=0; i<ng; i++) update_e <<<dim3(bpg),dim3(tpb)>>> (nx, ny, nz, ny*nz, i*bpg*tpb, hx_gpu, hy_gpu, hz_gpu, ex_gpu, ey_gpu, ez_gpu, cex_gpu, cey_gpu, cez_gpu); update_src <<<dim3(1),dim3(512)>>> (nx, ny, nz, ny*nz, tn, ez_gpu); } hipMemcpy( f, ez_gpu, n*sizeof(float), hipMemcpyDeviceToHost ); printf("Complete.\n"); return 0; }
.text .file "012-row_major.hip" .globl _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ # -- Begin function _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ .p2align 4, 0x90 .type _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_,@function _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_: # @_Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movq %r9, 72(%rsp) leaq 20(%rsp), %rax movq %rax, 80(%rsp) leaq 16(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) leaq 72(%rsp), %rax movq %rax, 120(%rsp) leaq 176(%rsp), %rax movq %rax, 128(%rsp) leaq 184(%rsp), %rax movq %rax, 136(%rsp) leaq 192(%rsp), %rax movq %rax, 144(%rsp) leaq 200(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8update_hiiiiiPfS_S_S_S_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_, .Lfunc_end0-_Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ .cfi_endproc # -- End function .globl _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ # -- Begin function _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ .p2align 4, 0x90 .type _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_,@function _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_: # @_Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ .cfi_startproc # %bb.0: subq $200, %rsp .cfi_def_cfa_offset 208 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movq %r9, 72(%rsp) leaq 20(%rsp), %rax movq %rax, 80(%rsp) leaq 16(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) leaq 72(%rsp), %rax movq %rax, 120(%rsp) leaq 208(%rsp), %rax movq %rax, 128(%rsp) leaq 216(%rsp), %rax movq %rax, 136(%rsp) leaq 224(%rsp), %rax movq %rax, 144(%rsp) leaq 232(%rsp), %rax movq %rax, 152(%rsp) leaq 240(%rsp), %rax movq %rax, 160(%rsp) leaq 248(%rsp), %rax movq %rax, 168(%rsp) leaq 256(%rsp), %rax movq %rax, 176(%rsp) leaq 264(%rsp), %rax movq %rax, 184(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $216, %rsp .cfi_adjust_cfa_offset -216 retq .Lfunc_end1: .size _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_, .Lfunc_end1-_Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ .cfi_endproc # -- End function .globl _Z25__device_stub__update_srciiiifPf # -- Begin function _Z25__device_stub__update_srciiiifPf .p2align 4, 0x90 .type _Z25__device_stub__update_srciiiifPf,@function _Z25__device_stub__update_srciiiifPf: # @_Z25__device_stub__update_srciiiifPf .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movss %xmm0, 4(%rsp) movq %r8, 72(%rsp) leaq 20(%rsp), %rax movq %rax, 80(%rsp) leaq 16(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) leaq 72(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z10update_srciiiifPf, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end2: .size _Z25__device_stub__update_srciiiifPf, .Lfunc_end2-_Z25__device_stub__update_srciiiifPf .cfi_endproc # -- End function .globl _Z24__device_stub__init_zeroiiPf # -- Begin function _Z24__device_stub__init_zeroiiPf .p2align 4, 0x90 .type _Z24__device_stub__init_zeroiiPf,@function _Z24__device_stub__init_zeroiiPf: # @_Z24__device_stub__init_zeroiiPf .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movl %edi, 4(%rsp) movl %esi, (%rsp) movq %rdx, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 64(%rsp) movq %rsp, %rax movq %rax, 72(%rsp) leaq 56(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z9init_zeroiiPf, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end3: .size _Z24__device_stub__init_zeroiiPf, .Lfunc_end3-_Z24__device_stub__init_zeroiiPf .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI4_0: .quad 0x4003c68000000000 # double 2.471923828125 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $360, %rsp # imm = 0x168 .cfi_def_cfa_offset 416 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $.Lstr, %edi callq puts@PLT xorl %r14d, %r14d movl $.L.str.1, %edi movl $320, %esi # imm = 0x140 movl $480, %edx # imm = 0x1E0 movl $480, %ecx # imm = 0x1E0 xorl %eax, %eax callq printf movsd .LCPI4_0(%rip), %xmm0 # xmm0 = mem[0],zero movl $.L.str.2, %edi movb $1, %al callq printf movl $.L.str.3, %edi movl $100000, %esi # imm = 0x186A0 xorl %eax, %eax callq printf movl $73728000, %edi # imm = 0x4650000 movl $4, %esi callq calloc movq %rax, 328(%rsp) # 8-byte Spill movl $73728000, %edi # imm = 0x4650000 movl $4, %esi callq calloc movq %rax, %rbx .p2align 4, 0x90 .LBB4_1: # =>This Inner Loop Header: Depth=1 movl $1056964608, (%rbx,%r14,4) # imm = 0x3F000000 incq %r14 cmpq $73728000, %r14 # imm = 0x4650000 jne .LBB4_1 # %bb.2: movabsq $4294967552, %r14 # imm = 0x100000100 leaq 128(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 120(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 112(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 104(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 96(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 80(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 320(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 312(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc leaq 304(%rsp), %rdi movl $294912000, %esi # imm = 0x11940000 callq hipMalloc movq 320(%rsp), %rdi movl $294912000, %edx # imm = 0x11940000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 312(%rsp), %rdi movl $294912000, %edx # imm = 0x11940000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 304(%rsp), %rdi movl $294912000, %edx # imm = 0x11940000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy xorl %r12d, %r12d leaq 47744(%r14), %r15 leaq 56(%rsp), %r13 leaq 40(%rsp), %rbx leaq 192(%rsp), %rbp jmp .LBB4_3 .p2align 4, 0x90 .LBB4_15: # in Loop: Header=BB4_3 Depth=1 addl $12288000, %r12d # imm = 0xBB8000 cmpl $73728000, %r12d # imm = 0x4650000 je .LBB4_16 .LBB4_3: # =>This Inner Loop Header: Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_5 # %bb.4: # in Loop: Header=BB4_3 Depth=1 movq 128(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB4_5: # in Loop: Header=BB4_3 Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_7 # %bb.6: # in Loop: Header=BB4_3 Depth=1 movq 120(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB4_7: # in Loop: Header=BB4_3 Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_9 # %bb.8: # in Loop: Header=BB4_3 Depth=1 movq 112(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB4_9: # in Loop: Header=BB4_3 Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_11 # %bb.10: # in Loop: Header=BB4_3 Depth=1 movq 104(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB4_11: # in Loop: Header=BB4_3 Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_13 # %bb.12: # in Loop: Header=BB4_3 Depth=1 movq 96(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB4_13: # in Loop: Header=BB4_3 Depth=1 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_15 # %bb.14: # in Loop: Header=BB4_3 Depth=1 movq 80(%rsp), %rax movl $73728000, 8(%rsp) # imm = 0x4650000 movl %r12d, (%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 32(%rsp), %rax movq %rax, 208(%rsp) movq %r13, %rdi movq %rbx, %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z9init_zeroiiPf, %edi movq %rbp, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB4_15 .LBB4_16: # %.preheader255 xorl %r13d, %r13d leaq 168(%rsp), %rbx leaq 160(%rsp), %rbp leaq 192(%rsp), %r12 leaq -255(%r14), %rax movq %rax, 344(%rsp) # 8-byte Spill leaq 256(%r14), %rax movq %rax, 336(%rsp) # 8-byte Spill jmp .LBB4_17 .p2align 4, 0x90 .LBB4_25: # in Loop: Header=BB4_17 Depth=1 movq 344(%rsp), %rdi # 8-byte Reload movl $1, %esi movq 336(%rsp), %rdx # 8-byte Reload movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax movq 352(%rsp), %r13 # 8-byte Reload je .LBB4_26 .LBB4_27: # in Loop: Header=BB4_17 Depth=1 incl %r13d cmpl $100000, %r13d # imm = 0x186A0 je .LBB4_28 .LBB4_17: # %.preheader254 # =>This Loop Header: Depth=1 # Child Loop BB4_18 Depth 2 # Child Loop BB4_22 Depth 2 movq %r13, 352(%rsp) # 8-byte Spill xorl %r13d, %r13d jmp .LBB4_18 .p2align 4, 0x90 .LBB4_20: # in Loop: Header=BB4_18 Depth=2 addl $12288000, %r13d # imm = 0xBB8000 cmpl $73728000, %r13d # imm = 0x4650000 je .LBB4_21 .LBB4_18: # Parent Loop BB4_17 Depth=1 # => This Inner Loop Header: Depth=2 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_20 # %bb.19: # in Loop: Header=BB4_18 Depth=2 movq 128(%rsp), %rax movq %rax, 32(%rsp) movq 120(%rsp), %rax movq %rax, 24(%rsp) movq 112(%rsp), %rax movq %rax, 16(%rsp) movq 104(%rsp), %rax movq %rax, 8(%rsp) movq 96(%rsp), %rax movq %rax, (%rsp) movq 80(%rsp), %rax movq %rax, 88(%rsp) movl $320, 152(%rsp) # imm = 0x140 movl $480, 144(%rsp) # imm = 0x1E0 movl $480, 136(%rsp) # imm = 0x1E0 movl $230400, 76(%rsp) # imm = 0x38400 movl %r13d, 72(%rsp) leaq 152(%rsp), %rax movq %rax, 192(%rsp) leaq 144(%rsp), %rax movq %rax, 200(%rsp) leaq 136(%rsp), %rax movq %rax, 208(%rsp) leaq 76(%rsp), %rax movq %rax, 216(%rsp) leaq 72(%rsp), %rax movq %rax, 224(%rsp) leaq 32(%rsp), %rax movq %rax, 232(%rsp) leaq 24(%rsp), %rax movq %rax, 240(%rsp) leaq 16(%rsp), %rax movq %rax, 248(%rsp) leaq 8(%rsp), %rax movq %rax, 256(%rsp) movq %rsp, %rax movq %rax, 264(%rsp) leaq 88(%rsp), %rax movq %rax, 272(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi movq %rbx, %rdx movq %rbp, %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z8update_hiiiiiPfS_S_S_S_S_, %edi movq %r12, %r9 pushq 160(%rsp) .cfi_adjust_cfa_offset 8 pushq 176(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB4_20 .p2align 4, 0x90 .LBB4_21: # %.preheader.preheader # in Loop: Header=BB4_17 Depth=1 xorl %r13d, %r13d jmp .LBB4_22 .p2align 4, 0x90 .LBB4_24: # in Loop: Header=BB4_22 Depth=2 addl $12288000, %r13d # imm = 0xBB8000 cmpl $73728000, %r13d # imm = 0x4650000 je .LBB4_25 .LBB4_22: # %.preheader # Parent Loop BB4_17 Depth=1 # => This Inner Loop Header: Depth=2 movq %r15, %rdi movl $1, %esi movq %r14, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB4_24 # %bb.23: # in Loop: Header=BB4_22 Depth=2 movq 128(%rsp), %rax movq %rax, 32(%rsp) movq 120(%rsp), %rax movq %rax, 24(%rsp) movq 112(%rsp), %rax movq %rax, 16(%rsp) movq 104(%rsp), %rax movq %rax, 8(%rsp) movq 96(%rsp), %rax movq %rax, (%rsp) movq 80(%rsp), %rax movq %rax, 88(%rsp) movq 320(%rsp), %rax movq %rax, 168(%rsp) movq 312(%rsp), %rax movq %rax, 160(%rsp) movq 304(%rsp), %rax movq %rax, 152(%rsp) movl $320, 76(%rsp) # imm = 0x140 movl $480, 72(%rsp) # imm = 0x1E0 movl $480, 188(%rsp) # imm = 0x1E0 movl $230400, 184(%rsp) # imm = 0x38400 movl %r13d, 180(%rsp) leaq 76(%rsp), %rax movq %rax, 192(%rsp) leaq 72(%rsp), %rax movq %rax, 200(%rsp) leaq 188(%rsp), %rax movq %rax, 208(%rsp) leaq 184(%rsp), %rax movq %rax, 216(%rsp) leaq 180(%rsp), %rax movq %rax, 224(%rsp) leaq 32(%rsp), %rax movq %rax, 232(%rsp) leaq 24(%rsp), %rax movq %rax, 240(%rsp) leaq 16(%rsp), %rax movq %rax, 248(%rsp) leaq 8(%rsp), %rax movq %rax, 256(%rsp) movq %rsp, %rax movq %rax, 264(%rsp) leaq 88(%rsp), %rax movq %rax, 272(%rsp) movq %rbx, 280(%rsp) movq %rbp, 288(%rsp) leaq 152(%rsp), %rax movq %rax, 296(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 144(%rsp), %rdx leaq 136(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, %edi movq %r12, %r9 pushq 136(%rsp) .cfi_adjust_cfa_offset 8 pushq 152(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB4_24 .p2align 4, 0x90 .LBB4_26: # in Loop: Header=BB4_17 Depth=1 cvtsi2ss %r13d, %xmm0 movq 80(%rsp), %rax movl $320, 8(%rsp) # imm = 0x140 movl $480, (%rsp) # imm = 0x1E0 movl $480, 88(%rsp) # imm = 0x1E0 movl $230400, 168(%rsp) # imm = 0x38400 movss %xmm0, 160(%rsp) movq %rax, 32(%rsp) leaq 8(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 88(%rsp), %rax movq %rax, 208(%rsp) movq %rbx, 216(%rsp) movq %rbp, 224(%rsp) leaq 32(%rsp), %rax movq %rax, 232(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z10update_srciiiifPf, %edi movq %r12, %r9 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB4_27 .LBB4_28: movq 80(%rsp), %rsi movl $294912000, %edx # imm = 0x11940000 movq 328(%rsp), %rdi # 8-byte Reload movl $2, %ecx callq hipMemcpy movl $.Lstr.1, %edi callq puts@PLT xorl %eax, %eax addq $360, %rsp # imm = 0x168 .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size main, .Lfunc_end4-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8update_hiiiiiPfS_S_S_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z10update_srciiiifPf, %esi movl $.L__unnamed_3, %edx movl $.L__unnamed_3, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z9init_zeroiiPf, %esi movl $.L__unnamed_4, %edx movl $.L__unnamed_4, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end5: .size __hip_module_ctor, .Lfunc_end5-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB6_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB6_2: retq .Lfunc_end6: .size __hip_module_dtor, .Lfunc_end6-__hip_module_dtor .cfi_endproc # -- End function .type _Z8update_hiiiiiPfS_S_S_S_S_,@object # @_Z8update_hiiiiiPfS_S_S_S_S_ .section .rodata,"a",@progbits .globl _Z8update_hiiiiiPfS_S_S_S_S_ .p2align 3, 0x0 _Z8update_hiiiiiPfS_S_S_S_S_: .quad _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ .size _Z8update_hiiiiiPfS_S_S_S_S_, 8 .type _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_,@object # @_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_ .globl _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_ .p2align 3, 0x0 _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_: .quad _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ .size _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_, 8 .type _Z10update_srciiiifPf,@object # @_Z10update_srciiiifPf .globl _Z10update_srciiiifPf .p2align 3, 0x0 _Z10update_srciiiifPf: .quad _Z25__device_stub__update_srciiiifPf .size _Z10update_srciiiifPf, 8 .type _Z9init_zeroiiPf,@object # @_Z9init_zeroiiPf .globl _Z9init_zeroiiPf .p2align 3, 0x0 _Z9init_zeroiiPf: .quad _Z24__device_stub__init_zeroiiPf .size _Z9init_zeroiiPf, 8 .type .L.str.1,@object # @.str.1 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.1: .asciz "Array size : %dx%dx%d\n" .size .L.str.1, 23 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Total used memory : %1.2f GB\n" .size .L.str.2, 30 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Iteration : %d step\n" .size .L.str.3, 21 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z8update_hiiiiiPfS_S_S_S_S_" .size .L__unnamed_1, 29 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z8update_eiiiiiPfS_S_S_S_S_S_S_S_" .size .L__unnamed_2, 35 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z10update_srciiiifPf" .size .L__unnamed_3, 22 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z9init_zeroiiPf" .size .L__unnamed_4, 17 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Simple FDTD simulation" .size .Lstr, 23 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Complete." .size .Lstr.1, 10 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__update_hiiiiiPfS_S_S_S_S_ .addrsig_sym _Z23__device_stub__update_eiiiiiPfS_S_S_S_S_S_S_S_ .addrsig_sym _Z25__device_stub__update_srciiiifPf .addrsig_sym _Z24__device_stub__init_zeroiiPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8update_hiiiiiPfS_S_S_S_S_ .addrsig_sym _Z8update_eiiiiiPfS_S_S_S_S_S_S_S_ .addrsig_sym _Z10update_srciiiifPf .addrsig_sym _Z9init_zeroiiPf .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
// Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; cudaSetDevice(2); cudaMalloc((void**)&d_A, m*k * sizeof(double)); cudaMalloc((void**)&d_B, k*n * sizeof(double)); cudaMalloc((void**)&d_C, m*n * sizeof(double)); cudaMemcpy(d_A, A, m*k * sizeof(double), cudaMemcpyHostToDevice); cudaMemcpy(d_B, B, k*n * sizeof(double), cudaMemcpyHostToDevice); // Initialize the output matrix with zeroes. cudaMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); cudaDeviceSynchronize(); cudaMemcpy(C, d_C, m*n * sizeof(double), cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } }
code for sm_80 Function : _Z2m1iiiPdS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ IMAD.MOV.U32 R0, RZ, RZ, 0x1 ; /* 0x00000001ff007424 */ /* 0x000fca00078e00ff */ /*0020*/ ISETP.LE.AND P0, PT, R0, c[0x0][0x160], PT ; /* 0x0000580000007a0c */ /* 0x000fda0003f03270 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ ISETP.LE.AND P0, PT, R0, c[0x0][0x168], PT ; /* 0x00005a0000007a0c */ /* 0x000fda0003f03270 */ /*0050*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0060*/ IMAD.MOV.U32 R0, RZ, RZ, 0x1 ; /* 0x00000001ff007424 */ /* 0x000fe200078e00ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0080*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff027624 */ /* 0x000fc600078e00ff */ /*0090*/ IADD3 R0, -R0, c[0x0][0x164], RZ ; /* 0x0000590000007a10 */ /* 0x000fc80007ffe1ff */ /*00a0*/ ISETP.GE.U32.AND P1, PT, R0, 0x3, PT ; /* 0x000000030000780c */ /* 0x000fe40003f26070 */ /*00b0*/ LOP3.LUT R0, R2, 0x3, RZ, 0xc0, !PT ; /* 0x0000000302007812 */ /* 0x000fe200078ec0ff */ /*00c0*/ IMAD.MOV.U32 R2, RZ, RZ, RZ ; /* 0x000000ffff027224 */ /* 0x000fc600078e00ff */ /*00d0*/ IADD3 R3, -R0, c[0x0][0x164], RZ ; /* 0x0000590000037a10 */ /* 0x000fe40007ffe1ff */ /*00e0*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff047624 */ /* 0x000fca00078e00ff */ /*00f0*/ ISETP.GE.AND P0, PT, R4, 0x1, PT ; /* 0x000000010400780c */ /* 0x000fda0003f06270 */ /*0100*/ @!P0 BRA 0xf00 ; /* 0x00000df000008947 */ /* 0x003fea0003800000 */ /*0110*/ IMAD R4, R2, c[0x0][0x164], RZ ; /* 0x0000590002047a24 */ /* 0x000fe400078e02ff */ /*0120*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc800078e00ff */ /*0130*/ IMAD R10, R2, c[0x0][0x168], R5 ; /* 0x00005a00020a7a24 */ /* 0x000fe400078e0205 */ /*0140*/ IMAD R6, R5.reuse, c[0x0][0x164], RZ ; /* 0x0000590005067a24 */ /* 0x041fe200078e02ff */ /*0150*/ IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105057810 */ /* 0x000fe20007ffe0ff */ /*0160*/ IMAD.MOV.U32 R7, RZ, RZ, 0x8 ; /* 0x00000008ff077424 */ /* 0x000fe400078e00ff */ /*0170*/ IMAD.MOV.U32 R9, RZ, RZ, RZ ; /* 0x000000ffff097224 */ /* 0x000fe200078e00ff */ /*0180*/ ISETP.GE.AND P2, PT, R5, c[0x0][0x168], PT ; /* 0x00005a0005007a0c */ /* 0x000fe20003f46270 */ /*0190*/ IMAD.WIDE R10, R10, R7, c[0x0][0x170] ; /* 0x00005c000a0a7625 */ /* 0x000fe200078e0207 */ /*01a0*/ @!P1 BRA 0xd60 ; /* 0x00000bb000009947 */ /* 0x002fea0003800000 */ /*01b0*/ ISETP.GT.AND P0, PT, R3, RZ, PT ; /* 0x000000ff0300720c */ /* 0x000fe20003f04270 */ /*01c0*/ IMAD.MOV.U32 R9, RZ, RZ, RZ ; /* 0x000000ffff097224 */ /* 0x000fc400078e00ff */ /*01d0*/ IMAD.MOV.U32 R22, RZ, RZ, c[0x0][0x180] ; /* 0x00006000ff167624 */ /* 0x000fe400078e00ff */ /*01e0*/ IMAD.MOV.U32 R23, RZ, RZ, c[0x0][0x184] ; /* 0x00006100ff177624 */ /* 0x000fe400078e00ff */ /*01f0*/ IMAD.MOV.U32 R8, RZ, RZ, R3 ; /* 0x000000ffff087224 */ /* 0x000fe400078e0003 */ /*0200*/ IMAD.WIDE R12, R6, R7, c[0x0][0x178] ; /* 0x00005e00060c7625 */ /* 0x000fc800078e0207 */ /*0210*/ @!P0 BRA 0xb70 ; /* 0x0000095000008947 */ /* 0x000fea0003800000 */ /*0220*/ ISETP.GT.AND P3, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f64270 */ /*0230*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*0240*/ @!P3 BRA 0x810 ; /* 0x000005c00000b947 */ /* 0x000fea0003800000 */ /*0250*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0260*/ IMAD.WIDE R20, R4, 0x8, R22 ; /* 0x0000000804147825 */ /* 0x001fe200078e0216 */ /*0270*/ LDG.E.64 R18, [R12.64] ; /* 0x000000040c127981 */ /* 0x000ea8000c1e1b00 */ /*0280*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea8000c1e1b00 */ /*0290*/ LDG.E.64 R14, [R20.64] ; /* 0x00000004140e7981 */ /* 0x000ea4000c1e1b00 */ /*02a0*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x004044000000000e */ /*02b0*/ LDG.E.64 R14, [R20.64+0x8] ; /* 0x00000804140e7981 */ /* 0x001eaa000c1e1b00 */ /*02c0*/ STG.E.64 [R20.64], R28 ; /* 0x0000001c14007986 */ /* 0x0021e8000c101b04 */ /*02d0*/ LDG.E.64 R18, [R12.64+0x8] ; /* 0x000008040c127981 */ /* 0x000ea8000c1e1b00 */ /*02e0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*02f0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x004284000000000e */ /*0300*/ LDG.E.64 R14, [R20.64+0x10] ; /* 0x00001004140e7981 */ /* 0x002eea000c1e1b00 */ /*0310*/ STG.E.64 [R20.64+0x8], R26 ; /* 0x0000081a14007986 */ /* 0x0043e8000c101b04 */ /*0320*/ LDG.E.64 R18, [R12.64+0x10] ; /* 0x000010040c127981 */ /* 0x000ee8000c1e1b00 */ /*0330*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ee4000c1e1b00 */ /*0340*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0084c4000000000e */ /*0350*/ LDG.E.64 R14, [R20.64+0x18] ; /* 0x00001804140e7981 */ /* 0x004e2a000c1e1b00 */ /*0360*/ STG.E.64 [R20.64+0x10], R24 ; /* 0x0000101814007986 */ /* 0x0085e8000c101b04 */ /*0370*/ LDG.E.64 R18, [R12.64+0x18] ; /* 0x000018040c127981 */ /* 0x000e28000c1e1b00 */ /*0380*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0390*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*03a0*/ LDG.E.64 R14, [R20.64+0x20] ; /* 0x00002004140e7981 */ /* 0x001e6a000c1e1b00 */ /*03b0*/ STG.E.64 [R20.64+0x18], R28 ; /* 0x0000181c14007986 */ /* 0x0081e8000c101b04 */ /*03c0*/ LDG.E.64 R18, [R12.64+0x20] ; /* 0x000020040c127981 */ /* 0x000e68000c1e1b00 */ /*03d0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*03e0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*03f0*/ LDG.E.64 R14, [R20.64+0x28] ; /* 0x00002804140e7981 */ /* 0x002eaa000c1e1b00 */ /*0400*/ STG.E.64 [R20.64+0x20], R26 ; /* 0x0000201a14007986 */ /* 0x0083e8000c101b04 */ /*0410*/ LDG.E.64 R18, [R12.64+0x28] ; /* 0x000028040c127981 */ /* 0x000ea8000c1e1b00 */ /*0420*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0430*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0440*/ LDG.E.64 R14, [R20.64+0x30] ; /* 0x00003004140e7981 */ /* 0x004e2a000c1e1b00 */ /*0450*/ STG.E.64 [R20.64+0x28], R24 ; /* 0x0000281814007986 */ /* 0x0085e8000c101b04 */ /*0460*/ LDG.E.64 R18, [R12.64+0x30] ; /* 0x000030040c127981 */ /* 0x000e28000c1e1b00 */ /*0470*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0480*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0490*/ LDG.E.64 R14, [R20.64+0x38] ; /* 0x00003804140e7981 */ /* 0x001e6a000c1e1b00 */ /*04a0*/ STG.E.64 [R20.64+0x30], R28 ; /* 0x0000301c14007986 */ /* 0x0081e8000c101b04 */ /*04b0*/ LDG.E.64 R18, [R12.64+0x38] ; /* 0x000038040c127981 */ /* 0x000e68000c1e1b00 */ /*04c0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*04d0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*04e0*/ LDG.E.64 R14, [R20.64+0x40] ; /* 0x00004004140e7981 */ /* 0x002eaa000c1e1b00 */ /*04f0*/ STG.E.64 [R20.64+0x38], R26 ; /* 0x0000381a14007986 */ /* 0x0083e8000c101b04 */ /*0500*/ LDG.E.64 R18, [R12.64+0x40] ; /* 0x000040040c127981 */ /* 0x000ea8000c1e1b00 */ /*0510*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0520*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0530*/ LDG.E.64 R14, [R20.64+0x48] ; /* 0x00004804140e7981 */ /* 0x004e2a000c1e1b00 */ /*0540*/ STG.E.64 [R20.64+0x40], R24 ; /* 0x0000401814007986 */ /* 0x0085e8000c101b04 */ /*0550*/ LDG.E.64 R18, [R12.64+0x48] ; /* 0x000048040c127981 */ /* 0x000e28000c1e1b00 */ /*0560*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0570*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0580*/ LDG.E.64 R14, [R20.64+0x50] ; /* 0x00005004140e7981 */ /* 0x001e6a000c1e1b00 */ /*0590*/ STG.E.64 [R20.64+0x48], R28 ; /* 0x0000481c14007986 */ /* 0x0081e8000c101b04 */ /*05a0*/ LDG.E.64 R18, [R12.64+0x50] ; /* 0x000050040c127981 */ /* 0x000e68000c1e1b00 */ /*05b0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*05c0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*05d0*/ LDG.E.64 R14, [R20.64+0x58] ; /* 0x00005804140e7981 */ /* 0x002eaa000c1e1b00 */ /*05e0*/ STG.E.64 [R20.64+0x50], R26 ; /* 0x0000501a14007986 */ /* 0x0083e8000c101b04 */ /*05f0*/ LDG.E.64 R18, [R12.64+0x58] ; /* 0x000058040c127981 */ /* 0x000ea8000c1e1b00 */ /*0600*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0610*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0620*/ LDG.E.64 R14, [R20.64+0x60] ; /* 0x00006004140e7981 */ /* 0x004e2a000c1e1b00 */ /*0630*/ STG.E.64 [R20.64+0x58], R24 ; /* 0x0000581814007986 */ /* 0x0085e8000c101b04 */ /*0640*/ LDG.E.64 R18, [R12.64+0x60] ; /* 0x000060040c127981 */ /* 0x000e28000c1e1b00 */ /*0650*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0660*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0670*/ LDG.E.64 R14, [R20.64+0x68] ; /* 0x00006804140e7981 */ /* 0x001e6a000c1e1b00 */ /*0680*/ STG.E.64 [R20.64+0x60], R28 ; /* 0x0000601c14007986 */ /* 0x0081e8000c101b04 */ /*0690*/ LDG.E.64 R18, [R12.64+0x68] ; /* 0x000068040c127981 */ /* 0x000e68000c1e1b00 */ /*06a0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*06b0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*06c0*/ LDG.E.64 R14, [R20.64+0x70] ; /* 0x00007004140e7981 */ /* 0x002eaa000c1e1b00 */ /*06d0*/ STG.E.64 [R20.64+0x68], R26 ; /* 0x0000681a14007986 */ /* 0x0081e8000c101b04 */ /*06e0*/ LDG.E.64 R18, [R12.64+0x70] ; /* 0x000070040c127981 */ /* 0x000ea8000c1e1b00 */ /*06f0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea2000c1e1b00 */ /*0700*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fe20007ffe0ff */ /*0710*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x004284000000000e */ /*0720*/ LDG.E.64 R14, [R20.64+0x78] ; /* 0x00007804140e7981 */ /* 0x002eea000c1e1b00 */ /*0730*/ STG.E.64 [R20.64+0x70], R24 ; /* 0x0000701814007986 */ /* 0x0041e8000c101b04 */ /*0740*/ LDG.E.64 R18, [R12.64+0x78] ; /* 0x000078040c127981 */ /* 0x000ee8000c1e1b00 */ /*0750*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ee2000c1e1b00 */ /*0760*/ ISETP.GT.AND P3, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fc40003f64270 */ /*0770*/ IADD3 R22, P4, R22, 0x80, RZ ; /* 0x0000008016167810 */ /* 0x000fe40007f9e0ff */ /*0780*/ IADD3 R9, R9, 0x10, RZ ; /* 0x0000001009097810 */ /* 0x000fc60007ffe0ff */ /*0790*/ IMAD.X R23, RZ, RZ, R23, P4 ; /* 0x000000ffff177224 */ /* 0x000fe200020e0617 */ /*07a0*/ DFMA R14, R18, R16, R14 ; /* 0x00000010120e722b */ /* 0x0082a4000000000e */ /*07b0*/ IADD3 R16, P5, R12, 0x80, RZ ; /* 0x000000800c107810 */ /* 0x002fc80007fbe0ff */ /*07c0*/ IADD3.X R17, RZ, R13, RZ, P5, !PT ; /* 0x0000000dff117210 */ /* 0x000fe20002ffe4ff */ /*07d0*/ STG.E.64 [R20.64+0x78], R14 ; /* 0x0000780e14007986 */ /* 0x0041e2000c101b04 */ /*07e0*/ IMAD.MOV.U32 R12, RZ, RZ, R16 ; /* 0x000000ffff0c7224 */ /* 0x000fc600078e0010 */ /*07f0*/ IMAD.MOV.U32 R13, RZ, RZ, R17 ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e0011 */ /*0800*/ @P3 BRA 0x260 ; /* 0xfffffa5000003947 */ /* 0x000fea000383ffff */ /*0810*/ ISETP.GT.AND P3, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f64270 */ /*0820*/ @!P3 BRA 0xb50 ; /* 0x000003200000b947 */ /* 0x000fea0003800000 */ /*0830*/ IMAD.WIDE R14, R4, 0x8, R22 ; /* 0x00000008040e7825 */ /* 0x001fe200078e0216 */ /*0840*/ LDG.E.64 R20, [R12.64] ; /* 0x000000040c147981 */ /* 0x000ea8000c1e1b00 */ /*0850*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea8000c1e1b00 */ /*0860*/ LDG.E.64 R16, [R14.64] ; /* 0x000000040e107981 */ /* 0x000ea4000c1e1b00 */ /*0870*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0040440000000010 */ /*0880*/ LDG.E.64 R16, [R14.64+0x8] ; /* 0x000008040e107981 */ /* 0x001eaa000c1e1b00 */ /*0890*/ STG.E.64 [R14.64], R24 ; /* 0x000000180e007986 */ /* 0x0021e8000c101b04 */ /*08a0*/ LDG.E.64 R20, [R12.64+0x8] ; /* 0x000008040c147981 */ /* 0x000ea8000c1e1b00 */ /*08b0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*08c0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0042840000000010 */ /*08d0*/ LDG.E.64 R16, [R14.64+0x10] ; /* 0x000010040e107981 */ /* 0x002eea000c1e1b00 */ /*08e0*/ STG.E.64 [R14.64+0x8], R28 ; /* 0x0000081c0e007986 */ /* 0x0043e8000c101b04 */ /*08f0*/ LDG.E.64 R20, [R12.64+0x10] ; /* 0x000010040c147981 */ /* 0x000ee8000c1e1b00 */ /*0900*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee4000c1e1b00 */ /*0910*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0084c40000000010 */ /*0920*/ LDG.E.64 R16, [R14.64+0x18] ; /* 0x000018040e107981 */ /* 0x004e2a000c1e1b00 */ /*0930*/ STG.E.64 [R14.64+0x10], R26 ; /* 0x0000101a0e007986 */ /* 0x0085e8000c101b04 */ /*0940*/ LDG.E.64 R20, [R12.64+0x18] ; /* 0x000018040c147981 */ /* 0x000e28000c1e1b00 */ /*0950*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000e24000c1e1b00 */ /*0960*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0010c40000000010 */ /*0970*/ LDG.E.64 R16, [R14.64+0x20] ; /* 0x000020040e107981 */ /* 0x001e6a000c1e1b00 */ /*0980*/ STG.E.64 [R14.64+0x18], R24 ; /* 0x000018180e007986 */ /* 0x0081e8000c101b04 */ /*0990*/ LDG.E.64 R20, [R12.64+0x20] ; /* 0x000020040c147981 */ /* 0x000e68000c1e1b00 */ /*09a0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000e64000c1e1b00 */ /*09b0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0022c40000000010 */ /*09c0*/ LDG.E.64 R16, [R14.64+0x28] ; /* 0x000028040e107981 */ /* 0x002eaa000c1e1b00 */ /*09d0*/ STG.E.64 [R14.64+0x20], R28 ; /* 0x0000201c0e007986 */ /* 0x0083e8000c101b04 */ /*09e0*/ LDG.E.64 R20, [R12.64+0x28] ; /* 0x000028040c147981 */ /* 0x000ea8000c1e1b00 */ /*09f0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*0a00*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0044c40000000010 */ /*0a10*/ LDG.E.64 R16, [R14.64+0x30] ; /* 0x000030040e107981 */ /* 0x004e2a000c1e1b00 */ /*0a20*/ STG.E.64 [R14.64+0x28], R26 ; /* 0x0000281a0e007986 */ /* 0x0083e8000c101b04 */ /*0a30*/ LDG.E.64 R18, [R12.64+0x30] ; /* 0x000030040c127981 */ /* 0x000e28000c1e1b00 */ /*0a40*/ LDG.E.64 R20, [R10.64] ; /* 0x000000040a147981 */ /* 0x000e24000c1e1b00 */ /*0a50*/ DFMA R24, R18, R20, R16 ; /* 0x000000141218722b */ /* 0x0010840000000010 */ /*0a60*/ LDG.E.64 R16, [R14.64+0x38] ; /* 0x000038040e107981 */ /* 0x001eea000c1e1b00 */ /*0a70*/ STG.E.64 [R14.64+0x30], R24 ; /* 0x000030180e007986 */ /* 0x0043e8000c101b04 */ /*0a80*/ LDG.E.64 R18, [R12.64+0x38] ; /* 0x000038040c127981 */ /* 0x000ee8000c1e1b00 */ /*0a90*/ LDG.E.64 R20, [R10.64] ; /* 0x000000040a147981 */ /* 0x000ee2000c1e1b00 */ /*0aa0*/ IADD3 R22, P3, R22, 0x40, RZ ; /* 0x0000004016167810 */ /* 0x000fc40007f7e0ff */ /*0ab0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0ac0*/ IADD3 R9, R9, 0x8, RZ ; /* 0x0000000809097810 */ /* 0x000fe20007ffe0ff */ /*0ad0*/ IMAD.X R23, RZ, RZ, R23, P3 ; /* 0x000000ffff177224 */ /* 0x000fe200018e0617 */ /*0ae0*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe20007ffe0ff */ /*0af0*/ DFMA R16, R18, R20, R16 ; /* 0x000000141210722b */ /* 0x0080a40000000010 */ /*0b00*/ IADD3 R18, P4, R12, 0x40, RZ ; /* 0x000000400c127810 */ /* 0x001fca0007f9e0ff */ /*0b10*/ STG.E.64 [R14.64+0x38], R16 ; /* 0x000038100e007986 */ /* 0x0043e2000c101b04 */ /*0b20*/ IMAD.X R19, RZ, RZ, R13, P4 ; /* 0x000000ffff137224 */ /* 0x000fe400020e060d */ /*0b30*/ IMAD.MOV.U32 R12, RZ, RZ, R18 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0012 */ /*0b40*/ IMAD.MOV.U32 R13, RZ, RZ, R19 ; /* 0x000000ffff0d7224 */ /* 0x000fe400078e0013 */ /*0b50*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0b60*/ @!P0 BRA 0xd60 ; /* 0x000001f000008947 */ /* 0x000fea0003800000 */ /*0b70*/ IMAD.WIDE R14, R4, 0x8, R22 ; /* 0x00000008040e7825 */ /* 0x003fe200078e0216 */ /*0b80*/ LDG.E.64 R20, [R12.64] ; /* 0x000000040c147981 */ /* 0x000ea8000c1e1b00 */ /*0b90*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea8000c1e1b00 */ /*0ba0*/ LDG.E.64 R16, [R14.64] ; /* 0x000000040e107981 */ /* 0x000ea4000c1e1b00 */ /*0bb0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0040440000000010 */ /*0bc0*/ LDG.E.64 R16, [R14.64+0x8] ; /* 0x000008040e107981 */ /* 0x001eaa000c1e1b00 */ /*0bd0*/ STG.E.64 [R14.64], R28 ; /* 0x0000001c0e007986 */ /* 0x0021e8000c101b04 */ /*0be0*/ LDG.E.64 R20, [R12.64+0x8] ; /* 0x000008040c147981 */ /* 0x000ea8000c1e1b00 */ /*0bf0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*0c00*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0042840000000010 */ /*0c10*/ LDG.E.64 R16, [R14.64+0x10] ; /* 0x000010040e107981 */ /* 0x002eea000c1e1b00 */ /*0c20*/ STG.E.64 [R14.64+0x8], R26 ; /* 0x0000081a0e007986 */ /* 0x0041e8000c101b04 */ /*0c30*/ LDG.E.64 R20, [R12.64+0x10] ; /* 0x000010040c147981 */ /* 0x000ee8000c1e1b00 */ /*0c40*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee2000c1e1b00 */ /*0c50*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fe20007ffe0ff */ /*0c60*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0082840000000010 */ /*0c70*/ LDG.E.64 R16, [R14.64+0x18] ; /* 0x000018040e107981 */ /* 0x002eea000c1e1b00 */ /*0c80*/ STG.E.64 [R14.64+0x10], R24 ; /* 0x000010180e007986 */ /* 0x0041e8000c101b04 */ /*0c90*/ LDG.E.64 R20, [R12.64+0x18] ; /* 0x000018040c147981 */ /* 0x000ee8000c1e1b00 */ /*0ca0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee2000c1e1b00 */ /*0cb0*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fc40003f05270 */ /*0cc0*/ IADD3 R22, P3, R22, 0x20, RZ ; /* 0x0000002016167810 */ /* 0x000fe40007f7e0ff */ /*0cd0*/ IADD3 R9, R9, 0x4, RZ ; /* 0x0000000409097810 */ /* 0x000fc60007ffe0ff */ /*0ce0*/ IMAD.X R23, RZ, RZ, R23, P3 ; /* 0x000000ffff177224 */ /* 0x000fe200018e0617 */ /*0cf0*/ DFMA R16, R20, R18, R16 ; /* 0x000000121410722b */ /* 0x0082a40000000010 */ /*0d00*/ IADD3 R18, P4, R12, 0x20, RZ ; /* 0x000000200c127810 */ /* 0x002fc80007f9e0ff */ /*0d10*/ IADD3.X R19, RZ, R13, RZ, P4, !PT ; /* 0x0000000dff137210 */ /* 0x000fe200027fe4ff */ /*0d20*/ STG.E.64 [R14.64+0x18], R16 ; /* 0x000018100e007986 */ /* 0x0041e2000c101b04 */ /*0d30*/ IMAD.MOV.U32 R12, RZ, RZ, R18 ; /* 0x000000ffff0c7224 */ /* 0x000fc600078e0012 */ /*0d40*/ IMAD.MOV.U32 R13, RZ, RZ, R19 ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e0013 */ /*0d50*/ @P0 BRA 0xb70 ; /* 0xfffffe1000000947 */ /* 0x001fea000383ffff */ /*0d60*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f05270 */ /*0d70*/ @!P0 BRA 0xef0 ; /* 0x0000017000008947 */ /* 0x000fea0003800000 */ /*0d80*/ IMAD.IADD R8, R6, 0x1, R9.reuse ; /* 0x0000000106087824 */ /* 0x100fe200078e0209 */ /*0d90*/ LDG.E.64 R14, [R10.64] ; /* 0x000000040a0e7981 */ /* 0x003ea2000c1e1b00 */ /*0da0*/ IMAD.IADD R6, R4, 0x1, R9 ; /* 0x0000000104067824 */ /* 0x000fe400078e0209 */ /*0db0*/ IMAD.WIDE R8, R8, R7, c[0x0][0x178] ; /* 0x00005e0008087625 */ /* 0x000fc800078e0207 */ /*0dc0*/ IMAD.WIDE R6, R6, R7, c[0x0][0x180] ; /* 0x0000600006067625 */ /* 0x000fe200078e0207 */ /*0dd0*/ LDG.E.64 R12, [R8.64] ; /* 0x00000004080c7981 */ /* 0x000ea8000c1e1b00 */ /*0de0*/ LDG.E.64 R16, [R6.64] ; /* 0x0000000406107981 */ /* 0x000ea2000c1e1b00 */ /*0df0*/ ISETP.NE.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fe20003f05270 */ /*0e00*/ DFMA R12, R12, R14, R16 ; /* 0x0000000e0c0c722b */ /* 0x004e0e0000000010 */ /*0e10*/ STG.E.64 [R6.64], R12 ; /* 0x0000000c06007986 */ /* 0x0011ea000c101b04 */ /*0e20*/ @!P0 BRA 0xef0 ; /* 0x000000c000008947 */ /* 0x000fea0003800000 */ /*0e30*/ LDG.E.64 R12, [R8.64+0x8] ; /* 0x00000804080c7981 */ /* 0x001ea8000c1e1b00 */ /*0e40*/ LDG.E.64 R14, [R10.64] ; /* 0x000000040a0e7981 */ /* 0x000ea8000c1e1b00 */ /*0e50*/ LDG.E.64 R16, [R6.64+0x8] ; /* 0x0000080406107981 */ /* 0x000ea2000c1e1b00 */ /*0e60*/ ISETP.NE.AND P0, PT, R0, 0x2, PT ; /* 0x000000020000780c */ /* 0x000fe20003f05270 */ /*0e70*/ DFMA R12, R12, R14, R16 ; /* 0x0000000e0c0c722b */ /* 0x004e0e0000000010 */ /*0e80*/ STG.E.64 [R6.64+0x8], R12 ; /* 0x0000080c06007986 */ /* 0x0011ea000c101b04 */ /*0e90*/ @!P0 BRA 0xef0 ; /* 0x0000005000008947 */ /* 0x000fea0003800000 */ /*0ea0*/ LDG.E.64 R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ea8000c1e1b00 */ /*0eb0*/ LDG.E.64 R8, [R8.64+0x10] ; /* 0x0000100408087981 */ /* 0x000ea8000c1e1b00 */ /*0ec0*/ LDG.E.64 R12, [R6.64+0x10] ; /* 0x00001004060c7981 */ /* 0x001ea4000c1e1b00 */ /*0ed0*/ DFMA R12, R10, R8, R12 ; /* 0x000000080a0c722b */ /* 0x004e0e000000000c */ /*0ee0*/ STG.E.64 [R6.64+0x10], R12 ; /* 0x0000100c06007986 */ /* 0x0011e8000c101b04 */ /*0ef0*/ @!P2 BRA 0x130 ; /* 0xfffff2300000a947 */ /* 0x000fea000383ffff */ /*0f00*/ IADD3 R2, R2, 0x1, RZ ; /* 0x0000000102027810 */ /* 0x000fc80007ffe0ff */ /*0f10*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x160], PT ; /* 0x0000580002007a0c */ /* 0x000fda0003f06270 */ /*0f20*/ @P0 CALL.REL.NOINC 0xf40 ; /* 0x0000001000000944 */ /* 0x000fe20003c00000 */ /*0f30*/ BRA 0xe0 ; /* 0xfffff1a000007947 */ /* 0x000fea000383ffff */ /*0f40*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0f50*/ BRA 0xf50; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0f60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fe0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ff0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
// Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; cudaSetDevice(2); cudaMalloc((void**)&d_A, m*k * sizeof(double)); cudaMalloc((void**)&d_B, k*n * sizeof(double)); cudaMalloc((void**)&d_C, m*n * sizeof(double)); cudaMemcpy(d_A, A, m*k * sizeof(double), cudaMemcpyHostToDevice); cudaMemcpy(d_B, B, k*n * sizeof(double), cudaMemcpyHostToDevice); // Initialize the output matrix with zeroes. cudaMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); cudaDeviceSynchronize(); cudaMemcpy(C, d_C, m*n * sizeof(double), cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } }
.file "tmpxft_000b0dc7_00000000-6_matmult_gpu1.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ .type _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_, @function _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_: .LFB2052: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 44(%rsp) movl %esi, 40(%rsp) movl %edx, 36(%rsp) movq %rcx, 24(%rsp) movq %r8, 16(%rsp) movq %r9, 8(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 44(%rsp), %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rax movq %rax, 120(%rsp) leaq 36(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z2m1iiiPdS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2052: .size _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_, .-_Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ .globl _Z2m1iiiPdS_S_ .type _Z2m1iiiPdS_S_, @function _Z2m1iiiPdS_S_: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _Z2m1iiiPdS_S_, .-_Z2m1iiiPdS_S_ .globl matmult_gpu1 .type matmult_gpu1, @function matmult_gpu1: .LFB2027: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $104, %rsp .cfi_def_cfa_offset 160 movl %edi, %r13d movl %esi, %r14d movl %edx, %r15d movq %rcx, 8(%rsp) movq %r8, 16(%rsp) movq %r9, 24(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax movl $2, %edi call cudaSetDevice@PLT movl %r13d, %r12d imull %r15d, %r12d movslq %r12d, %r12 salq $3, %r12 leaq 40(%rsp), %rdi movq %r12, %rsi call cudaMalloc@PLT movl %r15d, %ebp imull %r14d, %ebp movslq %ebp, %rbp salq $3, %rbp leaq 48(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl %r13d, %ebx imull %r14d, %ebx movslq %ebx, %rbx salq $3, %rbx leaq 56(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r12, %rdx movq 8(%rsp), %rsi movq 40(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbp, %rdx movq 16(%rsp), %rsi movq 48(%rsp), %rdi call cudaMemcpy@PLT movq %rbx, %rdx movl $0, %esi movq 56(%rsp), %rdi call cudaMemset@PLT movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $0, %r9d movl $0, %r8d movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L15 .L12: call cudaDeviceSynchronize@PLT movl $2, %ecx movq %rbx, %rdx movq 56(%rsp), %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movq 40(%rsp), %rdi call cudaFree@PLT movq 48(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rdi call cudaFree@PLT movq 88(%rsp), %rax subq %fs:40, %rax jne .L16 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state movq 56(%rsp), %r9 movq 48(%rsp), %r8 movq 40(%rsp), %rcx movl %r15d, %edx movl %r14d, %esi movl %r13d, %edi call _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2027: .size matmult_gpu1, .-matmult_gpu1 .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z2m1iiiPdS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2055: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z2m1iiiPdS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2055: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
// Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; cudaSetDevice(2); cudaMalloc((void**)&d_A, m*k * sizeof(double)); cudaMalloc((void**)&d_B, k*n * sizeof(double)); cudaMalloc((void**)&d_C, m*n * sizeof(double)); cudaMemcpy(d_A, A, m*k * sizeof(double), cudaMemcpyHostToDevice); cudaMemcpy(d_B, B, k*n * sizeof(double), cudaMemcpyHostToDevice); // Initialize the output matrix with zeroes. cudaMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); cudaDeviceSynchronize(); cudaMemcpy(C, d_C, m*n * sizeof(double), cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } }
#include <hip/hip_runtime.h> // Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; hipSetDevice(2); hipMalloc((void**)&d_A, m*k * sizeof(double)); hipMalloc((void**)&d_B, k*n * sizeof(double)); hipMalloc((void**)&d_C, m*n * sizeof(double)); hipMemcpy(d_A, A, m*k * sizeof(double), hipMemcpyHostToDevice); hipMemcpy(d_B, B, k*n * sizeof(double), hipMemcpyHostToDevice); // Initialize the output matrix with zeroes. hipMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); hipDeviceSynchronize(); hipMemcpy(C, d_C, m*n * sizeof(double), hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> // Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; hipSetDevice(2); hipMalloc((void**)&d_A, m*k * sizeof(double)); hipMalloc((void**)&d_B, k*n * sizeof(double)); hipMalloc((void**)&d_C, m*n * sizeof(double)); hipMemcpy(d_A, A, m*k * sizeof(double), hipMemcpyHostToDevice); hipMemcpy(d_B, B, k*n * sizeof(double), hipMemcpyHostToDevice); // Initialize the output matrix with zeroes. hipMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); hipDeviceSynchronize(); hipMemcpy(C, d_C, m*n * sizeof(double), hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z2m1iiiPdS_S_ .globl _Z2m1iiiPdS_S_ .p2align 8 .type _Z2m1iiiPdS_S_,@function _Z2m1iiiPdS_S_: s_load_b32 s22, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s22, 1 s_cbranch_scc1 .LBB0_9 s_clause 0x2 s_load_b64 s[2:3], s[0:1], 0x4 s_load_b128 s[4:7], s[0:1], 0x10 s_load_b64 s[8:9], s[0:1], 0x20 s_mov_b32 s11, 0 s_mov_b32 s23, 0 s_mov_b32 s12, s11 s_waitcnt lgkmcnt(0) s_cmp_gt_i32 s3, 0 s_cselect_b32 s0, -1, 0 s_cmp_gt_i32 s2, 0 v_cndmask_b32_e64 v0, 0, 1, s0 s_cselect_b32 s0, -1, 0 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e64 v1, 0, 1, s0 v_cmp_ne_u32_e64 s0, 1, v0 v_mov_b32_e32 v0, 0 s_delay_alu instid0(VALU_DEP_3) v_cmp_ne_u32_e64 s1, 1, v1 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_add_i32 s23, s23, 1 s_add_i32 s12, s12, s2 s_cmp_eq_u32 s23, s22 s_cbranch_scc1 .LBB0_9 .LBB0_3: s_and_b32 vcc_lo, exec_lo, s0 s_cbranch_vccnz .LBB0_2 s_mov_b32 s13, s11 s_mov_b32 s10, 0 s_lshl_b64 s[14:15], s[12:13], 3 s_mul_i32 s24, s23, s3 s_add_u32 s14, s8, s14 s_addc_u32 s15, s9, s15 s_mov_b32 s13, s10 s_branch .LBB0_6 .p2align 6 .LBB0_5: s_add_i32 s13, s13, 1 s_add_i32 s10, s10, s2 s_cmp_eq_u32 s13, s3 s_cbranch_scc1 .LBB0_2 .LBB0_6: s_and_b32 vcc_lo, exec_lo, s1 s_cbranch_vccnz .LBB0_5 s_lshl_b64 s[16:17], s[10:11], 3 s_mov_b32 s19, s11 s_add_u32 s16, s6, s16 s_addc_u32 s17, s7, s17 s_add_i32 s18, s13, s24 s_mov_b32 s25, s2 s_lshl_b64 s[20:21], s[18:19], 3 s_mov_b64 s[18:19], s[14:15] s_add_u32 s20, s4, s20 s_addc_u32 s21, s5, s21 .p2align 6 .LBB0_8: s_clause 0x2 global_load_b64 v[1:2], v0, s[20:21] global_load_b64 v[3:4], v0, s[16:17] global_load_b64 v[5:6], v0, s[18:19] s_add_i32 s25, s25, -1 s_waitcnt vmcnt(0) v_fma_f64 v[1:2], v[1:2], v[3:4], v[5:6] global_store_b64 v0, v[1:2], s[18:19] s_add_u32 s18, s18, 8 s_addc_u32 s19, s19, 0 s_add_u32 s16, s16, 8 s_addc_u32 s17, s17, 0 s_cmp_eq_u32 s25, 0 s_cbranch_scc0 .LBB0_8 s_branch .LBB0_5 .LBB0_9: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z2m1iiiPdS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 40 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 26 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z2m1iiiPdS_S_, .Lfunc_end0-_Z2m1iiiPdS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .offset: 4 .size: 4 .value_kind: by_value - .offset: 8 .size: 4 .value_kind: by_value - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 40 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z2m1iiiPdS_S_ .private_segment_fixed_size: 0 .sgpr_count: 28 .sgpr_spill_count: 0 .symbol: _Z2m1iiiPdS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 7 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> // Write a first sequential implementation (matmult gpu1()) of matrix multiplication on the // GPU that uses only a single thread. It should work for all matrix sizes. Hints: // – You need CUDA code to allocate memory on the GPU, transfer A and B to the // GPU, transfer C back to the CPU, and free the allocated memory. // // Time your kernel for small matrix sizes and compare to the reference DGEMM on the CPU. // matrix times matrix // m represents the number of rows (the vertical length) of A and C, // k represents the number of columns of A and the n. of rows of B, // n represents the number of columns (the horizontal length) of B and C. // ____k____ ____n____ ____n____ // | | | | | | // m | A | X k | B | = m | C | // | | | | | | // --------- --------- --------- __global__ void m1(int m, int n, int k, double *A, double *B, double *C) { // for (int i = 0; i < m; i++) { // for (int j = 0; j < n; j++) { // C[i*n + j] = 0; // } // } for (int i = 0; i < m; i++) { for (int h = 0; h < k; h++){ for (int j = 0; j < n; j++) { C[i*n + j] += A[i*k + h] * B[h*n + j]; } } } } extern "C" { void matmult_gpu1(int m, int n, int k, double *A, double *B, double *C) { double* d_A, * d_B, * d_C; hipSetDevice(2); hipMalloc((void**)&d_A, m*k * sizeof(double)); hipMalloc((void**)&d_B, k*n * sizeof(double)); hipMalloc((void**)&d_C, m*n * sizeof(double)); hipMemcpy(d_A, A, m*k * sizeof(double), hipMemcpyHostToDevice); hipMemcpy(d_B, B, k*n * sizeof(double), hipMemcpyHostToDevice); // Initialize the output matrix with zeroes. hipMemset(d_C, 0, m*n * sizeof(double)); m1<<<1,1>>>(m, n, k, d_A, d_B, d_C); hipDeviceSynchronize(); hipMemcpy(C, d_C, m*n * sizeof(double), hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } }
.text .file "matmult_gpu1.hip" .globl _Z17__device_stub__m1iiiPdS_S_ # -- Begin function _Z17__device_stub__m1iiiPdS_S_ .p2align 4, 0x90 .type _Z17__device_stub__m1iiiPdS_S_,@function _Z17__device_stub__m1iiiPdS_S_: # @_Z17__device_stub__m1iiiPdS_S_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movq %rcx, 88(%rsp) movq %r8, 80(%rsp) movq %r9, 72(%rsp) leaq 20(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 88(%rsp), %rax movq %rax, 120(%rsp) leaq 80(%rsp), %rax movq %rax, 128(%rsp) leaq 72(%rsp), %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z2m1iiiPdS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z17__device_stub__m1iiiPdS_S_, .Lfunc_end0-_Z17__device_stub__m1iiiPdS_S_ .cfi_endproc # -- End function .globl matmult_gpu1 # -- Begin function matmult_gpu1 .p2align 4, 0x90 .type matmult_gpu1,@function matmult_gpu1: # @matmult_gpu1 .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $200, %rsp .cfi_def_cfa_offset 256 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movq %r9, 64(%rsp) # 8-byte Spill movq %r8, 56(%rsp) # 8-byte Spill movq %rcx, 48(%rsp) # 8-byte Spill movl %edx, %ebp movl %esi, %r15d movl %edi, %r12d movl $2, %edi callq hipSetDevice movl %ebp, %eax imull %r12d, %eax movslq %eax, %r13 shlq $3, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi callq hipMalloc movl %ebp, %eax imull %r15d, %eax movslq %eax, %rbx shlq $3, %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi callq hipMalloc movl %r15d, %eax imull %r12d, %eax movslq %eax, %r14 shlq $3, %r14 leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 24(%rsp), %rdi movq 48(%rsp), %rsi # 8-byte Reload movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 16(%rsp), %rdi movq 56(%rsp), %rsi # 8-byte Reload movq %rbx, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi xorl %esi, %esi movq %r14, %rdx callq hipMemset movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 24(%rsp), %rax movq 16(%rsp), %rcx movq 8(%rsp), %rdx movl %r12d, 44(%rsp) movl %r15d, 40(%rsp) movl %ebp, 36(%rsp) movq %rax, 136(%rsp) movq %rcx, 128(%rsp) movq %rdx, 120(%rsp) leaq 44(%rsp), %rax movq %rax, 144(%rsp) leaq 40(%rsp), %rax movq %rax, 152(%rsp) leaq 36(%rsp), %rax movq %rax, 160(%rsp) leaq 136(%rsp), %rax movq %rax, 168(%rsp) leaq 128(%rsp), %rax movq %rax, 176(%rsp) leaq 120(%rsp), %rax movq %rax, 184(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z2m1iiiPdS_S_, %edi pushq 72(%rsp) .cfi_adjust_cfa_offset 8 pushq 88(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize movq 8(%rsp), %rsi movq 64(%rsp), %rdi # 8-byte Reload movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree addq $200, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size matmult_gpu1, .Lfunc_end1-matmult_gpu1 .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z2m1iiiPdS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z2m1iiiPdS_S_,@object # @_Z2m1iiiPdS_S_ .section .rodata,"a",@progbits .globl _Z2m1iiiPdS_S_ .p2align 3, 0x0 _Z2m1iiiPdS_S_: .quad _Z17__device_stub__m1iiiPdS_S_ .size _Z2m1iiiPdS_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z2m1iiiPdS_S_" .size .L__unnamed_1, 15 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z17__device_stub__m1iiiPdS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z2m1iiiPdS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z2m1iiiPdS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ IMAD.MOV.U32 R0, RZ, RZ, 0x1 ; /* 0x00000001ff007424 */ /* 0x000fca00078e00ff */ /*0020*/ ISETP.LE.AND P0, PT, R0, c[0x0][0x160], PT ; /* 0x0000580000007a0c */ /* 0x000fda0003f03270 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ ISETP.LE.AND P0, PT, R0, c[0x0][0x168], PT ; /* 0x00005a0000007a0c */ /* 0x000fda0003f03270 */ /*0050*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0060*/ IMAD.MOV.U32 R0, RZ, RZ, 0x1 ; /* 0x00000001ff007424 */ /* 0x000fe200078e00ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0080*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff027624 */ /* 0x000fc600078e00ff */ /*0090*/ IADD3 R0, -R0, c[0x0][0x164], RZ ; /* 0x0000590000007a10 */ /* 0x000fc80007ffe1ff */ /*00a0*/ ISETP.GE.U32.AND P1, PT, R0, 0x3, PT ; /* 0x000000030000780c */ /* 0x000fe40003f26070 */ /*00b0*/ LOP3.LUT R0, R2, 0x3, RZ, 0xc0, !PT ; /* 0x0000000302007812 */ /* 0x000fe200078ec0ff */ /*00c0*/ IMAD.MOV.U32 R2, RZ, RZ, RZ ; /* 0x000000ffff027224 */ /* 0x000fc600078e00ff */ /*00d0*/ IADD3 R3, -R0, c[0x0][0x164], RZ ; /* 0x0000590000037a10 */ /* 0x000fe40007ffe1ff */ /*00e0*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff047624 */ /* 0x000fca00078e00ff */ /*00f0*/ ISETP.GE.AND P0, PT, R4, 0x1, PT ; /* 0x000000010400780c */ /* 0x000fda0003f06270 */ /*0100*/ @!P0 BRA 0xf00 ; /* 0x00000df000008947 */ /* 0x003fea0003800000 */ /*0110*/ IMAD R4, R2, c[0x0][0x164], RZ ; /* 0x0000590002047a24 */ /* 0x000fe400078e02ff */ /*0120*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc800078e00ff */ /*0130*/ IMAD R10, R2, c[0x0][0x168], R5 ; /* 0x00005a00020a7a24 */ /* 0x000fe400078e0205 */ /*0140*/ IMAD R6, R5.reuse, c[0x0][0x164], RZ ; /* 0x0000590005067a24 */ /* 0x041fe200078e02ff */ /*0150*/ IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105057810 */ /* 0x000fe20007ffe0ff */ /*0160*/ IMAD.MOV.U32 R7, RZ, RZ, 0x8 ; /* 0x00000008ff077424 */ /* 0x000fe400078e00ff */ /*0170*/ IMAD.MOV.U32 R9, RZ, RZ, RZ ; /* 0x000000ffff097224 */ /* 0x000fe200078e00ff */ /*0180*/ ISETP.GE.AND P2, PT, R5, c[0x0][0x168], PT ; /* 0x00005a0005007a0c */ /* 0x000fe20003f46270 */ /*0190*/ IMAD.WIDE R10, R10, R7, c[0x0][0x170] ; /* 0x00005c000a0a7625 */ /* 0x000fe200078e0207 */ /*01a0*/ @!P1 BRA 0xd60 ; /* 0x00000bb000009947 */ /* 0x002fea0003800000 */ /*01b0*/ ISETP.GT.AND P0, PT, R3, RZ, PT ; /* 0x000000ff0300720c */ /* 0x000fe20003f04270 */ /*01c0*/ IMAD.MOV.U32 R9, RZ, RZ, RZ ; /* 0x000000ffff097224 */ /* 0x000fc400078e00ff */ /*01d0*/ IMAD.MOV.U32 R22, RZ, RZ, c[0x0][0x180] ; /* 0x00006000ff167624 */ /* 0x000fe400078e00ff */ /*01e0*/ IMAD.MOV.U32 R23, RZ, RZ, c[0x0][0x184] ; /* 0x00006100ff177624 */ /* 0x000fe400078e00ff */ /*01f0*/ IMAD.MOV.U32 R8, RZ, RZ, R3 ; /* 0x000000ffff087224 */ /* 0x000fe400078e0003 */ /*0200*/ IMAD.WIDE R12, R6, R7, c[0x0][0x178] ; /* 0x00005e00060c7625 */ /* 0x000fc800078e0207 */ /*0210*/ @!P0 BRA 0xb70 ; /* 0x0000095000008947 */ /* 0x000fea0003800000 */ /*0220*/ ISETP.GT.AND P3, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f64270 */ /*0230*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*0240*/ @!P3 BRA 0x810 ; /* 0x000005c00000b947 */ /* 0x000fea0003800000 */ /*0250*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0260*/ IMAD.WIDE R20, R4, 0x8, R22 ; /* 0x0000000804147825 */ /* 0x001fe200078e0216 */ /*0270*/ LDG.E.64 R18, [R12.64] ; /* 0x000000040c127981 */ /* 0x000ea8000c1e1b00 */ /*0280*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea8000c1e1b00 */ /*0290*/ LDG.E.64 R14, [R20.64] ; /* 0x00000004140e7981 */ /* 0x000ea4000c1e1b00 */ /*02a0*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x004044000000000e */ /*02b0*/ LDG.E.64 R14, [R20.64+0x8] ; /* 0x00000804140e7981 */ /* 0x001eaa000c1e1b00 */ /*02c0*/ STG.E.64 [R20.64], R28 ; /* 0x0000001c14007986 */ /* 0x0021e8000c101b04 */ /*02d0*/ LDG.E.64 R18, [R12.64+0x8] ; /* 0x000008040c127981 */ /* 0x000ea8000c1e1b00 */ /*02e0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*02f0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x004284000000000e */ /*0300*/ LDG.E.64 R14, [R20.64+0x10] ; /* 0x00001004140e7981 */ /* 0x002eea000c1e1b00 */ /*0310*/ STG.E.64 [R20.64+0x8], R26 ; /* 0x0000081a14007986 */ /* 0x0043e8000c101b04 */ /*0320*/ LDG.E.64 R18, [R12.64+0x10] ; /* 0x000010040c127981 */ /* 0x000ee8000c1e1b00 */ /*0330*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ee4000c1e1b00 */ /*0340*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0084c4000000000e */ /*0350*/ LDG.E.64 R14, [R20.64+0x18] ; /* 0x00001804140e7981 */ /* 0x004e2a000c1e1b00 */ /*0360*/ STG.E.64 [R20.64+0x10], R24 ; /* 0x0000101814007986 */ /* 0x0085e8000c101b04 */ /*0370*/ LDG.E.64 R18, [R12.64+0x18] ; /* 0x000018040c127981 */ /* 0x000e28000c1e1b00 */ /*0380*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0390*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*03a0*/ LDG.E.64 R14, [R20.64+0x20] ; /* 0x00002004140e7981 */ /* 0x001e6a000c1e1b00 */ /*03b0*/ STG.E.64 [R20.64+0x18], R28 ; /* 0x0000181c14007986 */ /* 0x0081e8000c101b04 */ /*03c0*/ LDG.E.64 R18, [R12.64+0x20] ; /* 0x000020040c127981 */ /* 0x000e68000c1e1b00 */ /*03d0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*03e0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*03f0*/ LDG.E.64 R14, [R20.64+0x28] ; /* 0x00002804140e7981 */ /* 0x002eaa000c1e1b00 */ /*0400*/ STG.E.64 [R20.64+0x20], R26 ; /* 0x0000201a14007986 */ /* 0x0083e8000c101b04 */ /*0410*/ LDG.E.64 R18, [R12.64+0x28] ; /* 0x000028040c127981 */ /* 0x000ea8000c1e1b00 */ /*0420*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0430*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0440*/ LDG.E.64 R14, [R20.64+0x30] ; /* 0x00003004140e7981 */ /* 0x004e2a000c1e1b00 */ /*0450*/ STG.E.64 [R20.64+0x28], R24 ; /* 0x0000281814007986 */ /* 0x0085e8000c101b04 */ /*0460*/ LDG.E.64 R18, [R12.64+0x30] ; /* 0x000030040c127981 */ /* 0x000e28000c1e1b00 */ /*0470*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0480*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0490*/ LDG.E.64 R14, [R20.64+0x38] ; /* 0x00003804140e7981 */ /* 0x001e6a000c1e1b00 */ /*04a0*/ STG.E.64 [R20.64+0x30], R28 ; /* 0x0000301c14007986 */ /* 0x0081e8000c101b04 */ /*04b0*/ LDG.E.64 R18, [R12.64+0x38] ; /* 0x000038040c127981 */ /* 0x000e68000c1e1b00 */ /*04c0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*04d0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*04e0*/ LDG.E.64 R14, [R20.64+0x40] ; /* 0x00004004140e7981 */ /* 0x002eaa000c1e1b00 */ /*04f0*/ STG.E.64 [R20.64+0x38], R26 ; /* 0x0000381a14007986 */ /* 0x0083e8000c101b04 */ /*0500*/ LDG.E.64 R18, [R12.64+0x40] ; /* 0x000040040c127981 */ /* 0x000ea8000c1e1b00 */ /*0510*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0520*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0530*/ LDG.E.64 R14, [R20.64+0x48] ; /* 0x00004804140e7981 */ /* 0x004e2a000c1e1b00 */ /*0540*/ STG.E.64 [R20.64+0x40], R24 ; /* 0x0000401814007986 */ /* 0x0085e8000c101b04 */ /*0550*/ LDG.E.64 R18, [R12.64+0x48] ; /* 0x000048040c127981 */ /* 0x000e28000c1e1b00 */ /*0560*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0570*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0580*/ LDG.E.64 R14, [R20.64+0x50] ; /* 0x00005004140e7981 */ /* 0x001e6a000c1e1b00 */ /*0590*/ STG.E.64 [R20.64+0x48], R28 ; /* 0x0000481c14007986 */ /* 0x0081e8000c101b04 */ /*05a0*/ LDG.E.64 R18, [R12.64+0x50] ; /* 0x000050040c127981 */ /* 0x000e68000c1e1b00 */ /*05b0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*05c0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*05d0*/ LDG.E.64 R14, [R20.64+0x58] ; /* 0x00005804140e7981 */ /* 0x002eaa000c1e1b00 */ /*05e0*/ STG.E.64 [R20.64+0x50], R26 ; /* 0x0000501a14007986 */ /* 0x0083e8000c101b04 */ /*05f0*/ LDG.E.64 R18, [R12.64+0x58] ; /* 0x000058040c127981 */ /* 0x000ea8000c1e1b00 */ /*0600*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea4000c1e1b00 */ /*0610*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x0044c4000000000e */ /*0620*/ LDG.E.64 R14, [R20.64+0x60] ; /* 0x00006004140e7981 */ /* 0x004e2a000c1e1b00 */ /*0630*/ STG.E.64 [R20.64+0x58], R24 ; /* 0x0000581814007986 */ /* 0x0085e8000c101b04 */ /*0640*/ LDG.E.64 R18, [R12.64+0x60] ; /* 0x000060040c127981 */ /* 0x000e28000c1e1b00 */ /*0650*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e24000c1e1b00 */ /*0660*/ DFMA R28, R18, R16, R14 ; /* 0x00000010121c722b */ /* 0x0010c4000000000e */ /*0670*/ LDG.E.64 R14, [R20.64+0x68] ; /* 0x00006804140e7981 */ /* 0x001e6a000c1e1b00 */ /*0680*/ STG.E.64 [R20.64+0x60], R28 ; /* 0x0000601c14007986 */ /* 0x0081e8000c101b04 */ /*0690*/ LDG.E.64 R18, [R12.64+0x68] ; /* 0x000068040c127981 */ /* 0x000e68000c1e1b00 */ /*06a0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000e64000c1e1b00 */ /*06b0*/ DFMA R26, R18, R16, R14 ; /* 0x00000010121a722b */ /* 0x0022c4000000000e */ /*06c0*/ LDG.E.64 R14, [R20.64+0x70] ; /* 0x00007004140e7981 */ /* 0x002eaa000c1e1b00 */ /*06d0*/ STG.E.64 [R20.64+0x68], R26 ; /* 0x0000681a14007986 */ /* 0x0081e8000c101b04 */ /*06e0*/ LDG.E.64 R18, [R12.64+0x70] ; /* 0x000070040c127981 */ /* 0x000ea8000c1e1b00 */ /*06f0*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ea2000c1e1b00 */ /*0700*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fe20007ffe0ff */ /*0710*/ DFMA R24, R18, R16, R14 ; /* 0x000000101218722b */ /* 0x004284000000000e */ /*0720*/ LDG.E.64 R14, [R20.64+0x78] ; /* 0x00007804140e7981 */ /* 0x002eea000c1e1b00 */ /*0730*/ STG.E.64 [R20.64+0x70], R24 ; /* 0x0000701814007986 */ /* 0x0041e8000c101b04 */ /*0740*/ LDG.E.64 R18, [R12.64+0x78] ; /* 0x000078040c127981 */ /* 0x000ee8000c1e1b00 */ /*0750*/ LDG.E.64 R16, [R10.64] ; /* 0x000000040a107981 */ /* 0x000ee2000c1e1b00 */ /*0760*/ ISETP.GT.AND P3, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fc40003f64270 */ /*0770*/ IADD3 R22, P4, R22, 0x80, RZ ; /* 0x0000008016167810 */ /* 0x000fe40007f9e0ff */ /*0780*/ IADD3 R9, R9, 0x10, RZ ; /* 0x0000001009097810 */ /* 0x000fc60007ffe0ff */ /*0790*/ IMAD.X R23, RZ, RZ, R23, P4 ; /* 0x000000ffff177224 */ /* 0x000fe200020e0617 */ /*07a0*/ DFMA R14, R18, R16, R14 ; /* 0x00000010120e722b */ /* 0x0082a4000000000e */ /*07b0*/ IADD3 R16, P5, R12, 0x80, RZ ; /* 0x000000800c107810 */ /* 0x002fc80007fbe0ff */ /*07c0*/ IADD3.X R17, RZ, R13, RZ, P5, !PT ; /* 0x0000000dff117210 */ /* 0x000fe20002ffe4ff */ /*07d0*/ STG.E.64 [R20.64+0x78], R14 ; /* 0x0000780e14007986 */ /* 0x0041e2000c101b04 */ /*07e0*/ IMAD.MOV.U32 R12, RZ, RZ, R16 ; /* 0x000000ffff0c7224 */ /* 0x000fc600078e0010 */ /*07f0*/ IMAD.MOV.U32 R13, RZ, RZ, R17 ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e0011 */ /*0800*/ @P3 BRA 0x260 ; /* 0xfffffa5000003947 */ /* 0x000fea000383ffff */ /*0810*/ ISETP.GT.AND P3, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f64270 */ /*0820*/ @!P3 BRA 0xb50 ; /* 0x000003200000b947 */ /* 0x000fea0003800000 */ /*0830*/ IMAD.WIDE R14, R4, 0x8, R22 ; /* 0x00000008040e7825 */ /* 0x001fe200078e0216 */ /*0840*/ LDG.E.64 R20, [R12.64] ; /* 0x000000040c147981 */ /* 0x000ea8000c1e1b00 */ /*0850*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea8000c1e1b00 */ /*0860*/ LDG.E.64 R16, [R14.64] ; /* 0x000000040e107981 */ /* 0x000ea4000c1e1b00 */ /*0870*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0040440000000010 */ /*0880*/ LDG.E.64 R16, [R14.64+0x8] ; /* 0x000008040e107981 */ /* 0x001eaa000c1e1b00 */ /*0890*/ STG.E.64 [R14.64], R24 ; /* 0x000000180e007986 */ /* 0x0021e8000c101b04 */ /*08a0*/ LDG.E.64 R20, [R12.64+0x8] ; /* 0x000008040c147981 */ /* 0x000ea8000c1e1b00 */ /*08b0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*08c0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0042840000000010 */ /*08d0*/ LDG.E.64 R16, [R14.64+0x10] ; /* 0x000010040e107981 */ /* 0x002eea000c1e1b00 */ /*08e0*/ STG.E.64 [R14.64+0x8], R28 ; /* 0x0000081c0e007986 */ /* 0x0043e8000c101b04 */ /*08f0*/ LDG.E.64 R20, [R12.64+0x10] ; /* 0x000010040c147981 */ /* 0x000ee8000c1e1b00 */ /*0900*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee4000c1e1b00 */ /*0910*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0084c40000000010 */ /*0920*/ LDG.E.64 R16, [R14.64+0x18] ; /* 0x000018040e107981 */ /* 0x004e2a000c1e1b00 */ /*0930*/ STG.E.64 [R14.64+0x10], R26 ; /* 0x0000101a0e007986 */ /* 0x0085e8000c101b04 */ /*0940*/ LDG.E.64 R20, [R12.64+0x18] ; /* 0x000018040c147981 */ /* 0x000e28000c1e1b00 */ /*0950*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000e24000c1e1b00 */ /*0960*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0010c40000000010 */ /*0970*/ LDG.E.64 R16, [R14.64+0x20] ; /* 0x000020040e107981 */ /* 0x001e6a000c1e1b00 */ /*0980*/ STG.E.64 [R14.64+0x18], R24 ; /* 0x000018180e007986 */ /* 0x0081e8000c101b04 */ /*0990*/ LDG.E.64 R20, [R12.64+0x20] ; /* 0x000020040c147981 */ /* 0x000e68000c1e1b00 */ /*09a0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000e64000c1e1b00 */ /*09b0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0022c40000000010 */ /*09c0*/ LDG.E.64 R16, [R14.64+0x28] ; /* 0x000028040e107981 */ /* 0x002eaa000c1e1b00 */ /*09d0*/ STG.E.64 [R14.64+0x20], R28 ; /* 0x0000201c0e007986 */ /* 0x0083e8000c101b04 */ /*09e0*/ LDG.E.64 R20, [R12.64+0x28] ; /* 0x000028040c147981 */ /* 0x000ea8000c1e1b00 */ /*09f0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*0a00*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0044c40000000010 */ /*0a10*/ LDG.E.64 R16, [R14.64+0x30] ; /* 0x000030040e107981 */ /* 0x004e2a000c1e1b00 */ /*0a20*/ STG.E.64 [R14.64+0x28], R26 ; /* 0x0000281a0e007986 */ /* 0x0083e8000c101b04 */ /*0a30*/ LDG.E.64 R18, [R12.64+0x30] ; /* 0x000030040c127981 */ /* 0x000e28000c1e1b00 */ /*0a40*/ LDG.E.64 R20, [R10.64] ; /* 0x000000040a147981 */ /* 0x000e24000c1e1b00 */ /*0a50*/ DFMA R24, R18, R20, R16 ; /* 0x000000141218722b */ /* 0x0010840000000010 */ /*0a60*/ LDG.E.64 R16, [R14.64+0x38] ; /* 0x000038040e107981 */ /* 0x001eea000c1e1b00 */ /*0a70*/ STG.E.64 [R14.64+0x30], R24 ; /* 0x000030180e007986 */ /* 0x0043e8000c101b04 */ /*0a80*/ LDG.E.64 R18, [R12.64+0x38] ; /* 0x000038040c127981 */ /* 0x000ee8000c1e1b00 */ /*0a90*/ LDG.E.64 R20, [R10.64] ; /* 0x000000040a147981 */ /* 0x000ee2000c1e1b00 */ /*0aa0*/ IADD3 R22, P3, R22, 0x40, RZ ; /* 0x0000004016167810 */ /* 0x000fc40007f7e0ff */ /*0ab0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0ac0*/ IADD3 R9, R9, 0x8, RZ ; /* 0x0000000809097810 */ /* 0x000fe20007ffe0ff */ /*0ad0*/ IMAD.X R23, RZ, RZ, R23, P3 ; /* 0x000000ffff177224 */ /* 0x000fe200018e0617 */ /*0ae0*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe20007ffe0ff */ /*0af0*/ DFMA R16, R18, R20, R16 ; /* 0x000000141210722b */ /* 0x0080a40000000010 */ /*0b00*/ IADD3 R18, P4, R12, 0x40, RZ ; /* 0x000000400c127810 */ /* 0x001fca0007f9e0ff */ /*0b10*/ STG.E.64 [R14.64+0x38], R16 ; /* 0x000038100e007986 */ /* 0x0043e2000c101b04 */ /*0b20*/ IMAD.X R19, RZ, RZ, R13, P4 ; /* 0x000000ffff137224 */ /* 0x000fe400020e060d */ /*0b30*/ IMAD.MOV.U32 R12, RZ, RZ, R18 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0012 */ /*0b40*/ IMAD.MOV.U32 R13, RZ, RZ, R19 ; /* 0x000000ffff0d7224 */ /* 0x000fe400078e0013 */ /*0b50*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0b60*/ @!P0 BRA 0xd60 ; /* 0x000001f000008947 */ /* 0x000fea0003800000 */ /*0b70*/ IMAD.WIDE R14, R4, 0x8, R22 ; /* 0x00000008040e7825 */ /* 0x003fe200078e0216 */ /*0b80*/ LDG.E.64 R20, [R12.64] ; /* 0x000000040c147981 */ /* 0x000ea8000c1e1b00 */ /*0b90*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea8000c1e1b00 */ /*0ba0*/ LDG.E.64 R16, [R14.64] ; /* 0x000000040e107981 */ /* 0x000ea4000c1e1b00 */ /*0bb0*/ DFMA R28, R20, R18, R16 ; /* 0x00000012141c722b */ /* 0x0040440000000010 */ /*0bc0*/ LDG.E.64 R16, [R14.64+0x8] ; /* 0x000008040e107981 */ /* 0x001eaa000c1e1b00 */ /*0bd0*/ STG.E.64 [R14.64], R28 ; /* 0x0000001c0e007986 */ /* 0x0021e8000c101b04 */ /*0be0*/ LDG.E.64 R20, [R12.64+0x8] ; /* 0x000008040c147981 */ /* 0x000ea8000c1e1b00 */ /*0bf0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ea4000c1e1b00 */ /*0c00*/ DFMA R26, R20, R18, R16 ; /* 0x00000012141a722b */ /* 0x0042840000000010 */ /*0c10*/ LDG.E.64 R16, [R14.64+0x10] ; /* 0x000010040e107981 */ /* 0x002eea000c1e1b00 */ /*0c20*/ STG.E.64 [R14.64+0x8], R26 ; /* 0x0000081a0e007986 */ /* 0x0041e8000c101b04 */ /*0c30*/ LDG.E.64 R20, [R12.64+0x10] ; /* 0x000010040c147981 */ /* 0x000ee8000c1e1b00 */ /*0c40*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee2000c1e1b00 */ /*0c50*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fe20007ffe0ff */ /*0c60*/ DFMA R24, R20, R18, R16 ; /* 0x000000121418722b */ /* 0x0082840000000010 */ /*0c70*/ LDG.E.64 R16, [R14.64+0x18] ; /* 0x000018040e107981 */ /* 0x002eea000c1e1b00 */ /*0c80*/ STG.E.64 [R14.64+0x10], R24 ; /* 0x000010180e007986 */ /* 0x0041e8000c101b04 */ /*0c90*/ LDG.E.64 R20, [R12.64+0x18] ; /* 0x000018040c147981 */ /* 0x000ee8000c1e1b00 */ /*0ca0*/ LDG.E.64 R18, [R10.64] ; /* 0x000000040a127981 */ /* 0x000ee2000c1e1b00 */ /*0cb0*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fc40003f05270 */ /*0cc0*/ IADD3 R22, P3, R22, 0x20, RZ ; /* 0x0000002016167810 */ /* 0x000fe40007f7e0ff */ /*0cd0*/ IADD3 R9, R9, 0x4, RZ ; /* 0x0000000409097810 */ /* 0x000fc60007ffe0ff */ /*0ce0*/ IMAD.X R23, RZ, RZ, R23, P3 ; /* 0x000000ffff177224 */ /* 0x000fe200018e0617 */ /*0cf0*/ DFMA R16, R20, R18, R16 ; /* 0x000000121410722b */ /* 0x0082a40000000010 */ /*0d00*/ IADD3 R18, P4, R12, 0x20, RZ ; /* 0x000000200c127810 */ /* 0x002fc80007f9e0ff */ /*0d10*/ IADD3.X R19, RZ, R13, RZ, P4, !PT ; /* 0x0000000dff137210 */ /* 0x000fe200027fe4ff */ /*0d20*/ STG.E.64 [R14.64+0x18], R16 ; /* 0x000018100e007986 */ /* 0x0041e2000c101b04 */ /*0d30*/ IMAD.MOV.U32 R12, RZ, RZ, R18 ; /* 0x000000ffff0c7224 */ /* 0x000fc600078e0012 */ /*0d40*/ IMAD.MOV.U32 R13, RZ, RZ, R19 ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e0013 */ /*0d50*/ @P0 BRA 0xb70 ; /* 0xfffffe1000000947 */ /* 0x001fea000383ffff */ /*0d60*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f05270 */ /*0d70*/ @!P0 BRA 0xef0 ; /* 0x0000017000008947 */ /* 0x000fea0003800000 */ /*0d80*/ IMAD.IADD R8, R6, 0x1, R9.reuse ; /* 0x0000000106087824 */ /* 0x100fe200078e0209 */ /*0d90*/ LDG.E.64 R14, [R10.64] ; /* 0x000000040a0e7981 */ /* 0x003ea2000c1e1b00 */ /*0da0*/ IMAD.IADD R6, R4, 0x1, R9 ; /* 0x0000000104067824 */ /* 0x000fe400078e0209 */ /*0db0*/ IMAD.WIDE R8, R8, R7, c[0x0][0x178] ; /* 0x00005e0008087625 */ /* 0x000fc800078e0207 */ /*0dc0*/ IMAD.WIDE R6, R6, R7, c[0x0][0x180] ; /* 0x0000600006067625 */ /* 0x000fe200078e0207 */ /*0dd0*/ LDG.E.64 R12, [R8.64] ; /* 0x00000004080c7981 */ /* 0x000ea8000c1e1b00 */ /*0de0*/ LDG.E.64 R16, [R6.64] ; /* 0x0000000406107981 */ /* 0x000ea2000c1e1b00 */ /*0df0*/ ISETP.NE.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fe20003f05270 */ /*0e00*/ DFMA R12, R12, R14, R16 ; /* 0x0000000e0c0c722b */ /* 0x004e0e0000000010 */ /*0e10*/ STG.E.64 [R6.64], R12 ; /* 0x0000000c06007986 */ /* 0x0011ea000c101b04 */ /*0e20*/ @!P0 BRA 0xef0 ; /* 0x000000c000008947 */ /* 0x000fea0003800000 */ /*0e30*/ LDG.E.64 R12, [R8.64+0x8] ; /* 0x00000804080c7981 */ /* 0x001ea8000c1e1b00 */ /*0e40*/ LDG.E.64 R14, [R10.64] ; /* 0x000000040a0e7981 */ /* 0x000ea8000c1e1b00 */ /*0e50*/ LDG.E.64 R16, [R6.64+0x8] ; /* 0x0000080406107981 */ /* 0x000ea2000c1e1b00 */ /*0e60*/ ISETP.NE.AND P0, PT, R0, 0x2, PT ; /* 0x000000020000780c */ /* 0x000fe20003f05270 */ /*0e70*/ DFMA R12, R12, R14, R16 ; /* 0x0000000e0c0c722b */ /* 0x004e0e0000000010 */ /*0e80*/ STG.E.64 [R6.64+0x8], R12 ; /* 0x0000080c06007986 */ /* 0x0011ea000c101b04 */ /*0e90*/ @!P0 BRA 0xef0 ; /* 0x0000005000008947 */ /* 0x000fea0003800000 */ /*0ea0*/ LDG.E.64 R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ea8000c1e1b00 */ /*0eb0*/ LDG.E.64 R8, [R8.64+0x10] ; /* 0x0000100408087981 */ /* 0x000ea8000c1e1b00 */ /*0ec0*/ LDG.E.64 R12, [R6.64+0x10] ; /* 0x00001004060c7981 */ /* 0x001ea4000c1e1b00 */ /*0ed0*/ DFMA R12, R10, R8, R12 ; /* 0x000000080a0c722b */ /* 0x004e0e000000000c */ /*0ee0*/ STG.E.64 [R6.64+0x10], R12 ; /* 0x0000100c06007986 */ /* 0x0011e8000c101b04 */ /*0ef0*/ @!P2 BRA 0x130 ; /* 0xfffff2300000a947 */ /* 0x000fea000383ffff */ /*0f00*/ IADD3 R2, R2, 0x1, RZ ; /* 0x0000000102027810 */ /* 0x000fc80007ffe0ff */ /*0f10*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x160], PT ; /* 0x0000580002007a0c */ /* 0x000fda0003f06270 */ /*0f20*/ @P0 CALL.REL.NOINC 0xf40 ; /* 0x0000001000000944 */ /* 0x000fe20003c00000 */ /*0f30*/ BRA 0xe0 ; /* 0xfffff1a000007947 */ /* 0x000fea000383ffff */ /*0f40*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0f50*/ BRA 0xf50; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0f60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0f90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0fe0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ff0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z2m1iiiPdS_S_ .globl _Z2m1iiiPdS_S_ .p2align 8 .type _Z2m1iiiPdS_S_,@function _Z2m1iiiPdS_S_: s_load_b32 s22, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s22, 1 s_cbranch_scc1 .LBB0_9 s_clause 0x2 s_load_b64 s[2:3], s[0:1], 0x4 s_load_b128 s[4:7], s[0:1], 0x10 s_load_b64 s[8:9], s[0:1], 0x20 s_mov_b32 s11, 0 s_mov_b32 s23, 0 s_mov_b32 s12, s11 s_waitcnt lgkmcnt(0) s_cmp_gt_i32 s3, 0 s_cselect_b32 s0, -1, 0 s_cmp_gt_i32 s2, 0 v_cndmask_b32_e64 v0, 0, 1, s0 s_cselect_b32 s0, -1, 0 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e64 v1, 0, 1, s0 v_cmp_ne_u32_e64 s0, 1, v0 v_mov_b32_e32 v0, 0 s_delay_alu instid0(VALU_DEP_3) v_cmp_ne_u32_e64 s1, 1, v1 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_add_i32 s23, s23, 1 s_add_i32 s12, s12, s2 s_cmp_eq_u32 s23, s22 s_cbranch_scc1 .LBB0_9 .LBB0_3: s_and_b32 vcc_lo, exec_lo, s0 s_cbranch_vccnz .LBB0_2 s_mov_b32 s13, s11 s_mov_b32 s10, 0 s_lshl_b64 s[14:15], s[12:13], 3 s_mul_i32 s24, s23, s3 s_add_u32 s14, s8, s14 s_addc_u32 s15, s9, s15 s_mov_b32 s13, s10 s_branch .LBB0_6 .p2align 6 .LBB0_5: s_add_i32 s13, s13, 1 s_add_i32 s10, s10, s2 s_cmp_eq_u32 s13, s3 s_cbranch_scc1 .LBB0_2 .LBB0_6: s_and_b32 vcc_lo, exec_lo, s1 s_cbranch_vccnz .LBB0_5 s_lshl_b64 s[16:17], s[10:11], 3 s_mov_b32 s19, s11 s_add_u32 s16, s6, s16 s_addc_u32 s17, s7, s17 s_add_i32 s18, s13, s24 s_mov_b32 s25, s2 s_lshl_b64 s[20:21], s[18:19], 3 s_mov_b64 s[18:19], s[14:15] s_add_u32 s20, s4, s20 s_addc_u32 s21, s5, s21 .p2align 6 .LBB0_8: s_clause 0x2 global_load_b64 v[1:2], v0, s[20:21] global_load_b64 v[3:4], v0, s[16:17] global_load_b64 v[5:6], v0, s[18:19] s_add_i32 s25, s25, -1 s_waitcnt vmcnt(0) v_fma_f64 v[1:2], v[1:2], v[3:4], v[5:6] global_store_b64 v0, v[1:2], s[18:19] s_add_u32 s18, s18, 8 s_addc_u32 s19, s19, 0 s_add_u32 s16, s16, 8 s_addc_u32 s17, s17, 0 s_cmp_eq_u32 s25, 0 s_cbranch_scc0 .LBB0_8 s_branch .LBB0_5 .LBB0_9: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z2m1iiiPdS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 40 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 26 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z2m1iiiPdS_S_, .Lfunc_end0-_Z2m1iiiPdS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .offset: 4 .size: 4 .value_kind: by_value - .offset: 8 .size: 4 .value_kind: by_value - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 40 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z2m1iiiPdS_S_ .private_segment_fixed_size: 0 .sgpr_count: 28 .sgpr_spill_count: 0 .symbol: _Z2m1iiiPdS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 7 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000b0dc7_00000000-6_matmult_gpu1.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ .type _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_, @function _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_: .LFB2052: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 44(%rsp) movl %esi, 40(%rsp) movl %edx, 36(%rsp) movq %rcx, 24(%rsp) movq %r8, 16(%rsp) movq %r9, 8(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 44(%rsp), %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rax movq %rax, 120(%rsp) leaq 36(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z2m1iiiPdS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2052: .size _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_, .-_Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ .globl _Z2m1iiiPdS_S_ .type _Z2m1iiiPdS_S_, @function _Z2m1iiiPdS_S_: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _Z2m1iiiPdS_S_, .-_Z2m1iiiPdS_S_ .globl matmult_gpu1 .type matmult_gpu1, @function matmult_gpu1: .LFB2027: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $104, %rsp .cfi_def_cfa_offset 160 movl %edi, %r13d movl %esi, %r14d movl %edx, %r15d movq %rcx, 8(%rsp) movq %r8, 16(%rsp) movq %r9, 24(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax movl $2, %edi call cudaSetDevice@PLT movl %r13d, %r12d imull %r15d, %r12d movslq %r12d, %r12 salq $3, %r12 leaq 40(%rsp), %rdi movq %r12, %rsi call cudaMalloc@PLT movl %r15d, %ebp imull %r14d, %ebp movslq %ebp, %rbp salq $3, %rbp leaq 48(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl %r13d, %ebx imull %r14d, %ebx movslq %ebx, %rbx salq $3, %rbx leaq 56(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r12, %rdx movq 8(%rsp), %rsi movq 40(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbp, %rdx movq 16(%rsp), %rsi movq 48(%rsp), %rdi call cudaMemcpy@PLT movq %rbx, %rdx movl $0, %esi movq 56(%rsp), %rdi call cudaMemset@PLT movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $0, %r9d movl $0, %r8d movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L15 .L12: call cudaDeviceSynchronize@PLT movl $2, %ecx movq %rbx, %rdx movq 56(%rsp), %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movq 40(%rsp), %rdi call cudaFree@PLT movq 48(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rdi call cudaFree@PLT movq 88(%rsp), %rax subq %fs:40, %rax jne .L16 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state movq 56(%rsp), %r9 movq 48(%rsp), %r8 movq 40(%rsp), %rcx movl %r15d, %edx movl %r14d, %esi movl %r13d, %edi call _Z28__device_stub__Z2m1iiiPdS_S_iiiPdS_S_ jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2027: .size matmult_gpu1, .-matmult_gpu1 .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z2m1iiiPdS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2055: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z2m1iiiPdS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2055: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "matmult_gpu1.hip" .globl _Z17__device_stub__m1iiiPdS_S_ # -- Begin function _Z17__device_stub__m1iiiPdS_S_ .p2align 4, 0x90 .type _Z17__device_stub__m1iiiPdS_S_,@function _Z17__device_stub__m1iiiPdS_S_: # @_Z17__device_stub__m1iiiPdS_S_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movq %rcx, 88(%rsp) movq %r8, 80(%rsp) movq %r9, 72(%rsp) leaq 20(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 88(%rsp), %rax movq %rax, 120(%rsp) leaq 80(%rsp), %rax movq %rax, 128(%rsp) leaq 72(%rsp), %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z2m1iiiPdS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z17__device_stub__m1iiiPdS_S_, .Lfunc_end0-_Z17__device_stub__m1iiiPdS_S_ .cfi_endproc # -- End function .globl matmult_gpu1 # -- Begin function matmult_gpu1 .p2align 4, 0x90 .type matmult_gpu1,@function matmult_gpu1: # @matmult_gpu1 .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $200, %rsp .cfi_def_cfa_offset 256 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movq %r9, 64(%rsp) # 8-byte Spill movq %r8, 56(%rsp) # 8-byte Spill movq %rcx, 48(%rsp) # 8-byte Spill movl %edx, %ebp movl %esi, %r15d movl %edi, %r12d movl $2, %edi callq hipSetDevice movl %ebp, %eax imull %r12d, %eax movslq %eax, %r13 shlq $3, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi callq hipMalloc movl %ebp, %eax imull %r15d, %eax movslq %eax, %rbx shlq $3, %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi callq hipMalloc movl %r15d, %eax imull %r12d, %eax movslq %eax, %r14 shlq $3, %r14 leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 24(%rsp), %rdi movq 48(%rsp), %rsi # 8-byte Reload movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 16(%rsp), %rdi movq 56(%rsp), %rsi # 8-byte Reload movq %rbx, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi xorl %esi, %esi movq %r14, %rdx callq hipMemset movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 24(%rsp), %rax movq 16(%rsp), %rcx movq 8(%rsp), %rdx movl %r12d, 44(%rsp) movl %r15d, 40(%rsp) movl %ebp, 36(%rsp) movq %rax, 136(%rsp) movq %rcx, 128(%rsp) movq %rdx, 120(%rsp) leaq 44(%rsp), %rax movq %rax, 144(%rsp) leaq 40(%rsp), %rax movq %rax, 152(%rsp) leaq 36(%rsp), %rax movq %rax, 160(%rsp) leaq 136(%rsp), %rax movq %rax, 168(%rsp) leaq 128(%rsp), %rax movq %rax, 176(%rsp) leaq 120(%rsp), %rax movq %rax, 184(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z2m1iiiPdS_S_, %edi pushq 72(%rsp) .cfi_adjust_cfa_offset 8 pushq 88(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize movq 8(%rsp), %rsi movq 64(%rsp), %rdi # 8-byte Reload movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree addq $200, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size matmult_gpu1, .Lfunc_end1-matmult_gpu1 .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z2m1iiiPdS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z2m1iiiPdS_S_,@object # @_Z2m1iiiPdS_S_ .section .rodata,"a",@progbits .globl _Z2m1iiiPdS_S_ .p2align 3, 0x0 _Z2m1iiiPdS_S_: .quad _Z17__device_stub__m1iiiPdS_S_ .size _Z2m1iiiPdS_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z2m1iiiPdS_S_" .size .L__unnamed_1, 15 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z17__device_stub__m1iiiPdS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z2m1iiiPdS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
__global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
code for sm_80 Function : _Z5inoutPsS_s .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_TID.X ; /* 0x0000000000047919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 1.1920928955078125e-07 ; /* 0x00000002ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fcc00078e0205 */ /*0050*/ LDG.E.U16 R2, [R2.64] ; /* 0x0000000602027981 */ /* 0x000ea2000c1e1500 */ /*0060*/ ULDC.U16 UR4, c[0x0][0x170] ; /* 0x00005c0000047ab9 */ /* 0x000fe20000000400 */ /*0070*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0080*/ UPRMT UR4, UR4, 0x9910, URZ ; /* 0x0000991004047896 */ /* 0x000fe2000800003f */ /*0090*/ PRMT R0, R2, 0x9910, RZ ; /* 0x0000991002007816 */ /* 0x004fca00000000ff */ /*00a0*/ IMAD R0, R0, UR4, RZ ; /* 0x0000000400007c24 */ /* 0x000fc8000f8e02ff */ /*00b0*/ IMAD.HI R0, R0, 0x431bde83, RZ ; /* 0x431bde8300007827 */ /* 0x000fca00078e02ff */ /*00c0*/ SHF.R.U32.HI R7, RZ, 0x1f, R0 ; /* 0x0000001fff077819 */ /* 0x000fc80000011600 */ /*00d0*/ LEA.HI.SX32 R7, R0, R7, 0xe ; /* 0x0000000700077211 */ /* 0x000fca00078f72ff */ /*00e0*/ STG.E.U16 [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101506 */ /*00f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0100*/ BRA 0x100; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
__global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
.file "tmpxft_00026e87_00000000-6_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z5inoutPsS_sPsS_s .type _Z27__device_stub__Z5inoutPsS_sPsS_s, @function _Z27__device_stub__Z5inoutPsS_sPsS_s: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movw %dx, 12(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z5inoutPsS_s(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z27__device_stub__Z5inoutPsS_sPsS_s, .-_Z27__device_stub__Z5inoutPsS_sPsS_s .globl _Z5inoutPsS_s .type _Z5inoutPsS_s, @function _Z5inoutPsS_s: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movswl %dx, %edx call _Z27__device_stub__Z5inoutPsS_sPsS_s addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z5inoutPsS_s, .-_Z5inoutPsS_s .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z5inoutPsS_s" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z5inoutPsS_s(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
__global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
#include <hip/hip_runtime.h> __global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> __global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5inoutPsS_s .globl _Z5inoutPsS_s .p2align 8 .type _Z5inoutPsS_s,@function _Z5inoutPsS_s: s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 1, v0 s_load_b32 s0, s[0:1], 0x10 s_waitcnt lgkmcnt(0) global_load_i16 v1, v0, s[6:7] s_sext_i32_i16 s0, s0 s_waitcnt vmcnt(0) v_mul_i32_i24_e32 v1, s0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_i32 v1, v1, 0x431bde83 v_lshrrev_b32_e32 v2, 31, v1 v_ashrrev_i32_e32 v1, 18, v1 s_delay_alu instid0(VALU_DEP_1) v_add_nc_u32_e32 v1, v1, v2 global_store_b16 v0, v1, s[4:5] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5inoutPsS_s .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 20 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z5inoutPsS_s, .Lfunc_end0-_Z5inoutPsS_s .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 2 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 20 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5inoutPsS_s .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z5inoutPsS_s.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> __global__ void inout(short int * out, short int * in, short Res) { const int i = threadIdx.x; out[i] = in[i] * Res/1000000; }
.text .file "test.hip" .globl _Z20__device_stub__inoutPsS_s # -- Begin function _Z20__device_stub__inoutPsS_s .p2align 4, 0x90 .type _Z20__device_stub__inoutPsS_s,@function _Z20__device_stub__inoutPsS_s: # @_Z20__device_stub__inoutPsS_s .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movw %dx, 14(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 14(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z5inoutPsS_s, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z20__device_stub__inoutPsS_s, .Lfunc_end0-_Z20__device_stub__inoutPsS_s .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5inoutPsS_s, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z5inoutPsS_s,@object # @_Z5inoutPsS_s .section .rodata,"a",@progbits .globl _Z5inoutPsS_s .p2align 3, 0x0 _Z5inoutPsS_s: .quad _Z20__device_stub__inoutPsS_s .size _Z5inoutPsS_s, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5inoutPsS_s" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__inoutPsS_s .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5inoutPsS_s .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z5inoutPsS_s .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_TID.X ; /* 0x0000000000047919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 1.1920928955078125e-07 ; /* 0x00000002ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fcc00078e0205 */ /*0050*/ LDG.E.U16 R2, [R2.64] ; /* 0x0000000602027981 */ /* 0x000ea2000c1e1500 */ /*0060*/ ULDC.U16 UR4, c[0x0][0x170] ; /* 0x00005c0000047ab9 */ /* 0x000fe20000000400 */ /*0070*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0080*/ UPRMT UR4, UR4, 0x9910, URZ ; /* 0x0000991004047896 */ /* 0x000fe2000800003f */ /*0090*/ PRMT R0, R2, 0x9910, RZ ; /* 0x0000991002007816 */ /* 0x004fca00000000ff */ /*00a0*/ IMAD R0, R0, UR4, RZ ; /* 0x0000000400007c24 */ /* 0x000fc8000f8e02ff */ /*00b0*/ IMAD.HI R0, R0, 0x431bde83, RZ ; /* 0x431bde8300007827 */ /* 0x000fca00078e02ff */ /*00c0*/ SHF.R.U32.HI R7, RZ, 0x1f, R0 ; /* 0x0000001fff077819 */ /* 0x000fc80000011600 */ /*00d0*/ LEA.HI.SX32 R7, R0, R7, 0xe ; /* 0x0000000700077211 */ /* 0x000fca00078f72ff */ /*00e0*/ STG.E.U16 [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101506 */ /*00f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0100*/ BRA 0x100; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5inoutPsS_s .globl _Z5inoutPsS_s .p2align 8 .type _Z5inoutPsS_s,@function _Z5inoutPsS_s: s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 1, v0 s_load_b32 s0, s[0:1], 0x10 s_waitcnt lgkmcnt(0) global_load_i16 v1, v0, s[6:7] s_sext_i32_i16 s0, s0 s_waitcnt vmcnt(0) v_mul_i32_i24_e32 v1, s0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_i32 v1, v1, 0x431bde83 v_lshrrev_b32_e32 v2, 31, v1 v_ashrrev_i32_e32 v1, 18, v1 s_delay_alu instid0(VALU_DEP_1) v_add_nc_u32_e32 v1, v1, v2 global_store_b16 v0, v1, s[4:5] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5inoutPsS_s .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 20 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z5inoutPsS_s, .Lfunc_end0-_Z5inoutPsS_s .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 2 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 20 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5inoutPsS_s .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z5inoutPsS_s.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00026e87_00000000-6_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z5inoutPsS_sPsS_s .type _Z27__device_stub__Z5inoutPsS_sPsS_s, @function _Z27__device_stub__Z5inoutPsS_sPsS_s: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movw %dx, 12(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z5inoutPsS_s(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z27__device_stub__Z5inoutPsS_sPsS_s, .-_Z27__device_stub__Z5inoutPsS_sPsS_s .globl _Z5inoutPsS_s .type _Z5inoutPsS_s, @function _Z5inoutPsS_s: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movswl %dx, %edx call _Z27__device_stub__Z5inoutPsS_sPsS_s addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z5inoutPsS_s, .-_Z5inoutPsS_s .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z5inoutPsS_s" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z5inoutPsS_s(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "test.hip" .globl _Z20__device_stub__inoutPsS_s # -- Begin function _Z20__device_stub__inoutPsS_s .p2align 4, 0x90 .type _Z20__device_stub__inoutPsS_s,@function _Z20__device_stub__inoutPsS_s: # @_Z20__device_stub__inoutPsS_s .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movw %dx, 14(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 14(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z5inoutPsS_s, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z20__device_stub__inoutPsS_s, .Lfunc_end0-_Z20__device_stub__inoutPsS_s .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5inoutPsS_s, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z5inoutPsS_s,@object # @_Z5inoutPsS_s .section .rodata,"a",@progbits .globl _Z5inoutPsS_s .p2align 3, 0x0 _Z5inoutPsS_s: .quad _Z20__device_stub__inoutPsS_s .size _Z5inoutPsS_s, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5inoutPsS_s" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__inoutPsS_s .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5inoutPsS_s .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cuda.h> #include <curand.h> #include <curand_kernel.h> #include <stdio.h> #include <ctime> __global__ void setup_random_kernel(curandState *state, int length, int offset) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curand_init((unsigned long long) clock(), idx, 0, &state[idx]); } } __global__ void get_random_array(curandState *state, int length, double *out_array) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; out_array[idx] = curand_uniform_double(&localState); } } /* * PHASE ONE */ __global__ void phase_one_shift(double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Compute block and compute its offset if (idx < length) { curandState localState = state[idx]; shift[idx] = (int) __double2int_rn( block_i[idx] + __double2int_rd(curand_uniform(&localState) * block_b[idx]) * block_n[idx]); state[idx] = localState; } } __global__ void phase_one_i(int *i, double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Choose first node if (idx < length) { curandState localState = state[idx]; i[idx] = (int) __double2int_rn(__double2int_rd(curand_uniform(&localState) * block_n[idx]) + shift[idx]); state[idx] = localState; } } __global__ void phase_one_j(int *i, int *j, double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; // Choose second node // "Without replacement" j[idx] = (int) __double2int_rn(__double2int_rd(curand_uniform(&localState) * (block_n[idx] - 1)) + shift[idx]); // Remove loops if (j[idx] >= i[idx]) { ++j[idx]; } state[idx] = localState; } } /* * PHASE TWO */ __global__ void phase_two_fill(double *phase_two_shift_fill, double *phase_two_sz_fill, double *phase_two_fill, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; phase_two_fill[idx] = phase_two_shift_fill[idx] + __double2int_rd(curand_uniform(&localState) * phase_two_sz_fill[idx]); state[idx] = localState; } } __global__ void phase_two_bulk(double *phase_two_shift_bulk, double *phase_two_sz_bulk, double *phase_two_bulk, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; phase_two_bulk[idx] = phase_two_shift_bulk[idx] + __double2int_rd(curand_uniform(&localState) * phase_two_sz_bulk[idx]); state[idx] = localState; } } __global__ void phase_two_d(double *phase_two_fill, double *phase_two_bulk, int *phase_two, double *phase_two_rd_fill, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; if (curand_uniform(&localState) < phase_two_rd_fill[idx]) { phase_two[idx] = (int) __double2int_rn(phase_two_fill[idx]); } else { phase_two[idx] = (int) __double2int_rn(phase_two_bulk[idx]); } state[idx] = localState; } }
.file "tmpxft_00028b01_00000000-6_PhasesKernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2273: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2273: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii .type _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii, @function _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii: .LFB2295: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z19setup_random_kernelP17curandStateXORWOWii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2295: .size _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii, .-_Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii .globl _Z19setup_random_kernelP17curandStateXORWOWii .type _Z19setup_random_kernelP17curandStateXORWOWii, @function _Z19setup_random_kernelP17curandStateXORWOWii: .LFB2296: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2296: .size _Z19setup_random_kernelP17curandStateXORWOWii, .-_Z19setup_random_kernelP17curandStateXORWOWii .globl _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd .type _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd, @function _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd: .LFB2297: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movl %esi, 20(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 20(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 120(%rsp), %rax subq %fs:40, %rax jne .L16 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z16get_random_arrayP17curandStateXORWOWiPd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2297: .size _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd, .-_Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd .globl _Z16get_random_arrayP17curandStateXORWOWiPd .type _Z16get_random_arrayP17curandStateXORWOWiPd, @function _Z16get_random_arrayP17curandStateXORWOWiPd: .LFB2298: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2298: .size _Z16get_random_arrayP17curandStateXORWOWiPd, .-_Z16get_random_arrayP17curandStateXORWOWiPd .globl _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi .type _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi, @function _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi: .LFB2299: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movl %r9d, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) leaq 4(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L23 .L19: movq 168(%rsp), %rax subq %fs:40, %rax jne .L24 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L19 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2299: .size _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi, .-_Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi .globl _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi .type _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi, @function _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi: .LFB2300: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2300: .size _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi, .-_Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi .globl _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi .type _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi, @function _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi: .LFB2301: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movq %r9, (%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) movq %rsp, %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L31 .L27: movq 168(%rsp), %rax subq %fs:40, %rax jne .L32 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L31: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L27 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2301: .size _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi, .-_Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi .globl _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi .type _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi, @function _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi: .LFB2302: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2302: .size _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi, .-_Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi .globl _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi .type _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi, @function _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi: .LFB2303: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movq %rcx, 32(%rsp) movq %r8, 24(%rsp) movq %r9, 16(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) leaq 232(%rsp), %rax movq %rax, 184(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $1, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) leaq 72(%rsp), %rcx leaq 64(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L39 .L35: movq 200(%rsp), %rax subq %fs:40, %rax jne .L40 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L39: .cfi_restore_state pushq 72(%rsp) .cfi_def_cfa_offset 232 pushq 72(%rsp) .cfi_def_cfa_offset 240 leaq 144(%rsp), %r9 movq 108(%rsp), %rcx movl 116(%rsp), %r8d movq 96(%rsp), %rsi movl 104(%rsp), %edx leaq _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L35 .L40: call __stack_chk_fail@PLT .cfi_endproc .LFE2303: .size _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi, .-_Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi .globl _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi .type _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi, @function _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi: .LFB2304: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2304: .size _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi, .-_Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi .globl _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .type _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, @function _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi: .LFB2305: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L47 .L43: movq 152(%rsp), %rax subq %fs:40, %rax jne .L48 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L47: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 184 pushq 56(%rsp) .cfi_def_cfa_offset 192 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14phase_two_fillPdS_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L43 .L48: call __stack_chk_fail@PLT .cfi_endproc .LFE2305: .size _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, .-_Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .globl _Z14phase_two_fillPdS_S_P17curandStateXORWOWi .type _Z14phase_two_fillPdS_S_P17curandStateXORWOWi, @function _Z14phase_two_fillPdS_S_P17curandStateXORWOWi: .LFB2306: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2306: .size _Z14phase_two_fillPdS_S_P17curandStateXORWOWi, .-_Z14phase_two_fillPdS_S_P17curandStateXORWOWi .globl _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .type _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, @function _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi: .LFB2307: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L55 .L51: movq 152(%rsp), %rax subq %fs:40, %rax jne .L56 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L55: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 184 pushq 56(%rsp) .cfi_def_cfa_offset 192 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L51 .L56: call __stack_chk_fail@PLT .cfi_endproc .LFE2307: .size _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, .-_Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .globl _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi .type _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi, @function _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi: .LFB2308: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2308: .size _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi, .-_Z14phase_two_bulkPdS_S_P17curandStateXORWOWi .globl _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi .type _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi, @function _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi: .LFB2309: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movl %r9d, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) leaq 4(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L63 .L59: movq 168(%rsp), %rax subq %fs:40, %rax jne .L64 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L63: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L59 .L64: call __stack_chk_fail@PLT .cfi_endproc .LFE2309: .size _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi, .-_Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi .globl _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi .type _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi, @function _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi: .LFB2310: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2310: .size _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi, .-_Z11phase_two_dPdS_PiS_P17curandStateXORWOWi .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z11phase_two_dPdS_PiS_P17curandStateXORWOWi" .align 8 .LC1: .string "_Z14phase_two_bulkPdS_S_P17curandStateXORWOWi" .align 8 .LC2: .string "_Z14phase_two_fillPdS_S_P17curandStateXORWOWi" .align 8 .LC3: .string "_Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi" .align 8 .LC4: .string "_Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi" .align 8 .LC5: .string "_Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi" .align 8 .LC6: .string "_Z16get_random_arrayP17curandStateXORWOWiPd" .align 8 .LC7: .string "_Z19setup_random_kernelP17curandStateXORWOWii" .section .rodata.str1.1,"aMS",@progbits,1 .LC8: .string "precalc_xorwow_matrix" .LC9: .string "precalc_xorwow_offset_matrix" .LC10: .string "mrg32k3aM1" .LC11: .string "mrg32k3aM2" .LC12: .string "mrg32k3aM1SubSeq" .LC13: .string "mrg32k3aM2SubSeq" .LC14: .string "mrg32k3aM1Seq" .LC15: .string "mrg32k3aM2Seq" .LC16: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2312: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _Z14phase_two_fillPdS_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z16get_random_arrayP17curandStateXORWOWiPd(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z19setup_random_kernelP17curandStateXORWOWii(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _ZL21precalc_xorwow_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _ZL28precalc_xorwow_offset_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM1(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM2(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM1SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC13(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM2SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC14(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM1Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM2Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $72, %r9d movl $0, %r8d leaq .LC16(%rip), %rdx movq %rdx, %rcx leaq _ZL17__cr_lgamma_table(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2312: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <cuda.h> #include <curand.h> #include <curand_kernel.h> #include <stdio.h> #include <ctime> __global__ void setup_random_kernel(curandState *state, int length, int offset) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curand_init((unsigned long long) clock(), idx, 0, &state[idx]); } } __global__ void get_random_array(curandState *state, int length, double *out_array) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; out_array[idx] = curand_uniform_double(&localState); } } /* * PHASE ONE */ __global__ void phase_one_shift(double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Compute block and compute its offset if (idx < length) { curandState localState = state[idx]; shift[idx] = (int) __double2int_rn( block_i[idx] + __double2int_rd(curand_uniform(&localState) * block_b[idx]) * block_n[idx]); state[idx] = localState; } } __global__ void phase_one_i(int *i, double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Choose first node if (idx < length) { curandState localState = state[idx]; i[idx] = (int) __double2int_rn(__double2int_rd(curand_uniform(&localState) * block_n[idx]) + shift[idx]); state[idx] = localState; } } __global__ void phase_one_j(int *i, int *j, double *block_b, double *block_i, double *block_n, int *shift, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; // Choose second node // "Without replacement" j[idx] = (int) __double2int_rn(__double2int_rd(curand_uniform(&localState) * (block_n[idx] - 1)) + shift[idx]); // Remove loops if (j[idx] >= i[idx]) { ++j[idx]; } state[idx] = localState; } } /* * PHASE TWO */ __global__ void phase_two_fill(double *phase_two_shift_fill, double *phase_two_sz_fill, double *phase_two_fill, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; phase_two_fill[idx] = phase_two_shift_fill[idx] + __double2int_rd(curand_uniform(&localState) * phase_two_sz_fill[idx]); state[idx] = localState; } } __global__ void phase_two_bulk(double *phase_two_shift_bulk, double *phase_two_sz_bulk, double *phase_two_bulk, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; phase_two_bulk[idx] = phase_two_shift_bulk[idx] + __double2int_rd(curand_uniform(&localState) * phase_two_sz_bulk[idx]); state[idx] = localState; } } __global__ void phase_two_d(double *phase_two_fill, double *phase_two_bulk, int *phase_two, double *phase_two_rd_fill, curandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { curandState localState = state[idx]; if (curand_uniform(&localState) < phase_two_rd_fill[idx]) { phase_two[idx] = (int) __double2int_rn(phase_two_fill[idx]); } else { phase_two[idx] = (int) __double2int_rn(phase_two_bulk[idx]); } state[idx] = localState; } }
#include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> #include <stdio.h> #include <ctime> __global__ void setup_random_kernel(hiprandState *state, int length, int offset) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprand_init((unsigned long long) clock(), idx, 0, &state[idx]); } } __global__ void get_random_array(hiprandState *state, int length, double *out_array) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; out_array[idx] = hiprand_uniform_double(&localState); } } /* * PHASE ONE */ __global__ void phase_one_shift(double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Compute block and compute its offset if (idx < length) { hiprandState localState = state[idx]; shift[idx] = (int) __double2int_rn( block_i[idx] + __double2int_rd(hiprand_uniform(&localState) * block_b[idx]) * block_n[idx]); state[idx] = localState; } } __global__ void phase_one_i(int *i, double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Choose first node if (idx < length) { hiprandState localState = state[idx]; i[idx] = (int) __double2int_rn(__double2int_rd(hiprand_uniform(&localState) * block_n[idx]) + shift[idx]); state[idx] = localState; } } __global__ void phase_one_j(int *i, int *j, double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; // Choose second node // "Without replacement" j[idx] = (int) __double2int_rn(__double2int_rd(hiprand_uniform(&localState) * (block_n[idx] - 1)) + shift[idx]); // Remove loops if (j[idx] >= i[idx]) { ++j[idx]; } state[idx] = localState; } } /* * PHASE TWO */ __global__ void phase_two_fill(double *phase_two_shift_fill, double *phase_two_sz_fill, double *phase_two_fill, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; phase_two_fill[idx] = phase_two_shift_fill[idx] + __double2int_rd(hiprand_uniform(&localState) * phase_two_sz_fill[idx]); state[idx] = localState; } } __global__ void phase_two_bulk(double *phase_two_shift_bulk, double *phase_two_sz_bulk, double *phase_two_bulk, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; phase_two_bulk[idx] = phase_two_shift_bulk[idx] + __double2int_rd(hiprand_uniform(&localState) * phase_two_sz_bulk[idx]); state[idx] = localState; } } __global__ void phase_two_d(double *phase_two_fill, double *phase_two_bulk, int *phase_two, double *phase_two_rd_fill, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; if (hiprand_uniform(&localState) < phase_two_rd_fill[idx]) { phase_two[idx] = (int) __double2int_rn(phase_two_fill[idx]); } else { phase_two[idx] = (int) __double2int_rn(phase_two_bulk[idx]); } state[idx] = localState; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> #include <stdio.h> #include <ctime> __global__ void setup_random_kernel(hiprandState *state, int length, int offset) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprand_init((unsigned long long) clock(), idx, 0, &state[idx]); } } __global__ void get_random_array(hiprandState *state, int length, double *out_array) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; out_array[idx] = hiprand_uniform_double(&localState); } } /* * PHASE ONE */ __global__ void phase_one_shift(double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Compute block and compute its offset if (idx < length) { hiprandState localState = state[idx]; shift[idx] = (int) __double2int_rn( block_i[idx] + __double2int_rd(hiprand_uniform(&localState) * block_b[idx]) * block_n[idx]); state[idx] = localState; } } __global__ void phase_one_i(int *i, double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; // Choose first node if (idx < length) { hiprandState localState = state[idx]; i[idx] = (int) __double2int_rn(__double2int_rd(hiprand_uniform(&localState) * block_n[idx]) + shift[idx]); state[idx] = localState; } } __global__ void phase_one_j(int *i, int *j, double *block_b, double *block_i, double *block_n, int *shift, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; // Choose second node // "Without replacement" j[idx] = (int) __double2int_rn(__double2int_rd(hiprand_uniform(&localState) * (block_n[idx] - 1)) + shift[idx]); // Remove loops if (j[idx] >= i[idx]) { ++j[idx]; } state[idx] = localState; } } /* * PHASE TWO */ __global__ void phase_two_fill(double *phase_two_shift_fill, double *phase_two_sz_fill, double *phase_two_fill, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; phase_two_fill[idx] = phase_two_shift_fill[idx] + __double2int_rd(hiprand_uniform(&localState) * phase_two_sz_fill[idx]); state[idx] = localState; } } __global__ void phase_two_bulk(double *phase_two_shift_bulk, double *phase_two_sz_bulk, double *phase_two_bulk, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; phase_two_bulk[idx] = phase_two_shift_bulk[idx] + __double2int_rd(hiprand_uniform(&localState) * phase_two_sz_bulk[idx]); state[idx] = localState; } } __global__ void phase_two_d(double *phase_two_fill, double *phase_two_bulk, int *phase_two, double *phase_two_rd_fill, hiprandState *state, int length) { int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < length) { hiprandState localState = state[idx]; if (hiprand_uniform(&localState) < phase_two_rd_fill[idx]) { phase_two[idx] = (int) __double2int_rn(phase_two_fill[idx]); } else { phase_two[idx] = (int) __double2int_rn(phase_two_bulk[idx]); } state[idx] = localState; } }
.text .file "PhasesKernel.hip" .globl _Z34__device_stub__setup_random_kernelP12hiprandStateii # -- Begin function _Z34__device_stub__setup_random_kernelP12hiprandStateii .p2align 4, 0x90 .type _Z34__device_stub__setup_random_kernelP12hiprandStateii,@function _Z34__device_stub__setup_random_kernelP12hiprandStateii: # @_Z34__device_stub__setup_random_kernelP12hiprandStateii .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 4(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z19setup_random_kernelP12hiprandStateii, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z34__device_stub__setup_random_kernelP12hiprandStateii, .Lfunc_end0-_Z34__device_stub__setup_random_kernelP12hiprandStateii .cfi_endproc # -- End function .globl _Z31__device_stub__get_random_arrayP12hiprandStateiPd # -- Begin function _Z31__device_stub__get_random_arrayP12hiprandStateiPd .p2align 4, 0x90 .type _Z31__device_stub__get_random_arrayP12hiprandStateiPd,@function _Z31__device_stub__get_random_arrayP12hiprandStateiPd: # @_Z31__device_stub__get_random_arrayP12hiprandStateiPd .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movq %rdx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 64(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z16get_random_arrayP12hiprandStateiPd, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z31__device_stub__get_random_arrayP12hiprandStateiPd, .Lfunc_end1-_Z31__device_stub__get_random_arrayP12hiprandStateiPd .cfi_endproc # -- End function .globl _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei # -- Begin function _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .p2align 4, 0x90 .type _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei,@function _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei: # @_Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movl %r9d, 4(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 4(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z15phase_one_shiftPdS_S_PiP12hiprandStatei, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end2: .size _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei, .Lfunc_end2-_Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei,@function _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei: # @_Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end3: .size _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei, .Lfunc_end3-_Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei,@function _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei: # @_Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end4: .size _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, .Lfunc_end4-_Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei # -- Begin function _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .p2align 4, 0x90 .type _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei,@function _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei: # @_Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z14phase_two_fillPdS_S_P12hiprandStatei, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end5: .size _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei, .Lfunc_end5-_Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei # -- Begin function _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .p2align 4, 0x90 .type _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei,@function _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei: # @_Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z14phase_two_bulkPdS_S_P12hiprandStatei, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end6: .size _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei, .Lfunc_end6-_Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei,@function _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei: # @_Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movl %r9d, 4(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 4(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_two_dPdS_PiS_P12hiprandStatei, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end7: .size _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei, .Lfunc_end7-_Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB8_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB8_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19setup_random_kernelP12hiprandStateii, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z16get_random_arrayP12hiprandStateiPd, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z15phase_one_shiftPdS_S_PiP12hiprandStatei, %esi movl $.L__unnamed_3, %edx movl $.L__unnamed_3, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, %esi movl $.L__unnamed_4, %edx movl $.L__unnamed_4, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, %esi movl $.L__unnamed_5, %edx movl $.L__unnamed_5, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z14phase_two_fillPdS_S_P12hiprandStatei, %esi movl $.L__unnamed_6, %edx movl $.L__unnamed_6, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z14phase_two_bulkPdS_S_P12hiprandStatei, %esi movl $.L__unnamed_7, %edx movl $.L__unnamed_7, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_two_dPdS_PiS_P12hiprandStatei, %esi movl $.L__unnamed_8, %edx movl $.L__unnamed_8, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end8: .size __hip_module_ctor, .Lfunc_end8-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB9_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB9_2: retq .Lfunc_end9: .size __hip_module_dtor, .Lfunc_end9-__hip_module_dtor .cfi_endproc # -- End function .type _Z19setup_random_kernelP12hiprandStateii,@object # @_Z19setup_random_kernelP12hiprandStateii .section .rodata,"a",@progbits .globl _Z19setup_random_kernelP12hiprandStateii .p2align 3, 0x0 _Z19setup_random_kernelP12hiprandStateii: .quad _Z34__device_stub__setup_random_kernelP12hiprandStateii .size _Z19setup_random_kernelP12hiprandStateii, 8 .type _Z16get_random_arrayP12hiprandStateiPd,@object # @_Z16get_random_arrayP12hiprandStateiPd .globl _Z16get_random_arrayP12hiprandStateiPd .p2align 3, 0x0 _Z16get_random_arrayP12hiprandStateiPd: .quad _Z31__device_stub__get_random_arrayP12hiprandStateiPd .size _Z16get_random_arrayP12hiprandStateiPd, 8 .type _Z15phase_one_shiftPdS_S_PiP12hiprandStatei,@object # @_Z15phase_one_shiftPdS_S_PiP12hiprandStatei .globl _Z15phase_one_shiftPdS_S_PiP12hiprandStatei .p2align 3, 0x0 _Z15phase_one_shiftPdS_S_PiP12hiprandStatei: .quad _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .size _Z15phase_one_shiftPdS_S_PiP12hiprandStatei, 8 .type _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei,@object # @_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .globl _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .p2align 3, 0x0 _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei: .quad _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .size _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, 8 .type _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei,@object # @_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .globl _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .p2align 3, 0x0 _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei: .quad _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .size _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, 8 .type _Z14phase_two_fillPdS_S_P12hiprandStatei,@object # @_Z14phase_two_fillPdS_S_P12hiprandStatei .globl _Z14phase_two_fillPdS_S_P12hiprandStatei .p2align 3, 0x0 _Z14phase_two_fillPdS_S_P12hiprandStatei: .quad _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .size _Z14phase_two_fillPdS_S_P12hiprandStatei, 8 .type _Z14phase_two_bulkPdS_S_P12hiprandStatei,@object # @_Z14phase_two_bulkPdS_S_P12hiprandStatei .globl _Z14phase_two_bulkPdS_S_P12hiprandStatei .p2align 3, 0x0 _Z14phase_two_bulkPdS_S_P12hiprandStatei: .quad _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .size _Z14phase_two_bulkPdS_S_P12hiprandStatei, 8 .type _Z11phase_two_dPdS_PiS_P12hiprandStatei,@object # @_Z11phase_two_dPdS_PiS_P12hiprandStatei .globl _Z11phase_two_dPdS_PiS_P12hiprandStatei .p2align 3, 0x0 _Z11phase_two_dPdS_PiS_P12hiprandStatei: .quad _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .size _Z11phase_two_dPdS_PiS_P12hiprandStatei, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z19setup_random_kernelP12hiprandStateii" .size .L__unnamed_1, 41 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z16get_random_arrayP12hiprandStateiPd" .size .L__unnamed_2, 39 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z15phase_one_shiftPdS_S_PiP12hiprandStatei" .size .L__unnamed_3, 44 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei" .size .L__unnamed_4, 44 .type .L__unnamed_5,@object # @4 .L__unnamed_5: .asciz "_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei" .size .L__unnamed_5, 46 .type .L__unnamed_6,@object # @5 .L__unnamed_6: .asciz "_Z14phase_two_fillPdS_S_P12hiprandStatei" .size .L__unnamed_6, 41 .type .L__unnamed_7,@object # @6 .L__unnamed_7: .asciz "_Z14phase_two_bulkPdS_S_P12hiprandStatei" .size .L__unnamed_7, 41 .type .L__unnamed_8,@object # @7 .L__unnamed_8: .asciz "_Z11phase_two_dPdS_PiS_P12hiprandStatei" .size .L__unnamed_8, 40 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z34__device_stub__setup_random_kernelP12hiprandStateii .addrsig_sym _Z31__device_stub__get_random_arrayP12hiprandStateiPd .addrsig_sym _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .addrsig_sym _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .addrsig_sym _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .addrsig_sym _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .addrsig_sym _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .addrsig_sym _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19setup_random_kernelP12hiprandStateii .addrsig_sym _Z16get_random_arrayP12hiprandStateiPd .addrsig_sym _Z15phase_one_shiftPdS_S_PiP12hiprandStatei .addrsig_sym _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .addrsig_sym _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .addrsig_sym _Z14phase_two_fillPdS_S_P12hiprandStatei .addrsig_sym _Z14phase_two_bulkPdS_S_P12hiprandStatei .addrsig_sym _Z11phase_two_dPdS_PiS_P12hiprandStatei .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00028b01_00000000-6_PhasesKernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2273: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2273: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii .type _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii, @function _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii: .LFB2295: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z19setup_random_kernelP17curandStateXORWOWii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2295: .size _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii, .-_Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii .globl _Z19setup_random_kernelP17curandStateXORWOWii .type _Z19setup_random_kernelP17curandStateXORWOWii, @function _Z19setup_random_kernelP17curandStateXORWOWii: .LFB2296: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z19setup_random_kernelP17curandStateXORWOWiiP17curandStateXORWOWii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2296: .size _Z19setup_random_kernelP17curandStateXORWOWii, .-_Z19setup_random_kernelP17curandStateXORWOWii .globl _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd .type _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd, @function _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd: .LFB2297: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movl %esi, 20(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 20(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 120(%rsp), %rax subq %fs:40, %rax jne .L16 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z16get_random_arrayP17curandStateXORWOWiPd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2297: .size _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd, .-_Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd .globl _Z16get_random_arrayP17curandStateXORWOWiPd .type _Z16get_random_arrayP17curandStateXORWOWiPd, @function _Z16get_random_arrayP17curandStateXORWOWiPd: .LFB2298: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z57__device_stub__Z16get_random_arrayP17curandStateXORWOWiPdP17curandStateXORWOWiPd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2298: .size _Z16get_random_arrayP17curandStateXORWOWiPd, .-_Z16get_random_arrayP17curandStateXORWOWiPd .globl _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi .type _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi, @function _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi: .LFB2299: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movl %r9d, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) leaq 4(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L23 .L19: movq 168(%rsp), %rax subq %fs:40, %rax jne .L24 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L19 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2299: .size _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi, .-_Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi .globl _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi .type _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi, @function _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi: .LFB2300: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z62__device_stub__Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWiPdS_S_PiP17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2300: .size _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi, .-_Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi .globl _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi .type _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi, @function _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi: .LFB2301: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movq %r9, (%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) movq %rsp, %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L31 .L27: movq 168(%rsp), %rax subq %fs:40, %rax jne .L32 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L31: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L27 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2301: .size _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi, .-_Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi .globl _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi .type _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi, @function _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi: .LFB2302: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z62__device_stub__Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWiPiPdS0_S0_S_P17curandStateXORWOWi addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2302: .size _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi, .-_Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi .globl _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi .type _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi, @function _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi: .LFB2303: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movq %rcx, 32(%rsp) movq %r8, 24(%rsp) movq %r9, 16(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) leaq 232(%rsp), %rax movq %rax, 184(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $1, 92(%rsp) movl $1, 96(%rsp) movl $1, 100(%rsp) leaq 72(%rsp), %rcx leaq 64(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L39 .L35: movq 200(%rsp), %rax subq %fs:40, %rax jne .L40 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L39: .cfi_restore_state pushq 72(%rsp) .cfi_def_cfa_offset 232 pushq 72(%rsp) .cfi_def_cfa_offset 240 leaq 144(%rsp), %r9 movq 108(%rsp), %rcx movl 116(%rsp), %r8d movq 96(%rsp), %rsi movl 104(%rsp), %edx leaq _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L35 .L40: call __stack_chk_fail@PLT .cfi_endproc .LFE2303: .size _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi, .-_Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi .globl _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi .type _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi, @function _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi: .LFB2304: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z64__device_stub__Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWiPiS_PdS0_S0_S_P17curandStateXORWOWi addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2304: .size _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi, .-_Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi .globl _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .type _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, @function _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi: .LFB2305: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L47 .L43: movq 152(%rsp), %rax subq %fs:40, %rax jne .L48 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L47: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 184 pushq 56(%rsp) .cfi_def_cfa_offset 192 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14phase_two_fillPdS_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L43 .L48: call __stack_chk_fail@PLT .cfi_endproc .LFE2305: .size _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, .-_Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .globl _Z14phase_two_fillPdS_S_P17curandStateXORWOWi .type _Z14phase_two_fillPdS_S_P17curandStateXORWOWi, @function _Z14phase_two_fillPdS_S_P17curandStateXORWOWi: .LFB2306: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z14phase_two_fillPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2306: .size _Z14phase_two_fillPdS_S_P17curandStateXORWOWi, .-_Z14phase_two_fillPdS_S_P17curandStateXORWOWi .globl _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .type _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, @function _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi: .LFB2307: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L55 .L51: movq 152(%rsp), %rax subq %fs:40, %rax jne .L56 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L55: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 184 pushq 56(%rsp) .cfi_def_cfa_offset 192 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L51 .L56: call __stack_chk_fail@PLT .cfi_endproc .LFE2307: .size _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi, .-_Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi .globl _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi .type _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi, @function _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi: .LFB2308: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z59__device_stub__Z14phase_two_bulkPdS_S_P17curandStateXORWOWiPdS_S_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2308: .size _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi, .-_Z14phase_two_bulkPdS_S_P17curandStateXORWOWi .globl _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi .type _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi, @function _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi: .LFB2309: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movq %r8, 8(%rsp) movl %r9d, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 8(%rsp), %rax movq %rax, 144(%rsp) leaq 4(%rsp), %rax movq %rax, 152(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L63 .L59: movq 168(%rsp), %rax subq %fs:40, %rax jne .L64 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L63: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L59 .L64: call __stack_chk_fail@PLT .cfi_endproc .LFE2309: .size _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi, .-_Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi .globl _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi .type _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi, @function _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi: .LFB2310: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z58__device_stub__Z11phase_two_dPdS_PiS_P17curandStateXORWOWiPdS_PiS_P17curandStateXORWOWi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2310: .size _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi, .-_Z11phase_two_dPdS_PiS_P17curandStateXORWOWi .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z11phase_two_dPdS_PiS_P17curandStateXORWOWi" .align 8 .LC1: .string "_Z14phase_two_bulkPdS_S_P17curandStateXORWOWi" .align 8 .LC2: .string "_Z14phase_two_fillPdS_S_P17curandStateXORWOWi" .align 8 .LC3: .string "_Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi" .align 8 .LC4: .string "_Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi" .align 8 .LC5: .string "_Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi" .align 8 .LC6: .string "_Z16get_random_arrayP17curandStateXORWOWiPd" .align 8 .LC7: .string "_Z19setup_random_kernelP17curandStateXORWOWii" .section .rodata.str1.1,"aMS",@progbits,1 .LC8: .string "precalc_xorwow_matrix" .LC9: .string "precalc_xorwow_offset_matrix" .LC10: .string "mrg32k3aM1" .LC11: .string "mrg32k3aM2" .LC12: .string "mrg32k3aM1SubSeq" .LC13: .string "mrg32k3aM2SubSeq" .LC14: .string "mrg32k3aM1Seq" .LC15: .string "mrg32k3aM2Seq" .LC16: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2312: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_two_dPdS_PiS_P17curandStateXORWOWi(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z14phase_two_bulkPdS_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _Z14phase_two_fillPdS_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_one_jPiS_PdS0_S0_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z11phase_one_iPiPdS0_S0_S_P17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z15phase_one_shiftPdS_S_PiP17curandStateXORWOWi(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z16get_random_arrayP17curandStateXORWOWiPd(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z19setup_random_kernelP17curandStateXORWOWii(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _ZL21precalc_xorwow_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _ZL28precalc_xorwow_offset_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM1(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM2(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM1SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC13(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM2SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC14(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM1Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM2Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $72, %r9d movl $0, %r8d leaq .LC16(%rip), %rdx movq %rdx, %rcx leaq _ZL17__cr_lgamma_table(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2312: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "PhasesKernel.hip" .globl _Z34__device_stub__setup_random_kernelP12hiprandStateii # -- Begin function _Z34__device_stub__setup_random_kernelP12hiprandStateii .p2align 4, 0x90 .type _Z34__device_stub__setup_random_kernelP12hiprandStateii,@function _Z34__device_stub__setup_random_kernelP12hiprandStateii: # @_Z34__device_stub__setup_random_kernelP12hiprandStateii .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 4(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z19setup_random_kernelP12hiprandStateii, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z34__device_stub__setup_random_kernelP12hiprandStateii, .Lfunc_end0-_Z34__device_stub__setup_random_kernelP12hiprandStateii .cfi_endproc # -- End function .globl _Z31__device_stub__get_random_arrayP12hiprandStateiPd # -- Begin function _Z31__device_stub__get_random_arrayP12hiprandStateiPd .p2align 4, 0x90 .type _Z31__device_stub__get_random_arrayP12hiprandStateiPd,@function _Z31__device_stub__get_random_arrayP12hiprandStateiPd: # @_Z31__device_stub__get_random_arrayP12hiprandStateiPd .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movq %rdx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 64(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z16get_random_arrayP12hiprandStateiPd, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z31__device_stub__get_random_arrayP12hiprandStateiPd, .Lfunc_end1-_Z31__device_stub__get_random_arrayP12hiprandStateiPd .cfi_endproc # -- End function .globl _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei # -- Begin function _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .p2align 4, 0x90 .type _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei,@function _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei: # @_Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movl %r9d, 4(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 4(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z15phase_one_shiftPdS_S_PiP12hiprandStatei, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end2: .size _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei, .Lfunc_end2-_Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei,@function _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei: # @_Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end3: .size _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei, .Lfunc_end3-_Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei,@function _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei: # @_Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end4: .size _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, .Lfunc_end4-_Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei # -- Begin function _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .p2align 4, 0x90 .type _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei,@function _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei: # @_Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z14phase_two_fillPdS_S_P12hiprandStatei, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end5: .size _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei, .Lfunc_end5-_Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei # -- Begin function _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .p2align 4, 0x90 .type _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei,@function _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei: # @_Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z14phase_two_bulkPdS_S_P12hiprandStatei, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end6: .size _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei, .Lfunc_end6-_Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .cfi_endproc # -- End function .globl _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei # -- Begin function _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .p2align 4, 0x90 .type _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei,@function _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei: # @_Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movl %r9d, 4(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 4(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z11phase_two_dPdS_PiS_P12hiprandStatei, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end7: .size _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei, .Lfunc_end7-_Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB8_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB8_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19setup_random_kernelP12hiprandStateii, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z16get_random_arrayP12hiprandStateiPd, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z15phase_one_shiftPdS_S_PiP12hiprandStatei, %esi movl $.L__unnamed_3, %edx movl $.L__unnamed_3, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, %esi movl $.L__unnamed_4, %edx movl $.L__unnamed_4, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, %esi movl $.L__unnamed_5, %edx movl $.L__unnamed_5, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z14phase_two_fillPdS_S_P12hiprandStatei, %esi movl $.L__unnamed_6, %edx movl $.L__unnamed_6, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z14phase_two_bulkPdS_S_P12hiprandStatei, %esi movl $.L__unnamed_7, %edx movl $.L__unnamed_7, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11phase_two_dPdS_PiS_P12hiprandStatei, %esi movl $.L__unnamed_8, %edx movl $.L__unnamed_8, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end8: .size __hip_module_ctor, .Lfunc_end8-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB9_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB9_2: retq .Lfunc_end9: .size __hip_module_dtor, .Lfunc_end9-__hip_module_dtor .cfi_endproc # -- End function .type _Z19setup_random_kernelP12hiprandStateii,@object # @_Z19setup_random_kernelP12hiprandStateii .section .rodata,"a",@progbits .globl _Z19setup_random_kernelP12hiprandStateii .p2align 3, 0x0 _Z19setup_random_kernelP12hiprandStateii: .quad _Z34__device_stub__setup_random_kernelP12hiprandStateii .size _Z19setup_random_kernelP12hiprandStateii, 8 .type _Z16get_random_arrayP12hiprandStateiPd,@object # @_Z16get_random_arrayP12hiprandStateiPd .globl _Z16get_random_arrayP12hiprandStateiPd .p2align 3, 0x0 _Z16get_random_arrayP12hiprandStateiPd: .quad _Z31__device_stub__get_random_arrayP12hiprandStateiPd .size _Z16get_random_arrayP12hiprandStateiPd, 8 .type _Z15phase_one_shiftPdS_S_PiP12hiprandStatei,@object # @_Z15phase_one_shiftPdS_S_PiP12hiprandStatei .globl _Z15phase_one_shiftPdS_S_PiP12hiprandStatei .p2align 3, 0x0 _Z15phase_one_shiftPdS_S_PiP12hiprandStatei: .quad _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .size _Z15phase_one_shiftPdS_S_PiP12hiprandStatei, 8 .type _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei,@object # @_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .globl _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .p2align 3, 0x0 _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei: .quad _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .size _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei, 8 .type _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei,@object # @_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .globl _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .p2align 3, 0x0 _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei: .quad _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .size _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei, 8 .type _Z14phase_two_fillPdS_S_P12hiprandStatei,@object # @_Z14phase_two_fillPdS_S_P12hiprandStatei .globl _Z14phase_two_fillPdS_S_P12hiprandStatei .p2align 3, 0x0 _Z14phase_two_fillPdS_S_P12hiprandStatei: .quad _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .size _Z14phase_two_fillPdS_S_P12hiprandStatei, 8 .type _Z14phase_two_bulkPdS_S_P12hiprandStatei,@object # @_Z14phase_two_bulkPdS_S_P12hiprandStatei .globl _Z14phase_two_bulkPdS_S_P12hiprandStatei .p2align 3, 0x0 _Z14phase_two_bulkPdS_S_P12hiprandStatei: .quad _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .size _Z14phase_two_bulkPdS_S_P12hiprandStatei, 8 .type _Z11phase_two_dPdS_PiS_P12hiprandStatei,@object # @_Z11phase_two_dPdS_PiS_P12hiprandStatei .globl _Z11phase_two_dPdS_PiS_P12hiprandStatei .p2align 3, 0x0 _Z11phase_two_dPdS_PiS_P12hiprandStatei: .quad _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .size _Z11phase_two_dPdS_PiS_P12hiprandStatei, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z19setup_random_kernelP12hiprandStateii" .size .L__unnamed_1, 41 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z16get_random_arrayP12hiprandStateiPd" .size .L__unnamed_2, 39 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z15phase_one_shiftPdS_S_PiP12hiprandStatei" .size .L__unnamed_3, 44 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei" .size .L__unnamed_4, 44 .type .L__unnamed_5,@object # @4 .L__unnamed_5: .asciz "_Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei" .size .L__unnamed_5, 46 .type .L__unnamed_6,@object # @5 .L__unnamed_6: .asciz "_Z14phase_two_fillPdS_S_P12hiprandStatei" .size .L__unnamed_6, 41 .type .L__unnamed_7,@object # @6 .L__unnamed_7: .asciz "_Z14phase_two_bulkPdS_S_P12hiprandStatei" .size .L__unnamed_7, 41 .type .L__unnamed_8,@object # @7 .L__unnamed_8: .asciz "_Z11phase_two_dPdS_PiS_P12hiprandStatei" .size .L__unnamed_8, 40 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z34__device_stub__setup_random_kernelP12hiprandStateii .addrsig_sym _Z31__device_stub__get_random_arrayP12hiprandStateiPd .addrsig_sym _Z30__device_stub__phase_one_shiftPdS_S_PiP12hiprandStatei .addrsig_sym _Z26__device_stub__phase_one_iPiPdS0_S0_S_P12hiprandStatei .addrsig_sym _Z26__device_stub__phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .addrsig_sym _Z29__device_stub__phase_two_fillPdS_S_P12hiprandStatei .addrsig_sym _Z29__device_stub__phase_two_bulkPdS_S_P12hiprandStatei .addrsig_sym _Z26__device_stub__phase_two_dPdS_PiS_P12hiprandStatei .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19setup_random_kernelP12hiprandStateii .addrsig_sym _Z16get_random_arrayP12hiprandStateiPd .addrsig_sym _Z15phase_one_shiftPdS_S_PiP12hiprandStatei .addrsig_sym _Z11phase_one_iPiPdS0_S0_S_P12hiprandStatei .addrsig_sym _Z11phase_one_jPiS_PdS0_S0_S_P12hiprandStatei .addrsig_sym _Z14phase_two_fillPdS_S_P12hiprandStatei .addrsig_sym _Z14phase_two_bulkPdS_S_P12hiprandStatei .addrsig_sym _Z11phase_two_dPdS_PiS_P12hiprandStatei .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
code for sm_80 Function : slice_sparse_dense_row .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0040*/ IADD3 R2, R0, c[0x0][0x180], RZ ; /* 0x0000600000027a10 */ /* 0x000fc80007ffe0ff */ /*0050*/ ISETP.GT.AND P0, PT, R2, c[0x0][0x184], PT ; /* 0x0000610002007a0c */ /* 0x000fda0003f04270 */ /*0060*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0070*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fe200078e00ff */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R7, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fca00078e0207 */ /*00a0*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R9, [R2.64+0x4] ; /* 0x0000040402097981 */ /* 0x000ea4000c1e1900 */ /*00c0*/ ISETP.GE.AND P0, PT, R4, R9, PT ; /* 0x000000090400720c */ /* 0x004fda0003f06270 */ /*00d0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00e0*/ MOV R10, R4 ; /* 0x00000004000a7202 */ /* 0x000fe20000000f00 */ /*00f0*/ IMAD.MOV.U32 R5, RZ, RZ, 0x8 ; /* 0x00000008ff057424 */ /* 0x000fe200078e00ff */ /*0100*/ MOV R11, c[0x0][0x190] ; /* 0x00006400000b7a02 */ /* 0x000fc60000000f00 */ /*0110*/ IMAD.WIDE R4, R10, R5, c[0x0][0x160] ; /* 0x000058000a047625 */ /* 0x000fc800078e0205 */ /*0120*/ IMAD.WIDE R6, R10, R7, c[0x0][0x170] ; /* 0x00005c000a067625 */ /* 0x000fe200078e0207 */ /*0130*/ MOV R15, R5 ; /* 0x00000005000f7202 */ /* 0x000fc60000000f00 */ /*0140*/ IMAD R0, R0, R11, -c[0x0][0x188] ; /* 0x8000620000007624 */ /* 0x000fe200078e020b */ /*0150*/ MOV R13, R7 ; /* 0x00000007000d7202 */ /* 0x000fe20000000f00 */ /*0160*/ IMAD.MOV.U32 R12, RZ, RZ, R4 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0004 */ /*0170*/ IMAD.MOV.U32 R8, RZ, RZ, R6 ; /* 0x000000ffff087224 */ /* 0x000fe400078e0006 */ /*0180*/ IMAD.MOV.U32 R11, RZ, RZ, R9 ; /* 0x000000ffff0b7224 */ /* 0x000fe400078e0009 */ /*0190*/ MOV R9, R13 ; /* 0x0000000d00097202 */ /* 0x000fca0000000f00 */ /*01a0*/ LDG.E R7, [R8.64] ; /* 0x0000000408077981 */ /* 0x000ea4000c1e1900 */ /*01b0*/ ISETP.GT.AND P0, PT, R7, c[0x0][0x18c], PT ; /* 0x0000630007007a0c */ /* 0x004fc80003f04270 */ /*01c0*/ ISETP.LT.OR P0, PT, R7, c[0x0][0x188], P0 ; /* 0x0000620007007a0c */ /* 0x000fda0000701670 */ /*01d0*/ @!P0 IMAD.MOV.U32 R4, RZ, RZ, R12 ; /* 0x000000ffff048224 */ /* 0x000fe200078e000c */ /*01e0*/ @!P0 MOV R5, R15 ; /* 0x0000000f00058202 */ /* 0x000fcc0000000f00 */ /*01f0*/ @!P0 LDG.E.64 R4, [R4.64] ; /* 0x0000000404048981 */ /* 0x000ea2000c1e1b00 */ /*0200*/ @!P0 IMAD.IADD R6, R0, 0x1, R7 ; /* 0x0000000100068824 */ /* 0x000fe200078e0207 */ /*0210*/ @!P0 MOV R7, 0x8 ; /* 0x0000000800078802 */ /* 0x000fca0000000f00 */ /*0220*/ @!P0 IMAD.WIDE R6, R6, R7, c[0x0][0x178] ; /* 0x00005e0006068625 */ /* 0x000fca00078e0207 */ /*0230*/ @!P0 STG.E.64 [R6.64], R4 ; /* 0x0000000406008986 */ /* 0x0041e8000c101b04 */ /*0240*/ @!P0 LDG.E R11, [R2.64+0x4] ; /* 0x00000404020b8981 */ /* 0x000ea2000c1e1900 */ /*0250*/ IADD3 R10, R10, 0x1, RZ ; /* 0x000000010a0a7810 */ /* 0x000fe40007ffe0ff */ /*0260*/ IADD3 R8, P2, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x000fe40007f5e0ff */ /*0270*/ IADD3 R12, P1, R12, 0x8, RZ ; /* 0x000000080c0c7810 */ /* 0x000fc60007f3e0ff */ /*0280*/ IMAD.X R13, RZ, RZ, R13, P2 ; /* 0x000000ffff0d7224 */ /* 0x000fe200010e060d */ /*0290*/ IADD3.X R15, RZ, R15, RZ, P1, !PT ; /* 0x0000000fff0f7210 */ /* 0x000fe40000ffe4ff */ /*02a0*/ ISETP.GE.AND P0, PT, R10, R11, PT ; /* 0x0000000b0a00720c */ /* 0x004fda0003f06270 */ /*02b0*/ @!P0 BRA 0x190 ; /* 0xfffffed000008947 */ /* 0x001fea000383ffff */ /*02c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02d0*/ BRA 0x2d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*02e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0300*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0310*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0320*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0330*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0340*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0350*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0360*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0370*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
.file "tmpxft_00053c34_00000000-6_slice_sparse_dense_row.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii .type _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii, @function _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 216(%rsp), %rax movq %rax, 168(%rsp) leaq 224(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq slice_sparse_dense_row(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii, .-_Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii .globl slice_sparse_dense_row .type slice_sparse_dense_row, @function slice_sparse_dense_row: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 call _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii addq $40, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size slice_sparse_dense_row, .-slice_sparse_dense_row .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "slice_sparse_dense_row" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq slice_sparse_dense_row(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
#include <hip/hip_runtime.h> #include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected slice_sparse_dense_row .globl slice_sparse_dense_row .p2align 8 .type slice_sparse_dense_row,@function slice_sparse_dense_row: s_clause 0x1 s_load_b32 s4, s[0:1], 0x44 s_load_b64 s[2:3], s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, s15, s4, v[0:1] v_add_nc_u32_e32 v0, s2, v2 s_mov_b32 s2, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_ge_i32_e64 s3, v0 s_cbranch_execz .LBB0_6 s_load_b64 s[2:3], s[0:1], 0x8 v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo global_load_b64 v[0:1], v[0:1], off s_waitcnt vmcnt(0) v_cmp_lt_i32_e32 vcc_lo, v0, v1 s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_6 s_clause 0x3 s_load_b32 s8, s[0:1], 0x30 s_load_b128 s[4:7], s[0:1], 0x10 s_load_b64 s[2:3], s[0:1], 0x28 s_load_b64 s[0:1], s[0:1], 0x0 v_ashrrev_i32_e32 v4, 31, v0 v_mov_b32_e32 v3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_lshlrev_b64 v[5:6], 2, v[3:4] v_lshlrev_b64 v[7:8], 3, v[3:4] s_waitcnt lgkmcnt(0) v_mul_lo_u32 v9, v2, s8 v_add_co_u32 v2, vcc_lo, s4, v5 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v6, vcc_lo v_add_co_u32 v4, vcc_lo, s0, v7 v_add_co_ci_u32_e32 v5, vcc_lo, s1, v8, vcc_lo v_subrev_nc_u32_e32 v6, s2, v9 s_mov_b32 s1, 0 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_4 .p2align 6 .LBB0_3: s_or_b32 exec_lo, exec_lo, s0 v_add_nc_u32_e32 v0, 1, v0 v_add_co_u32 v2, vcc_lo, v2, 4 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_cmp_ge_i32_e32 vcc_lo, v0, v1 v_add_co_u32 v4, s0, v4, 8 v_add_co_ci_u32_e64 v5, s0, 0, v5, s0 s_or_b32 s1, vcc_lo, s1 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s1 s_cbranch_execz .LBB0_6 .LBB0_4: global_load_b32 v7, v[2:3], off s_waitcnt vmcnt(0) v_cmp_le_i32_e32 vcc_lo, s2, v7 v_cmp_ge_i32_e64 s0, s3, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s4, vcc_lo, s0 s_and_saveexec_b32 s0, s4 s_cbranch_execz .LBB0_3 global_load_b64 v[8:9], v[4:5], off v_add_nc_u32_e32 v10, v6, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v11, 31, v10 v_lshlrev_b64 v[10:11], 3, v[10:11] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v10, vcc_lo, s6, v10 v_add_co_ci_u32_e32 v11, vcc_lo, s7, v11, vcc_lo s_waitcnt vmcnt(0) global_store_b64 v[10:11], v[8:9], off s_branch .LBB0_3 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel slice_sparse_dense_row .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 312 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 12 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size slice_sparse_dense_row, .Lfunc_end0-slice_sparse_dense_row .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: by_value - .offset: 44 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: by_value - .offset: 56 .size: 4 .value_kind: hidden_block_count_x - .offset: 60 .size: 4 .value_kind: hidden_block_count_y - .offset: 64 .size: 4 .value_kind: hidden_block_count_z - .offset: 68 .size: 2 .value_kind: hidden_group_size_x - .offset: 70 .size: 2 .value_kind: hidden_group_size_y - .offset: 72 .size: 2 .value_kind: hidden_group_size_z - .offset: 74 .size: 2 .value_kind: hidden_remainder_x - .offset: 76 .size: 2 .value_kind: hidden_remainder_y - .offset: 78 .size: 2 .value_kind: hidden_remainder_z - .offset: 96 .size: 8 .value_kind: hidden_global_offset_x - .offset: 104 .size: 8 .value_kind: hidden_global_offset_y - .offset: 112 .size: 8 .value_kind: hidden_global_offset_z - .offset: 120 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 312 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: slice_sparse_dense_row .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: slice_sparse_dense_row.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 12 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ /********************************** When updating a kernel or adding a new one, please compile the ptx file and commit it: nvcc -ptx -arch=sm_30 SystemML.cu ***********************************/ /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: rows of output matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is sparse and the output matrix is dense. * This function avoids unnecessary sparse to dense conversion of the input matrix. * Parallelization: subset of number of non-zeroes of input matrix. * * @params inVal input val pointer * @params inRowPtr input row pointer * @params colInd input col index pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param retClen number of columns of output matrix */ extern "C" /** * Performs a slice operation where the input matrix is dense and the output matrix is dense. * * @params in dense input pointer * @params ret dense output pointer * @param rl row lower * @param ru row upper * @param cl column lower * @param cu column upper * @param inClen number of columns of input matrix * @param retRlen number of rows of output matrix * @param retClen number of columns of output matrix */ extern "C" /** * Does a copy of upper to lower triangle of the given matrix * @param ret the input and output array allocated on the GPU * @param dim the number of rows of the square matrix ret * @param N total number of elements of the matrix */ extern "C" extern "C" __global__ void slice_sparse_dense_row(double* inVal, int* inRowPtr, int* colInd, double* ret, int rl, int ru, int cl, int cu, int retClen) { int index = blockIdx.x * blockDim.x + threadIdx.x; int rowIndex = index + rl; if (rowIndex <= ru){ /* * TODO: Alternative approach: use dynamic parallelism. We are skipping this for now to avoid * the complexity of two-step separate compilation and linking process. * * extern "C" * __global__ void slice_sparse_dense_row_helper(double* inVal, int* inRowPtr, int* colInd, double* ret, * int rl, int ru, int cl, int cu, int retClen, int start, int end, int index) { * int i = blockIdx.x * blockDim.x + threadIdx.x + start; * // Only slice if the index falls into the given range * if(i < end && cl <= colInd[i] && colInd[i] <= cu) { * ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; * } * } * * int size = inRowPtr[rowIndex+1] - inRowPtr[rowIndex]; * double numThreads = (double)min(size, MAX_NUM_THREADS_CHILD_KERNEL); * slice_sparse_dense_row_helper<<< ceil(numThreads/ MAX_NUM_THREADS_CHILD_KERNEL), MAX_NUM_THREADS_CHILD_KERNEL>>>(inVal, inRowPtr, colInd, ret, * rl, ru, cl, cu, retClen, inRowPtr[rowIndex], inRowPtr[rowIndex+1], index); * * Two-step compilation and linking process in JCudaKernels's constructor: * cuLinkAddFile(linkState, CUjitInputType.CU_JIT_INPUT_LIBRARY, "/usr/local/cuda/lib64/libcudadevrt.a", jitOptions); */ // Iterate over elements of the row 'rowIndex'. for(int i = inRowPtr[rowIndex]; i < inRowPtr[rowIndex+1]; i++) { // Only slice if the index falls into the given range if(cl <= colInd[i] && colInd[i] <= cu) { ret[ index*retClen + (colInd[i] - cl) ] = inVal[i]; } } } }
.text .file "slice_sparse_dense_row.hip" .globl __device_stub__slice_sparse_dense_row # -- Begin function __device_stub__slice_sparse_dense_row .p2align 4, 0x90 .type __device_stub__slice_sparse_dense_row,@function __device_stub__slice_sparse_dense_row: # @__device_stub__slice_sparse_dense_row .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $slice_sparse_dense_row, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size __device_stub__slice_sparse_dense_row, .Lfunc_end0-__device_stub__slice_sparse_dense_row .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $slice_sparse_dense_row, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type slice_sparse_dense_row,@object # @slice_sparse_dense_row .section .rodata,"a",@progbits .globl slice_sparse_dense_row .p2align 3, 0x0 slice_sparse_dense_row: .quad __device_stub__slice_sparse_dense_row .size slice_sparse_dense_row, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "slice_sparse_dense_row" .size .L__unnamed_1, 23 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __device_stub__slice_sparse_dense_row .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym slice_sparse_dense_row .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : slice_sparse_dense_row .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0040*/ IADD3 R2, R0, c[0x0][0x180], RZ ; /* 0x0000600000027a10 */ /* 0x000fc80007ffe0ff */ /*0050*/ ISETP.GT.AND P0, PT, R2, c[0x0][0x184], PT ; /* 0x0000610002007a0c */ /* 0x000fda0003f04270 */ /*0060*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0070*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fe200078e00ff */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R7, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fca00078e0207 */ /*00a0*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R9, [R2.64+0x4] ; /* 0x0000040402097981 */ /* 0x000ea4000c1e1900 */ /*00c0*/ ISETP.GE.AND P0, PT, R4, R9, PT ; /* 0x000000090400720c */ /* 0x004fda0003f06270 */ /*00d0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00e0*/ MOV R10, R4 ; /* 0x00000004000a7202 */ /* 0x000fe20000000f00 */ /*00f0*/ IMAD.MOV.U32 R5, RZ, RZ, 0x8 ; /* 0x00000008ff057424 */ /* 0x000fe200078e00ff */ /*0100*/ MOV R11, c[0x0][0x190] ; /* 0x00006400000b7a02 */ /* 0x000fc60000000f00 */ /*0110*/ IMAD.WIDE R4, R10, R5, c[0x0][0x160] ; /* 0x000058000a047625 */ /* 0x000fc800078e0205 */ /*0120*/ IMAD.WIDE R6, R10, R7, c[0x0][0x170] ; /* 0x00005c000a067625 */ /* 0x000fe200078e0207 */ /*0130*/ MOV R15, R5 ; /* 0x00000005000f7202 */ /* 0x000fc60000000f00 */ /*0140*/ IMAD R0, R0, R11, -c[0x0][0x188] ; /* 0x8000620000007624 */ /* 0x000fe200078e020b */ /*0150*/ MOV R13, R7 ; /* 0x00000007000d7202 */ /* 0x000fe20000000f00 */ /*0160*/ IMAD.MOV.U32 R12, RZ, RZ, R4 ; /* 0x000000ffff0c7224 */ /* 0x000fe400078e0004 */ /*0170*/ IMAD.MOV.U32 R8, RZ, RZ, R6 ; /* 0x000000ffff087224 */ /* 0x000fe400078e0006 */ /*0180*/ IMAD.MOV.U32 R11, RZ, RZ, R9 ; /* 0x000000ffff0b7224 */ /* 0x000fe400078e0009 */ /*0190*/ MOV R9, R13 ; /* 0x0000000d00097202 */ /* 0x000fca0000000f00 */ /*01a0*/ LDG.E R7, [R8.64] ; /* 0x0000000408077981 */ /* 0x000ea4000c1e1900 */ /*01b0*/ ISETP.GT.AND P0, PT, R7, c[0x0][0x18c], PT ; /* 0x0000630007007a0c */ /* 0x004fc80003f04270 */ /*01c0*/ ISETP.LT.OR P0, PT, R7, c[0x0][0x188], P0 ; /* 0x0000620007007a0c */ /* 0x000fda0000701670 */ /*01d0*/ @!P0 IMAD.MOV.U32 R4, RZ, RZ, R12 ; /* 0x000000ffff048224 */ /* 0x000fe200078e000c */ /*01e0*/ @!P0 MOV R5, R15 ; /* 0x0000000f00058202 */ /* 0x000fcc0000000f00 */ /*01f0*/ @!P0 LDG.E.64 R4, [R4.64] ; /* 0x0000000404048981 */ /* 0x000ea2000c1e1b00 */ /*0200*/ @!P0 IMAD.IADD R6, R0, 0x1, R7 ; /* 0x0000000100068824 */ /* 0x000fe200078e0207 */ /*0210*/ @!P0 MOV R7, 0x8 ; /* 0x0000000800078802 */ /* 0x000fca0000000f00 */ /*0220*/ @!P0 IMAD.WIDE R6, R6, R7, c[0x0][0x178] ; /* 0x00005e0006068625 */ /* 0x000fca00078e0207 */ /*0230*/ @!P0 STG.E.64 [R6.64], R4 ; /* 0x0000000406008986 */ /* 0x0041e8000c101b04 */ /*0240*/ @!P0 LDG.E R11, [R2.64+0x4] ; /* 0x00000404020b8981 */ /* 0x000ea2000c1e1900 */ /*0250*/ IADD3 R10, R10, 0x1, RZ ; /* 0x000000010a0a7810 */ /* 0x000fe40007ffe0ff */ /*0260*/ IADD3 R8, P2, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x000fe40007f5e0ff */ /*0270*/ IADD3 R12, P1, R12, 0x8, RZ ; /* 0x000000080c0c7810 */ /* 0x000fc60007f3e0ff */ /*0280*/ IMAD.X R13, RZ, RZ, R13, P2 ; /* 0x000000ffff0d7224 */ /* 0x000fe200010e060d */ /*0290*/ IADD3.X R15, RZ, R15, RZ, P1, !PT ; /* 0x0000000fff0f7210 */ /* 0x000fe40000ffe4ff */ /*02a0*/ ISETP.GE.AND P0, PT, R10, R11, PT ; /* 0x0000000b0a00720c */ /* 0x004fda0003f06270 */ /*02b0*/ @!P0 BRA 0x190 ; /* 0xfffffed000008947 */ /* 0x001fea000383ffff */ /*02c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02d0*/ BRA 0x2d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*02e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0300*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0310*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0320*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0330*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0340*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0350*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0360*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0370*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected slice_sparse_dense_row .globl slice_sparse_dense_row .p2align 8 .type slice_sparse_dense_row,@function slice_sparse_dense_row: s_clause 0x1 s_load_b32 s4, s[0:1], 0x44 s_load_b64 s[2:3], s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, s15, s4, v[0:1] v_add_nc_u32_e32 v0, s2, v2 s_mov_b32 s2, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_ge_i32_e64 s3, v0 s_cbranch_execz .LBB0_6 s_load_b64 s[2:3], s[0:1], 0x8 v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo global_load_b64 v[0:1], v[0:1], off s_waitcnt vmcnt(0) v_cmp_lt_i32_e32 vcc_lo, v0, v1 s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_6 s_clause 0x3 s_load_b32 s8, s[0:1], 0x30 s_load_b128 s[4:7], s[0:1], 0x10 s_load_b64 s[2:3], s[0:1], 0x28 s_load_b64 s[0:1], s[0:1], 0x0 v_ashrrev_i32_e32 v4, 31, v0 v_mov_b32_e32 v3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_lshlrev_b64 v[5:6], 2, v[3:4] v_lshlrev_b64 v[7:8], 3, v[3:4] s_waitcnt lgkmcnt(0) v_mul_lo_u32 v9, v2, s8 v_add_co_u32 v2, vcc_lo, s4, v5 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v6, vcc_lo v_add_co_u32 v4, vcc_lo, s0, v7 v_add_co_ci_u32_e32 v5, vcc_lo, s1, v8, vcc_lo v_subrev_nc_u32_e32 v6, s2, v9 s_mov_b32 s1, 0 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_4 .p2align 6 .LBB0_3: s_or_b32 exec_lo, exec_lo, s0 v_add_nc_u32_e32 v0, 1, v0 v_add_co_u32 v2, vcc_lo, v2, 4 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_cmp_ge_i32_e32 vcc_lo, v0, v1 v_add_co_u32 v4, s0, v4, 8 v_add_co_ci_u32_e64 v5, s0, 0, v5, s0 s_or_b32 s1, vcc_lo, s1 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s1 s_cbranch_execz .LBB0_6 .LBB0_4: global_load_b32 v7, v[2:3], off s_waitcnt vmcnt(0) v_cmp_le_i32_e32 vcc_lo, s2, v7 v_cmp_ge_i32_e64 s0, s3, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s4, vcc_lo, s0 s_and_saveexec_b32 s0, s4 s_cbranch_execz .LBB0_3 global_load_b64 v[8:9], v[4:5], off v_add_nc_u32_e32 v10, v6, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v11, 31, v10 v_lshlrev_b64 v[10:11], 3, v[10:11] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v10, vcc_lo, s6, v10 v_add_co_ci_u32_e32 v11, vcc_lo, s7, v11, vcc_lo s_waitcnt vmcnt(0) global_store_b64 v[10:11], v[8:9], off s_branch .LBB0_3 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel slice_sparse_dense_row .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 312 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 12 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size slice_sparse_dense_row, .Lfunc_end0-slice_sparse_dense_row .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: by_value - .offset: 44 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: by_value - .offset: 56 .size: 4 .value_kind: hidden_block_count_x - .offset: 60 .size: 4 .value_kind: hidden_block_count_y - .offset: 64 .size: 4 .value_kind: hidden_block_count_z - .offset: 68 .size: 2 .value_kind: hidden_group_size_x - .offset: 70 .size: 2 .value_kind: hidden_group_size_y - .offset: 72 .size: 2 .value_kind: hidden_group_size_z - .offset: 74 .size: 2 .value_kind: hidden_remainder_x - .offset: 76 .size: 2 .value_kind: hidden_remainder_y - .offset: 78 .size: 2 .value_kind: hidden_remainder_z - .offset: 96 .size: 8 .value_kind: hidden_global_offset_x - .offset: 104 .size: 8 .value_kind: hidden_global_offset_y - .offset: 112 .size: 8 .value_kind: hidden_global_offset_z - .offset: 120 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 312 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: slice_sparse_dense_row .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: slice_sparse_dense_row.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 12 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00053c34_00000000-6_slice_sparse_dense_row.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii .type _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii, @function _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 216(%rsp), %rax movq %rax, 168(%rsp) leaq 224(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq slice_sparse_dense_row(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii, .-_Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii .globl slice_sparse_dense_row .type slice_sparse_dense_row, @function slice_sparse_dense_row: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 call _Z54__device_stub__Z22slice_sparse_dense_rowPdPiS0_S_iiiiiPdPiS0_S_iiiii addq $40, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size slice_sparse_dense_row, .-slice_sparse_dense_row .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "slice_sparse_dense_row" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq slice_sparse_dense_row(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "slice_sparse_dense_row.hip" .globl __device_stub__slice_sparse_dense_row # -- Begin function __device_stub__slice_sparse_dense_row .p2align 4, 0x90 .type __device_stub__slice_sparse_dense_row,@function __device_stub__slice_sparse_dense_row: # @__device_stub__slice_sparse_dense_row .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 96(%rsp), %r9 movl $slice_sparse_dense_row, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size __device_stub__slice_sparse_dense_row, .Lfunc_end0-__device_stub__slice_sparse_dense_row .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $slice_sparse_dense_row, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type slice_sparse_dense_row,@object # @slice_sparse_dense_row .section .rodata,"a",@progbits .globl slice_sparse_dense_row .p2align 3, 0x0 slice_sparse_dense_row: .quad __device_stub__slice_sparse_dense_row .size slice_sparse_dense_row, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "slice_sparse_dense_row" .size .L__unnamed_1, 23 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __device_stub__slice_sparse_dense_row .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym slice_sparse_dense_row .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <iostream> #include <cstdlib> #include <cstdio> #include <cuda_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; cudaMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { cudaDeviceSynchronize(); auto t = get_time(); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); cudaEventRecord(start); cudaDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); cudaEventRecord(stop); cudaEventSynchronize(stop); float milliseconds = 0; cudaEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
code for sm_80 Function : _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GT.U32.AND P0, PT, R2, 0xffffff, PT ; /* 0x00ffffff0200780c */ /* 0x000fda0003f04070 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R9, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff097435 */ /* 0x000fe200000001ff */ /*0070*/ IADD3 R4, R2.reuse, 0x1000000, RZ ; /* 0x0100000002047810 */ /* 0x040fe20007ffe0ff */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0090*/ IADD3 R6, R2.reuse, 0x2000000, RZ ; /* 0x0200000002067810 */ /* 0x040fe40007ffe0ff */ /*00a0*/ IADD3 R8, R2, 0x3000000, RZ ; /* 0x0300000002087810 */ /* 0x000fca0007ffe0ff */ /*00b0*/ IMAD.WIDE.U32 R2, R2, R9, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fc800078e0009 */ /*00c0*/ IMAD.WIDE.U32 R4, R4, R9.reuse, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x080fe200078e0009 */ /*00d0*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe6000c101904 */ /*00e0*/ IMAD.WIDE.U32 R6, R6, R9.reuse, c[0x0][0x160] ; /* 0x0000580006067625 */ /* 0x080fe200078e0009 */ /*00f0*/ STG.E [R4.64], RZ ; /* 0x000000ff04007986 */ /* 0x000fe6000c101904 */ /*0100*/ IMAD.WIDE.U32 R8, R8, R9, c[0x0][0x160] ; /* 0x0000580008087625 */ /* 0x000fe200078e0009 */ /*0110*/ STG.E [R6.64], RZ ; /* 0x000000ff06007986 */ /* 0x000fe8000c101904 */ /*0120*/ STG.E [R8.64], RZ ; /* 0x000000ff08007986 */ /* 0x000fe2000c101904 */ /*0130*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0140*/ BRA 0x140; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <iostream> #include <cstdlib> #include <cstdio> #include <cuda_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; cudaMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { cudaDeviceSynchronize(); auto t = get_time(); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); cudaEventRecord(start); cudaDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); cudaEventRecord(stop); cudaEventSynchronize(stop); float milliseconds = 0; cudaEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
.file "tmpxft_00122c85_00000000-6_fill.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3673: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3673: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z8get_timev .type _Z8get_timev, @function _Z8get_timev: .LFB3669: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $0, %esi call gettimeofday@PLT pxor %xmm0, %xmm0 cvtsi2sdq 8(%rsp), %xmm0 mulsd .LC0(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq (%rsp), %xmm1 addsd %xmm1, %xmm0 movq 24(%rsp), %rax subq %fs:40, %rax jne .L6 addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size _Z8get_timev, .-_Z8get_timev .globl _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f .type _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f, @function _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f: .LFB3695: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z4fillPA64_A64_A64_A4_A4_A4_A4_f(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE3695: .size _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f, .-_Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, @function _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: .LFB3696: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3696: .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, .-_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "bs = " .LC3: .string "device " .text .globl main .type main, @function main: .LFB3670: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $1, %edx movl $268435456, %esi call cudaMallocManaged@PLT leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movl $64, %esi call _ZNSolsEi@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $20, %r12d leaq 16(%rsp), %r15 leaq 24(%rsp), %r14 leaq _ZSt4cout(%rip), %r13 jmp .L21 .L27: movq 8(%rsp), %rdi call _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f jmp .L16 .L28: movq 56(%rsp), %rax subq %fs:40, %rax jne .L25 call _ZSt16__throw_bad_castv@PLT .L25: call __stack_chk_fail@PLT .L19: movq %rbp, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq 0(%rbp), %rax movl $10, %esi movq %rbp, %rdi call *48(%rax) movl %eax, %esi .L20: movsbl %sil, %esi movq %rbx, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT subl $1, %r12d je .L26 .L21: call cudaDeviceSynchronize@PLT call _Z8get_timev movq %r15, %rdi call cudaEventCreate@PLT movq %r14, %rdi call cudaEventCreate@PLT movl $0, %esi movq 16(%rsp), %rdi call cudaEventRecord@PLT call cudaDeviceSynchronize@PLT movl $64, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $262144, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L27 .L16: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movq 24(%rsp), %rdi call cudaEventSynchronize@PLT movl $0x00000000, 44(%rsp) leaq 44(%rsp), %rdi movq 24(%rsp), %rdx movq 16(%rsp), %rsi call cudaEventElapsedTime@PLT movl $8, %edx leaq .LC3(%rip), %rsi movq %r13, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT pxor %xmm0, %xmm0 cvtss2sd 44(%rsp), %xmm0 movq %r13, %rdi call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rbx movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %rbp testq %rbp, %rbp je .L28 cmpb $0, 56(%rbp) je .L19 movzbl 67(%rbp), %esi jmp .L20 .L26: leaq _ZSt4cout(%rip), %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L29 movl $0, %eax addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L29: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3670: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC4: .string "_Z4fillPA64_A64_A64_A4_A4_A4_A4_f" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3698: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z4fillPA64_A64_A64_A4_A4_A4_A4_f(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3698: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC0: .long -1598689907 .long 1051772663 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <iostream> #include <cstdlib> #include <cstdio> #include <cuda_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; cudaMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { cudaDeviceSynchronize(); auto t = get_time(); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); cudaEventRecord(start); cudaDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); cudaEventRecord(stop); cudaEventSynchronize(stop); float milliseconds = 0; cudaEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
#include <iostream> #include <cstdlib> #include <cstdio> #include <hip/hip_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; hipMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { hipDeviceSynchronize(); auto t = get_time(); hipEvent_t start, stop; hipEventCreate(&start); hipEventCreate(&stop); hipEventRecord(start); hipDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); hipEventRecord(stop); hipEventSynchronize(stop); float milliseconds = 0; hipEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <iostream> #include <cstdlib> #include <cstdio> #include <hip/hip_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; hipMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { hipDeviceSynchronize(); auto t = get_time(); hipEvent_t start, stop; hipEventCreate(&start); hipEventCreate(&stop); hipEventRecord(start); hipDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); hipEventRecord(stop); hipEventSynchronize(stop); float milliseconds = 0; hipEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 8 .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f,@function _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: s_load_b32 s2, s[0:1], 0x14 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e32 0x1000000, v1 s_cbranch_execz .LBB0_3 s_load_b64 s[0:1], s[0:1], 0x0 v_mov_b32_e32 v3, 0 s_mov_b32 s2, 0 .LBB0_2: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_add_nc_u32_e32 v2, s2, v1 s_add_i32 s2, s2, 0x1000000 s_cmp_lg_u32 s2, 0x4000000 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[2:3] s_waitcnt lgkmcnt(0) v_add_co_u32 v4, vcc_lo, s0, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v5, vcc_lo, s1, v5, vcc_lo global_store_b32 v[4:5], v3, off s_cbranch_scc1 .LBB0_2 .LBB0_3: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, .Lfunc_end0-_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4fillPA64_A64_A64_A4_A4_A4_A4_f.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <iostream> #include <cstdlib> #include <cstdio> #include <hip/hip_runtime.h> #include <sys/time.h> double get_time() { struct timeval tv; gettimeofday(&tv, nullptr); return tv.tv_sec + 1e-6 * tv.tv_usec; } constexpr int m = 256; constexpr int n = m * m * m; constexpr int block_size = 4; using grid_type = float[m / block_size][m / block_size][m / block_size][4] [block_size][block_size][block_size]; __global__ void fill(grid_type *grid_) { unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; if (i >= n) return; float *data = (float *)grid_; grid_type &grid = *grid_; for (int k = 0; k < 4; k++) { data[i + k * n] = 0; } } int main() { float *a; hipMallocManaged(&a, n * sizeof(float) * 4); auto bs = block_size * block_size * block_size; std::cout << "bs = " << bs << std::endl; for (int i = 0; i < 20; i++) { hipDeviceSynchronize(); auto t = get_time(); hipEvent_t start, stop; hipEventCreate(&start); hipEventCreate(&stop); hipEventRecord(start); hipDeviceSynchronize(); fill<<<n / bs, bs>>>((grid_type *)a); hipEventRecord(stop); hipEventSynchronize(stop); float milliseconds = 0; hipEventElapsedTime(&milliseconds, start, stop); std::cout << "device " << milliseconds << std::endl; } std::cout << std::endl; }
.text .file "fill.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z8get_timev .LCPI0_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .text .globl _Z8get_timev .p2align 4, 0x90 .type _Z8get_timev,@function _Z8get_timev: # @_Z8get_timev .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 leaq 8(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 8(%rsp), %xmm1 cvtsi2sdq 16(%rsp), %xmm0 mulsd .LCPI0_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z8get_timev, .Lfunc_end0-_Z8get_timev .cfi_endproc # -- End function .globl _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f # -- Begin function _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 4, 0x90 .type _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f,@function _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f: # @_Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movq %rdi, 64(%rsp) leaq 64(%rsp), %rax movq %rax, (%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d movq %rsp, %r9 movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end1: .size _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f, .Lfunc_end1-_Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $88, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 leaq 40(%rsp), %rdi movl $268435456, %esi # imm = 0x10000000 movl $1, %edx callq hipMallocManaged movl $_ZSt4cout, %edi movl $.L.str, %esi movl $5, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movl $_ZSt4cout, %edi movl $64, %esi callq _ZNSolsEi movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.1: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB2_3 # %bb.2: movzbl 67(%rbx), %ecx jmp .LBB2_4 .LBB2_3: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB2_4: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movabsq $4294967360, %rbx # imm = 0x100000040 movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movl $20, %r15d leaq 16(%rsp), %r14 movq %rbx, %r12 leaq 262080(%rbx), %r13 jmp .LBB2_5 .p2align 4, 0x90 .LBB2_15: # in Loop: Header=BB2_5 Depth=1 movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB2_16: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit17 # in Loop: Header=BB2_5 Depth=1 movsbl %al, %esi movq %rbp, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv decl %r15d je .LBB2_10 .LBB2_5: # =>This Inner Loop Header: Depth=1 callq hipDeviceSynchronize movq %r14, %rdi xorl %esi, %esi callq gettimeofday leaq 8(%rsp), %rdi callq hipEventCreate movq %rsp, %rdi callq hipEventCreate movq 8(%rsp), %rdi xorl %esi, %esi callq hipEventRecord callq hipDeviceSynchronize movq %r13, %rdi movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_7 # %bb.6: # in Loop: Header=BB2_5 Depth=1 movq 40(%rsp), %rax movq %rax, 80(%rsp) leaq 80(%rsp), %rax movq %rax, 32(%rsp) movq %r14, %rdi leaq 64(%rsp), %rsi leaq 56(%rsp), %rdx leaq 48(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %edi leaq 32(%rsp), %r9 pushq 48(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_7: # in Loop: Header=BB2_5 Depth=1 movq (%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq (%rsp), %rdi callq hipEventSynchronize movl $0, 16(%rsp) movq 8(%rsp), %rsi movq (%rsp), %rdx movq %r14, %rdi callq hipEventElapsedTime movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $8, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movss 16(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq %rax, %rbp movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbp,%rax), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i14 # in Loop: Header=BB2_5 Depth=1 cmpb $0, 56(%rbx) je .LBB2_15 # %bb.9: # in Loop: Header=BB2_5 Depth=1 movzbl 67(%rbx), %eax jmp .LBB2_16 .LBB2_10: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.11: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i9 cmpb $0, 56(%rbx) je .LBB2_13 # %bb.12: movzbl 67(%rbx), %eax jmp .LBB2_14 .LBB2_13: movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB2_14: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit12 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %eax, %eax addq $88, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB2_17: .cfi_def_cfa_offset 144 callq _ZSt16__throw_bad_castv .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f,@object # @_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .rodata,"a",@progbits .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 3, 0x0 _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: .quad _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "bs = " .size .L.str, 6 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "device " .size .L.str.1, 9 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z4fillPA64_A64_A64_A4_A4_A4_A4_f" .size .L__unnamed_1, 34 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GT.U32.AND P0, PT, R2, 0xffffff, PT ; /* 0x00ffffff0200780c */ /* 0x000fda0003f04070 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R9, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff097435 */ /* 0x000fe200000001ff */ /*0070*/ IADD3 R4, R2.reuse, 0x1000000, RZ ; /* 0x0100000002047810 */ /* 0x040fe20007ffe0ff */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0090*/ IADD3 R6, R2.reuse, 0x2000000, RZ ; /* 0x0200000002067810 */ /* 0x040fe40007ffe0ff */ /*00a0*/ IADD3 R8, R2, 0x3000000, RZ ; /* 0x0300000002087810 */ /* 0x000fca0007ffe0ff */ /*00b0*/ IMAD.WIDE.U32 R2, R2, R9, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fc800078e0009 */ /*00c0*/ IMAD.WIDE.U32 R4, R4, R9.reuse, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x080fe200078e0009 */ /*00d0*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe6000c101904 */ /*00e0*/ IMAD.WIDE.U32 R6, R6, R9.reuse, c[0x0][0x160] ; /* 0x0000580006067625 */ /* 0x080fe200078e0009 */ /*00f0*/ STG.E [R4.64], RZ ; /* 0x000000ff04007986 */ /* 0x000fe6000c101904 */ /*0100*/ IMAD.WIDE.U32 R8, R8, R9, c[0x0][0x160] ; /* 0x0000580008087625 */ /* 0x000fe200078e0009 */ /*0110*/ STG.E [R6.64], RZ ; /* 0x000000ff06007986 */ /* 0x000fe8000c101904 */ /*0120*/ STG.E [R8.64], RZ ; /* 0x000000ff08007986 */ /* 0x000fe2000c101904 */ /*0130*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0140*/ BRA 0x140; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 8 .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f,@function _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: s_load_b32 s2, s[0:1], 0x14 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e32 0x1000000, v1 s_cbranch_execz .LBB0_3 s_load_b64 s[0:1], s[0:1], 0x0 v_mov_b32_e32 v3, 0 s_mov_b32 s2, 0 .LBB0_2: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_add_nc_u32_e32 v2, s2, v1 s_add_i32 s2, s2, 0x1000000 s_cmp_lg_u32 s2, 0x4000000 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[2:3] s_waitcnt lgkmcnt(0) v_add_co_u32 v4, vcc_lo, s0, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v5, vcc_lo, s1, v5, vcc_lo global_store_b32 v[4:5], v3, off s_cbranch_scc1 .LBB0_2 .LBB0_3: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, .Lfunc_end0-_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4fillPA64_A64_A64_A4_A4_A4_A4_f.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00122c85_00000000-6_fill.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3673: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3673: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z8get_timev .type _Z8get_timev, @function _Z8get_timev: .LFB3669: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $0, %esi call gettimeofday@PLT pxor %xmm0, %xmm0 cvtsi2sdq 8(%rsp), %xmm0 mulsd .LC0(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq (%rsp), %xmm1 addsd %xmm1, %xmm0 movq 24(%rsp), %rax subq %fs:40, %rax jne .L6 addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size _Z8get_timev, .-_Z8get_timev .globl _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f .type _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f, @function _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f: .LFB3695: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z4fillPA64_A64_A64_A4_A4_A4_A4_f(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE3695: .size _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f, .-_Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, @function _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: .LFB3696: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3696: .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, .-_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "bs = " .LC3: .string "device " .text .globl main .type main, @function main: .LFB3670: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $1, %edx movl $268435456, %esi call cudaMallocManaged@PLT leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movl $64, %esi call _ZNSolsEi@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $20, %r12d leaq 16(%rsp), %r15 leaq 24(%rsp), %r14 leaq _ZSt4cout(%rip), %r13 jmp .L21 .L27: movq 8(%rsp), %rdi call _Z47__device_stub__Z4fillPA64_A64_A64_A4_A4_A4_A4_fPA64_A64_A64_A4_A4_A4_A4_f jmp .L16 .L28: movq 56(%rsp), %rax subq %fs:40, %rax jne .L25 call _ZSt16__throw_bad_castv@PLT .L25: call __stack_chk_fail@PLT .L19: movq %rbp, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq 0(%rbp), %rax movl $10, %esi movq %rbp, %rdi call *48(%rax) movl %eax, %esi .L20: movsbl %sil, %esi movq %rbx, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT subl $1, %r12d je .L26 .L21: call cudaDeviceSynchronize@PLT call _Z8get_timev movq %r15, %rdi call cudaEventCreate@PLT movq %r14, %rdi call cudaEventCreate@PLT movl $0, %esi movq 16(%rsp), %rdi call cudaEventRecord@PLT call cudaDeviceSynchronize@PLT movl $64, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $262144, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L27 .L16: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movq 24(%rsp), %rdi call cudaEventSynchronize@PLT movl $0x00000000, 44(%rsp) leaq 44(%rsp), %rdi movq 24(%rsp), %rdx movq 16(%rsp), %rsi call cudaEventElapsedTime@PLT movl $8, %edx leaq .LC3(%rip), %rsi movq %r13, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT pxor %xmm0, %xmm0 cvtss2sd 44(%rsp), %xmm0 movq %r13, %rdi call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rbx movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %rbp testq %rbp, %rbp je .L28 cmpb $0, 56(%rbp) je .L19 movzbl 67(%rbp), %esi jmp .L20 .L26: leaq _ZSt4cout(%rip), %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L29 movl $0, %eax addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L29: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3670: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC4: .string "_Z4fillPA64_A64_A64_A4_A4_A4_A4_f" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3698: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z4fillPA64_A64_A64_A4_A4_A4_A4_f(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3698: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC0: .long -1598689907 .long 1051772663 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "fill.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z8get_timev .LCPI0_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .text .globl _Z8get_timev .p2align 4, 0x90 .type _Z8get_timev,@function _Z8get_timev: # @_Z8get_timev .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 leaq 8(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 8(%rsp), %xmm1 cvtsi2sdq 16(%rsp), %xmm0 mulsd .LCPI0_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z8get_timev, .Lfunc_end0-_Z8get_timev .cfi_endproc # -- End function .globl _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f # -- Begin function _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 4, 0x90 .type _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f,@function _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f: # @_Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movq %rdi, 64(%rsp) leaq 64(%rsp), %rax movq %rax, (%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d movq %rsp, %r9 movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end1: .size _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f, .Lfunc_end1-_Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $88, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 leaq 40(%rsp), %rdi movl $268435456, %esi # imm = 0x10000000 movl $1, %edx callq hipMallocManaged movl $_ZSt4cout, %edi movl $.L.str, %esi movl $5, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movl $_ZSt4cout, %edi movl $64, %esi callq _ZNSolsEi movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.1: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB2_3 # %bb.2: movzbl 67(%rbx), %ecx jmp .LBB2_4 .LBB2_3: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB2_4: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movabsq $4294967360, %rbx # imm = 0x100000040 movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movl $20, %r15d leaq 16(%rsp), %r14 movq %rbx, %r12 leaq 262080(%rbx), %r13 jmp .LBB2_5 .p2align 4, 0x90 .LBB2_15: # in Loop: Header=BB2_5 Depth=1 movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB2_16: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit17 # in Loop: Header=BB2_5 Depth=1 movsbl %al, %esi movq %rbp, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv decl %r15d je .LBB2_10 .LBB2_5: # =>This Inner Loop Header: Depth=1 callq hipDeviceSynchronize movq %r14, %rdi xorl %esi, %esi callq gettimeofday leaq 8(%rsp), %rdi callq hipEventCreate movq %rsp, %rdi callq hipEventCreate movq 8(%rsp), %rdi xorl %esi, %esi callq hipEventRecord callq hipDeviceSynchronize movq %r13, %rdi movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_7 # %bb.6: # in Loop: Header=BB2_5 Depth=1 movq 40(%rsp), %rax movq %rax, 80(%rsp) leaq 80(%rsp), %rax movq %rax, 32(%rsp) movq %r14, %rdi leaq 64(%rsp), %rsi leaq 56(%rsp), %rdx leaq 48(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %edi leaq 32(%rsp), %r9 pushq 48(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_7: # in Loop: Header=BB2_5 Depth=1 movq (%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq (%rsp), %rdi callq hipEventSynchronize movl $0, 16(%rsp) movq 8(%rsp), %rsi movq (%rsp), %rdx movq %r14, %rdi callq hipEventElapsedTime movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $8, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movss 16(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq %rax, %rbp movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbp,%rax), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i14 # in Loop: Header=BB2_5 Depth=1 cmpb $0, 56(%rbx) je .LBB2_15 # %bb.9: # in Loop: Header=BB2_5 Depth=1 movzbl 67(%rbx), %eax jmp .LBB2_16 .LBB2_10: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %rbx testq %rbx, %rbx je .LBB2_17 # %bb.11: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i9 cmpb $0, 56(%rbx) je .LBB2_13 # %bb.12: movzbl 67(%rbx), %eax jmp .LBB2_14 .LBB2_13: movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB2_14: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit12 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %eax, %eax addq $88, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB2_17: .cfi_def_cfa_offset 144 callq _ZSt16__throw_bad_castv .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4fillPA64_A64_A64_A4_A4_A4_A4_f, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z4fillPA64_A64_A64_A4_A4_A4_A4_f,@object # @_Z4fillPA64_A64_A64_A4_A4_A4_A4_f .section .rodata,"a",@progbits .globl _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .p2align 3, 0x0 _Z4fillPA64_A64_A64_A4_A4_A4_A4_f: .quad _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .size _Z4fillPA64_A64_A64_A4_A4_A4_A4_f, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "bs = " .size .L.str, 6 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "device " .size .L.str.1, 9 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z4fillPA64_A64_A64_A4_A4_A4_A4_f" .size .L__unnamed_1, 34 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__fillPA64_A64_A64_A4_A4_A4_A4_f .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4fillPA64_A64_A64_A4_A4_A4_A4_f .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
code for sm_80 Function : _Z15MatrixMulDevicePfS_S_Pi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff027624 */ /* 0x000fe200078e00ff */ /*0020*/ MOV R3, c[0x0][0x17c] ; /* 0x00005f0000037a02 */ /* 0x000fe20000000f00 */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fc80000000a00 */ /*0040*/ LDG.E R0, [R2.64] ; /* 0x0000000602007981 */ /* 0x000ea2000c1e1900 */ /*0050*/ I2F.U32.RP R6, c[0x0][0xc] ; /* 0x0000030000067b06 */ /* 0x000e220000209000 */ /*0060*/ ISETP.NE.U32.AND P2, PT, RZ, c[0x0][0xc], PT ; /* 0x00000300ff007a0c */ /* 0x000fce0003f45070 */ /*0070*/ MUFU.RCP R6, R6 ; /* 0x0000000600067308 */ /* 0x001e240000001000 */ /*0080*/ IADD3 R4, R6, 0xffffffe, RZ ; /* 0x0ffffffe06047810 */ /* 0x001fcc0007ffe0ff */ /*0090*/ F2I.FTZ.U32.TRUNC.NTZ R5, R4 ; /* 0x0000000400057305 */ /* 0x000064000021f000 */ /*00a0*/ HFMA2.MMA R4, -RZ, RZ, 0, 0 ; /* 0x00000000ff047435 */ /* 0x001fe200000001ff */ /*00b0*/ IMAD.MOV R7, RZ, RZ, -R5 ; /* 0x000000ffff077224 */ /* 0x002fc800078e0a05 */ /*00c0*/ IMAD R7, R7, c[0x0][0xc], RZ ; /* 0x0000030007077a24 */ /* 0x000fca00078e02ff */ /*00d0*/ IMAD.HI.U32 R5, R5, R7, R4 ; /* 0x0000000705057227 */ /* 0x000fe400078e0004 */ /*00e0*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e280000002500 */ /*00f0*/ IMAD.HI.U32 R5, R5, R0, RZ ; /* 0x0000000005057227 */ /* 0x004fc800078e00ff */ /*0100*/ IMAD.MOV R7, RZ, RZ, -R5 ; /* 0x000000ffff077224 */ /* 0x000fc800078e0a05 */ /*0110*/ IMAD R0, R7, c[0x0][0xc], R0 ; /* 0x0000030007007a24 */ /* 0x000fca00078e0200 */ /*0120*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0xc], PT ; /* 0x0000030000007a0c */ /* 0x000fda0003f06070 */ /*0130*/ @P0 IADD3 R0, R0, -c[0x0][0xc], RZ ; /* 0x8000030000000a10 */ /* 0x000fe40007ffe0ff */ /*0140*/ @P0 IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105050810 */ /* 0x000fe40007ffe0ff */ /*0150*/ ISETP.GE.U32.AND P1, PT, R0, c[0x0][0xc], PT ; /* 0x0000030000007a0c */ /* 0x000fda0003f26070 */ /*0160*/ @P1 IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105051810 */ /* 0x000fe40007ffe0ff */ /*0170*/ @!P2 LOP3.LUT R5, RZ, c[0x0][0xc], RZ, 0x33, !PT ; /* 0x00000300ff05aa12 */ /* 0x000fca00078e33ff */ /*0180*/ IMAD R15, R5, R4, RZ ; /* 0x00000004050f7224 */ /* 0x001fca00078e02ff */ /*0190*/ IADD3 R0, R5, -0x1, R15 ; /* 0xffffffff05007810 */ /* 0x000fc80007ffe00f */ /*01a0*/ ISETP.GE.U32.AND P0, PT, R15, R0, PT ; /* 0x000000000f00720c */ /* 0x000fda0003f06070 */ /*01b0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*01c0*/ S2R R14, SR_TID.X ; /* 0x00000000000e7919 */ /* 0x000e240000002100 */ /*01d0*/ LDG.E R17, [R2.64] ; /* 0x0000000602117981 */ /* 0x000ea4000c1e1900 */ /*01e0*/ ISETP.GT.AND P0, PT, R17, RZ, PT ; /* 0x000000ff1100720c */ /* 0x004fda0003f04270 */ /*01f0*/ @!P0 MOV R21, RZ ; /* 0x000000ff00158202 */ /* 0x001fe20000000f00 */ /*0200*/ @!P0 BRA 0x6f0 ; /* 0x000004e000008947 */ /* 0x000fea0003800000 */ /*0210*/ IADD3 R4, R17.reuse, -0x1, RZ ; /* 0xffffffff11047810 */ /* 0x040fe20007ffe0ff */ /*0220*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0230*/ LOP3.LUT R16, R17.reuse, 0x3, RZ, 0xc0, !PT ; /* 0x0000000311107812 */ /* 0x040fe200078ec0ff */ /*0240*/ IMAD R18, R17, R15, RZ ; /* 0x0000000f11127224 */ /* 0x000fe200078e02ff */ /*0250*/ ISETP.GE.U32.AND P1, PT, R4, 0x3, PT ; /* 0x000000030400780c */ /* 0x000fe20003f26070 */ /*0260*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0270*/ ISETP.NE.AND P0, PT, R16, RZ, PT ; /* 0x000000ff1000720c */ /* 0x000fd60003f05270 */ /*0280*/ @!P1 BRA 0x530 ; /* 0x000002a000009947 */ /* 0x000fea0003800000 */ /*0290*/ HFMA2.MMA R19, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff137435 */ /* 0x000fe200000001ff */ /*02a0*/ IMAD.IADD R22, R17.reuse, 0x1, R14.reuse ; /* 0x0000000111167824 */ /* 0x141fe200078e020e */ /*02b0*/ LEA R20, R17.reuse, R14.reuse, 0x1 ; /* 0x0000000e11147211 */ /* 0x0c0fe200078e08ff */ /*02c0*/ IMAD R24, R17, 0x3, R14 ; /* 0x0000000311187824 */ /* 0x000fe200078e020e */ /*02d0*/ MOV R26, R14 ; /* 0x0000000e001a7202 */ /* 0x000fe20000000f00 */ /*02e0*/ IMAD.IADD R27, R16, 0x1, -R17 ; /* 0x00000001101b7824 */ /* 0x000fe200078e0a11 */ /*02f0*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fc80008000000 */ /*0300*/ IMAD.WIDE R4, R18, R19, c[0x0][0x160] ; /* 0x0000580012047625 */ /* 0x000fc800078e0213 */ /*0310*/ IMAD.WIDE.U32 R6, R26, R19.reuse, c[0x0][0x168] ; /* 0x00005a001a067625 */ /* 0x080fe200078e0013 */ /*0320*/ LDG.E R23, [R4.64] ; /* 0x0000000604177981 */ /* 0x0010a6000c1e1900 */ /*0330*/ IMAD.WIDE.U32 R8, R22, R19.reuse, c[0x0][0x168] ; /* 0x00005a0016087625 */ /* 0x080fe200078e0013 */ /*0340*/ LDG.E R25, [R4.64+0x4] ; /* 0x0000040604197981 */ /* 0x0000e8000c1e1900 */ /*0350*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea8000c1e1900 */ /*0360*/ LDG.E R8, [R8.64] ; /* 0x0000000608087981 */ /* 0x000ee2000c1e1900 */ /*0370*/ IMAD.WIDE.U32 R10, R20, R19, c[0x0][0x168] ; /* 0x00005a00140a7625 */ /* 0x000fc600078e0013 */ /*0380*/ LDG.E R29, [R4.64+0x8] ; /* 0x00000806041d7981 */ /* 0x000128000c1e1900 */ /*0390*/ LDG.E R10, [R10.64] ; /* 0x000000060a0a7981 */ /* 0x000f22000c1e1900 */ /*03a0*/ IMAD.WIDE.U32 R12, R24, R19, c[0x0][0x168] ; /* 0x00005a00180c7625 */ /* 0x000fc600078e0013 */ /*03b0*/ LDG.E R28, [R4.64+0xc] ; /* 0x00000c06041c7981 */ /* 0x000168000c1e1900 */ /*03c0*/ LDG.E R12, [R12.64] ; /* 0x000000060c0c7981 */ /* 0x000f62000c1e1900 */ /*03d0*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea20000201400 */ /*03e0*/ UIADD3 UR4, UR4, 0x4, URZ ; /* 0x0000000404047890 */ /* 0x000fe2000fffe03f */ /*03f0*/ IADD3 R4, P2, R4, 0x10, RZ ; /* 0x0000001004047810 */ /* 0x001fe20007f5e0ff */ /*0400*/ IMAD R24, R17.reuse, 0x4, R24 ; /* 0x0000000411187824 */ /* 0x040fe200078e0218 */ /*0410*/ LEA R22, R17.reuse, R22, 0x2 ; /* 0x0000001611167211 */ /* 0x040fe400078e10ff */ /*0420*/ LEA R20, R17.reuse, R20, 0x2 ; /* 0x0000001411147211 */ /* 0x040fe200078e10ff */ /*0430*/ IMAD.X R5, RZ, RZ, R5, P2 ; /* 0x000000ffff057224 */ /* 0x000fe200010e0605 */ /*0440*/ LEA R26, R17, R26, 0x2 ; /* 0x0000001a111a7211 */ /* 0x000fe200078e10ff */ /*0450*/ FFMA R6, R6, R23, R21 ; /* 0x0000001706067223 */ /* 0x004fcc0000000015 */ /*0460*/ F2I.TRUNC.NTZ R6, R6 ; /* 0x0000000600067305 */ /* 0x000e30000020f100 */ /*0470*/ I2F R7, R6 ; /* 0x0000000600077306 */ /* 0x001ee40000201400 */ /*0480*/ FFMA R7, R8, R25, R7 ; /* 0x0000001908077223 */ /* 0x008fcc0000000007 */ /*0490*/ F2I.TRUNC.NTZ R7, R7 ; /* 0x0000000700077305 */ /* 0x000e30000020f100 */ /*04a0*/ I2F R8, R7 ; /* 0x0000000700087306 */ /* 0x001f240000201400 */ /*04b0*/ FFMA R8, R10, R29, R8 ; /* 0x0000001d0a087223 */ /* 0x010fcc0000000008 */ /*04c0*/ F2I.TRUNC.NTZ R8, R8 ; /* 0x0000000800087305 */ /* 0x000e22000020f100 */ /*04d0*/ IADD3 R10, R27, UR4, RZ ; /* 0x000000041b0a7c10 */ /* 0x000fce000fffe0ff */ /*04e0*/ I2F R9, R8 ; /* 0x0000000800097306 */ /* 0x001f620000201400 */ /*04f0*/ ISETP.NE.AND P1, PT, R10, RZ, PT ; /* 0x000000ff0a00720c */ /* 0x000fe20003f25270 */ /*0500*/ FFMA R9, R12, R28, R9 ; /* 0x0000001c0c097223 */ /* 0x020fcc0000000009 */ /*0510*/ F2I.TRUNC.NTZ R21, R9 ; /* 0x0000000900157305 */ /* 0x00006c000020f100 */ /*0520*/ @P1 BRA 0x310 ; /* 0xfffffde000001947 */ /* 0x000fea000383ffff */ /*0530*/ @!P0 BRA 0x6f0 ; /* 0x000001b000008947 */ /* 0x000fea0003800000 */ /*0540*/ HFMA2.MMA R13, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0d7435 */ /* 0x000fe200000001ff */ /*0550*/ IADD3 R4, R18, UR4, RZ ; /* 0x0000000412047c10 */ /* 0x000fe2000fffe0ff */ /*0560*/ IMAD R8, R17, UR4, R14 ; /* 0x0000000411087c24 */ /* 0x001fd0000f8e020e */ /*0570*/ IMAD.WIDE.U32 R6, R8, R13, c[0x0][0x168] ; /* 0x00005a0008067625 */ /* 0x000fc800078e000d */ /*0580*/ IMAD.WIDE R4, R4, R13, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e020d */ /*0590*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea8000c1e1900 */ /*05a0*/ LDG.E R9, [R4.64] ; /* 0x0000000604097981 */ /* 0x000ea2000c1e1900 */ /*05b0*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea20000201400 */ /*05c0*/ ISETP.NE.AND P0, PT, R16, 0x1, PT ; /* 0x000000011000780c */ /* 0x000fe20003f05270 */ /*05d0*/ FFMA R9, R6, R9, R21 ; /* 0x0000000906097223 */ /* 0x004fcc0000000015 */ /*05e0*/ F2I.TRUNC.NTZ R21, R9 ; /* 0x0000000900157305 */ /* 0x00006c000020f100 */ /*05f0*/ @!P0 BRA 0x6f0 ; /* 0x000000f000008947 */ /* 0x000fea0003800000 */ /*0600*/ IMAD.IADD R8, R17, 0x1, R8 ; /* 0x0000000111087824 */ /* 0x000fe200078e0208 */ /*0610*/ LDG.E R11, [R4.64+0x4] ; /* 0x00000406040b7981 */ /* 0x000ea6000c1e1900 */ /*0620*/ IMAD.WIDE.U32 R6, R8, R13, c[0x0][0x168] ; /* 0x00005a0008067625 */ /* 0x000fcc00078e000d */ /*0630*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea2000c1e1900 */ /*0640*/ ISETP.NE.AND P0, PT, R16, 0x2, PT ; /* 0x000000021000780c */ /* 0x000fda0003f05270 */ /*0650*/ @P0 IADD3 R8, R17, R8, RZ ; /* 0x0000000811080210 */ /* 0x000fca0007ffe0ff */ /*0660*/ @P0 IMAD.WIDE.U32 R8, R8, R13, c[0x0][0x168] ; /* 0x00005a0008080625 */ /* 0x001fe400078e000d */ /*0670*/ @P0 LDG.E R13, [R4.64+0x8] ; /* 0x00000806040d0981 */ /* 0x000ee8000c1e1900 */ /*0680*/ @P0 LDG.E R8, [R8.64] ; /* 0x0000000608080981 */ /* 0x000ee2000c1e1900 */ /*0690*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea40000201400 */ /*06a0*/ FFMA R11, R6, R11, R21 ; /* 0x0000000b060b7223 */ /* 0x004fcc0000000015 */ /*06b0*/ F2I.TRUNC.NTZ R21, R11 ; /* 0x0000000b00157305 */ /* 0x000e30000020f100 */ /*06c0*/ @P0 I2F R10, R21 ; /* 0x00000015000a0306 */ /* 0x001ee40000201400 */ /*06d0*/ @P0 FFMA R13, R8, R13, R10 ; /* 0x0000000d080d0223 */ /* 0x008fcc000000000a */ /*06e0*/ @P0 F2I.TRUNC.NTZ R21, R13 ; /* 0x0000000d00150305 */ /* 0x000064000020f100 */ /*06f0*/ IMAD R4, R17, R15, R14 ; /* 0x0000000f11047224 */ /* 0x001fe200078e020e */ /*0700*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002e220000201400 */ /*0710*/ IADD3 R15, R15, 0x1, RZ ; /* 0x000000010f0f7810 */ /* 0x000fe40007ffe0ff */ /*0720*/ MOV R5, 0x4 ; /* 0x0000000400057802 */ /* 0x000fe40000000f00 */ /*0730*/ ISETP.GE.U32.AND P0, PT, R15, R0, PT ; /* 0x000000000f00720c */ /* 0x000fc60003f06070 */ /*0740*/ IMAD.WIDE.U32 R4, R4, R5, c[0x0][0x170] ; /* 0x00005c0004047625 */ /* 0x000fca00078e0005 */ /*0750*/ STG.E [R4.64], R21 ; /* 0x0000001504007986 */ /* 0x0011ea000c101906 */ /*0760*/ @!P0 BRA 0x1d0 ; /* 0xfffffa6000008947 */ /* 0x000fea000383ffff */ /*0770*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0780*/ BRA 0x780; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0790*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0800*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0810*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0820*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0830*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0840*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0850*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0860*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0870*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
.file "tmpxft_00003130_00000000-6_MatrixMulDevice.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi .type _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi, @function _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z15MatrixMulDevicePfS_S_Pi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi, .-_Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi .globl _Z15MatrixMulDevicePfS_S_Pi .type _Z15MatrixMulDevicePfS_S_Pi, @function _Z15MatrixMulDevicePfS_S_Pi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z15MatrixMulDevicePfS_S_Pi, .-_Z15MatrixMulDevicePfS_S_Pi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z15MatrixMulDevicePfS_S_Pi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z15MatrixMulDevicePfS_S_Pi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z15MatrixMulDevicePfS_S_Pi .globl _Z15MatrixMulDevicePfS_S_Pi .p2align 8 .type _Z15MatrixMulDevicePfS_S_Pi,@function _Z15MatrixMulDevicePfS_S_Pi: s_clause 0x1 s_load_b32 s4, s[0:1], 0x20 s_load_b64 s[2:3], s[0:1], 0x18 s_waitcnt lgkmcnt(0) v_cvt_f32_u32_e32 v1, s4 s_load_b32 s12, s[2:3], 0x0 s_sub_i32 s3, 0, s4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_iflag_f32_e32 v1, v1 s_waitcnt_depctr 0xfff v_mul_f32_e32 v1, 0x4f7ffffe, v1 v_cvt_u32_f32_e32 v1, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_readfirstlane_b32 s2, v1 s_mul_i32 s3, s3, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_mul_hi_u32 s3, s2, s3 s_add_i32 s2, s2, s3 s_waitcnt lgkmcnt(0) s_mul_hi_u32 s2, s12, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(SALU_CYCLE_1) s_mul_i32 s3, s2, s4 s_add_i32 s5, s2, 1 s_sub_i32 s3, s12, s3 s_sub_i32 s6, s3, s4 s_cmp_ge_u32 s3, s4 s_cselect_b32 s2, s5, s2 s_cselect_b32 s3, s6, s3 s_add_i32 s5, s2, 1 s_cmp_ge_u32 s3, s4 s_cselect_b32 s8, s5, s2 s_add_i32 s2, s15, 1 s_mul_i32 s13, s8, s15 s_mul_i32 s14, s8, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s14, s14, -1 s_cmp_ge_u32 s13, s14 s_cbranch_scc1 .LBB0_6 s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[2:3], s[0:1], 0x10 s_cmp_gt_i32 s12, 0 v_mov_b32_e32 v2, 0 s_cselect_b32 s0, -1, 0 s_mul_i32 s1, s8, s15 v_cndmask_b32_e64 v1, 0, 1, s0 s_mul_i32 s8, s1, s12 s_delay_alu instid0(VALU_DEP_1) v_cmp_ne_u32_e64 s0, 1, v1 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: v_mad_u64_u32 v[4:5], null, s13, s12, v[0:1] v_mov_b32_e32 v5, v2 v_cvt_f32_i32_e32 v1, v3 s_add_i32 s13, s13, 1 s_add_i32 s8, s8, s12 s_cmp_lt_u32 s13, s14 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[4:5] s_waitcnt lgkmcnt(0) v_add_co_u32 v3, vcc_lo, s2, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s3, v5, vcc_lo global_store_b32 v[3:4], v1, off s_cbranch_scc0 .LBB0_6 .LBB0_3: v_mov_b32_e32 v3, v2 s_delay_alu instid0(VALU_DEP_2) s_and_b32 vcc_lo, exec_lo, s0 s_cbranch_vccnz .LBB0_2 s_ashr_i32 s9, s8, 31 v_mov_b32_e32 v3, 0 s_lshl_b64 s[10:11], s[8:9], 2 v_mov_b32_e32 v1, v0 s_waitcnt lgkmcnt(0) s_add_u32 s10, s4, s10 s_addc_u32 s11, s5, s11 s_mov_b32 s1, s12 .p2align 6 .LBB0_5: v_lshlrev_b64 v[4:5], 2, v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_cvt_f32_i32_e32 v3, v3 s_add_i32 s1, s1, -1 v_add_nc_u32_e32 v1, s12, v1 v_add_co_u32 v4, vcc_lo, s6, v4 s_delay_alu instid0(VALU_DEP_4) v_add_co_ci_u32_e32 v5, vcc_lo, s7, v5, vcc_lo global_load_b32 v6, v2, s[10:11] global_load_b32 v4, v[4:5], off s_add_u32 s10, s10, 4 s_addc_u32 s11, s11, 0 s_cmp_lg_u32 s1, 0 s_waitcnt vmcnt(0) v_fmac_f32_e32 v3, v6, v4 s_delay_alu instid0(VALU_DEP_1) v_cvt_i32_f32_e32 v3, v3 s_cbranch_scc1 .LBB0_5 s_branch .LBB0_2 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z15MatrixMulDevicePfS_S_Pi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z15MatrixMulDevicePfS_S_Pi, .Lfunc_end0-_Z15MatrixMulDevicePfS_S_Pi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: hidden_block_count_x - .offset: 36 .size: 4 .value_kind: hidden_block_count_y - .offset: 40 .size: 4 .value_kind: hidden_block_count_z - .offset: 44 .size: 2 .value_kind: hidden_group_size_x - .offset: 46 .size: 2 .value_kind: hidden_group_size_y - .offset: 48 .size: 2 .value_kind: hidden_group_size_z - .offset: 50 .size: 2 .value_kind: hidden_remainder_x - .offset: 52 .size: 2 .value_kind: hidden_remainder_y - .offset: 54 .size: 2 .value_kind: hidden_remainder_z - .offset: 72 .size: 8 .value_kind: hidden_global_offset_x - .offset: 80 .size: 8 .value_kind: hidden_global_offset_y - .offset: 88 .size: 8 .value_kind: hidden_global_offset_z - .offset: 96 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z15MatrixMulDevicePfS_S_Pi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z15MatrixMulDevicePfS_S_Pi.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 7 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void MatrixMulDevice( float *A, float *B, float *C, int *matrixSize) { int chunk = (*matrixSize) / gridDim.x; int sum, i, k; for(i = blockIdx.x * chunk; i < blockIdx.x * chunk + chunk - 1; i++) { sum = 0; for(k = 0; k < *matrixSize; k++) { sum += A[i * *matrixSize + k] * B [k * *matrixSize + threadIdx.x]; } C[i * *matrixSize + threadIdx.x] = sum; } }
.text .file "MatrixMulDevice.hip" .globl _Z30__device_stub__MatrixMulDevicePfS_S_Pi # -- Begin function _Z30__device_stub__MatrixMulDevicePfS_S_Pi .p2align 4, 0x90 .type _Z30__device_stub__MatrixMulDevicePfS_S_Pi,@function _Z30__device_stub__MatrixMulDevicePfS_S_Pi: # @_Z30__device_stub__MatrixMulDevicePfS_S_Pi .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movq %rcx, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z15MatrixMulDevicePfS_S_Pi, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z30__device_stub__MatrixMulDevicePfS_S_Pi, .Lfunc_end0-_Z30__device_stub__MatrixMulDevicePfS_S_Pi .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z15MatrixMulDevicePfS_S_Pi, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z15MatrixMulDevicePfS_S_Pi,@object # @_Z15MatrixMulDevicePfS_S_Pi .section .rodata,"a",@progbits .globl _Z15MatrixMulDevicePfS_S_Pi .p2align 3, 0x0 _Z15MatrixMulDevicePfS_S_Pi: .quad _Z30__device_stub__MatrixMulDevicePfS_S_Pi .size _Z15MatrixMulDevicePfS_S_Pi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z15MatrixMulDevicePfS_S_Pi" .size .L__unnamed_1, 28 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z30__device_stub__MatrixMulDevicePfS_S_Pi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z15MatrixMulDevicePfS_S_Pi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z15MatrixMulDevicePfS_S_Pi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff027624 */ /* 0x000fe200078e00ff */ /*0020*/ MOV R3, c[0x0][0x17c] ; /* 0x00005f0000037a02 */ /* 0x000fe20000000f00 */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fc80000000a00 */ /*0040*/ LDG.E R0, [R2.64] ; /* 0x0000000602007981 */ /* 0x000ea2000c1e1900 */ /*0050*/ I2F.U32.RP R6, c[0x0][0xc] ; /* 0x0000030000067b06 */ /* 0x000e220000209000 */ /*0060*/ ISETP.NE.U32.AND P2, PT, RZ, c[0x0][0xc], PT ; /* 0x00000300ff007a0c */ /* 0x000fce0003f45070 */ /*0070*/ MUFU.RCP R6, R6 ; /* 0x0000000600067308 */ /* 0x001e240000001000 */ /*0080*/ IADD3 R4, R6, 0xffffffe, RZ ; /* 0x0ffffffe06047810 */ /* 0x001fcc0007ffe0ff */ /*0090*/ F2I.FTZ.U32.TRUNC.NTZ R5, R4 ; /* 0x0000000400057305 */ /* 0x000064000021f000 */ /*00a0*/ HFMA2.MMA R4, -RZ, RZ, 0, 0 ; /* 0x00000000ff047435 */ /* 0x001fe200000001ff */ /*00b0*/ IMAD.MOV R7, RZ, RZ, -R5 ; /* 0x000000ffff077224 */ /* 0x002fc800078e0a05 */ /*00c0*/ IMAD R7, R7, c[0x0][0xc], RZ ; /* 0x0000030007077a24 */ /* 0x000fca00078e02ff */ /*00d0*/ IMAD.HI.U32 R5, R5, R7, R4 ; /* 0x0000000705057227 */ /* 0x000fe400078e0004 */ /*00e0*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e280000002500 */ /*00f0*/ IMAD.HI.U32 R5, R5, R0, RZ ; /* 0x0000000005057227 */ /* 0x004fc800078e00ff */ /*0100*/ IMAD.MOV R7, RZ, RZ, -R5 ; /* 0x000000ffff077224 */ /* 0x000fc800078e0a05 */ /*0110*/ IMAD R0, R7, c[0x0][0xc], R0 ; /* 0x0000030007007a24 */ /* 0x000fca00078e0200 */ /*0120*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0xc], PT ; /* 0x0000030000007a0c */ /* 0x000fda0003f06070 */ /*0130*/ @P0 IADD3 R0, R0, -c[0x0][0xc], RZ ; /* 0x8000030000000a10 */ /* 0x000fe40007ffe0ff */ /*0140*/ @P0 IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105050810 */ /* 0x000fe40007ffe0ff */ /*0150*/ ISETP.GE.U32.AND P1, PT, R0, c[0x0][0xc], PT ; /* 0x0000030000007a0c */ /* 0x000fda0003f26070 */ /*0160*/ @P1 IADD3 R5, R5, 0x1, RZ ; /* 0x0000000105051810 */ /* 0x000fe40007ffe0ff */ /*0170*/ @!P2 LOP3.LUT R5, RZ, c[0x0][0xc], RZ, 0x33, !PT ; /* 0x00000300ff05aa12 */ /* 0x000fca00078e33ff */ /*0180*/ IMAD R15, R5, R4, RZ ; /* 0x00000004050f7224 */ /* 0x001fca00078e02ff */ /*0190*/ IADD3 R0, R5, -0x1, R15 ; /* 0xffffffff05007810 */ /* 0x000fc80007ffe00f */ /*01a0*/ ISETP.GE.U32.AND P0, PT, R15, R0, PT ; /* 0x000000000f00720c */ /* 0x000fda0003f06070 */ /*01b0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*01c0*/ S2R R14, SR_TID.X ; /* 0x00000000000e7919 */ /* 0x000e240000002100 */ /*01d0*/ LDG.E R17, [R2.64] ; /* 0x0000000602117981 */ /* 0x000ea4000c1e1900 */ /*01e0*/ ISETP.GT.AND P0, PT, R17, RZ, PT ; /* 0x000000ff1100720c */ /* 0x004fda0003f04270 */ /*01f0*/ @!P0 MOV R21, RZ ; /* 0x000000ff00158202 */ /* 0x001fe20000000f00 */ /*0200*/ @!P0 BRA 0x6f0 ; /* 0x000004e000008947 */ /* 0x000fea0003800000 */ /*0210*/ IADD3 R4, R17.reuse, -0x1, RZ ; /* 0xffffffff11047810 */ /* 0x040fe20007ffe0ff */ /*0220*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0230*/ LOP3.LUT R16, R17.reuse, 0x3, RZ, 0xc0, !PT ; /* 0x0000000311107812 */ /* 0x040fe200078ec0ff */ /*0240*/ IMAD R18, R17, R15, RZ ; /* 0x0000000f11127224 */ /* 0x000fe200078e02ff */ /*0250*/ ISETP.GE.U32.AND P1, PT, R4, 0x3, PT ; /* 0x000000030400780c */ /* 0x000fe20003f26070 */ /*0260*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0270*/ ISETP.NE.AND P0, PT, R16, RZ, PT ; /* 0x000000ff1000720c */ /* 0x000fd60003f05270 */ /*0280*/ @!P1 BRA 0x530 ; /* 0x000002a000009947 */ /* 0x000fea0003800000 */ /*0290*/ HFMA2.MMA R19, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff137435 */ /* 0x000fe200000001ff */ /*02a0*/ IMAD.IADD R22, R17.reuse, 0x1, R14.reuse ; /* 0x0000000111167824 */ /* 0x141fe200078e020e */ /*02b0*/ LEA R20, R17.reuse, R14.reuse, 0x1 ; /* 0x0000000e11147211 */ /* 0x0c0fe200078e08ff */ /*02c0*/ IMAD R24, R17, 0x3, R14 ; /* 0x0000000311187824 */ /* 0x000fe200078e020e */ /*02d0*/ MOV R26, R14 ; /* 0x0000000e001a7202 */ /* 0x000fe20000000f00 */ /*02e0*/ IMAD.IADD R27, R16, 0x1, -R17 ; /* 0x00000001101b7824 */ /* 0x000fe200078e0a11 */ /*02f0*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fc80008000000 */ /*0300*/ IMAD.WIDE R4, R18, R19, c[0x0][0x160] ; /* 0x0000580012047625 */ /* 0x000fc800078e0213 */ /*0310*/ IMAD.WIDE.U32 R6, R26, R19.reuse, c[0x0][0x168] ; /* 0x00005a001a067625 */ /* 0x080fe200078e0013 */ /*0320*/ LDG.E R23, [R4.64] ; /* 0x0000000604177981 */ /* 0x0010a6000c1e1900 */ /*0330*/ IMAD.WIDE.U32 R8, R22, R19.reuse, c[0x0][0x168] ; /* 0x00005a0016087625 */ /* 0x080fe200078e0013 */ /*0340*/ LDG.E R25, [R4.64+0x4] ; /* 0x0000040604197981 */ /* 0x0000e8000c1e1900 */ /*0350*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea8000c1e1900 */ /*0360*/ LDG.E R8, [R8.64] ; /* 0x0000000608087981 */ /* 0x000ee2000c1e1900 */ /*0370*/ IMAD.WIDE.U32 R10, R20, R19, c[0x0][0x168] ; /* 0x00005a00140a7625 */ /* 0x000fc600078e0013 */ /*0380*/ LDG.E R29, [R4.64+0x8] ; /* 0x00000806041d7981 */ /* 0x000128000c1e1900 */ /*0390*/ LDG.E R10, [R10.64] ; /* 0x000000060a0a7981 */ /* 0x000f22000c1e1900 */ /*03a0*/ IMAD.WIDE.U32 R12, R24, R19, c[0x0][0x168] ; /* 0x00005a00180c7625 */ /* 0x000fc600078e0013 */ /*03b0*/ LDG.E R28, [R4.64+0xc] ; /* 0x00000c06041c7981 */ /* 0x000168000c1e1900 */ /*03c0*/ LDG.E R12, [R12.64] ; /* 0x000000060c0c7981 */ /* 0x000f62000c1e1900 */ /*03d0*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea20000201400 */ /*03e0*/ UIADD3 UR4, UR4, 0x4, URZ ; /* 0x0000000404047890 */ /* 0x000fe2000fffe03f */ /*03f0*/ IADD3 R4, P2, R4, 0x10, RZ ; /* 0x0000001004047810 */ /* 0x001fe20007f5e0ff */ /*0400*/ IMAD R24, R17.reuse, 0x4, R24 ; /* 0x0000000411187824 */ /* 0x040fe200078e0218 */ /*0410*/ LEA R22, R17.reuse, R22, 0x2 ; /* 0x0000001611167211 */ /* 0x040fe400078e10ff */ /*0420*/ LEA R20, R17.reuse, R20, 0x2 ; /* 0x0000001411147211 */ /* 0x040fe200078e10ff */ /*0430*/ IMAD.X R5, RZ, RZ, R5, P2 ; /* 0x000000ffff057224 */ /* 0x000fe200010e0605 */ /*0440*/ LEA R26, R17, R26, 0x2 ; /* 0x0000001a111a7211 */ /* 0x000fe200078e10ff */ /*0450*/ FFMA R6, R6, R23, R21 ; /* 0x0000001706067223 */ /* 0x004fcc0000000015 */ /*0460*/ F2I.TRUNC.NTZ R6, R6 ; /* 0x0000000600067305 */ /* 0x000e30000020f100 */ /*0470*/ I2F R7, R6 ; /* 0x0000000600077306 */ /* 0x001ee40000201400 */ /*0480*/ FFMA R7, R8, R25, R7 ; /* 0x0000001908077223 */ /* 0x008fcc0000000007 */ /*0490*/ F2I.TRUNC.NTZ R7, R7 ; /* 0x0000000700077305 */ /* 0x000e30000020f100 */ /*04a0*/ I2F R8, R7 ; /* 0x0000000700087306 */ /* 0x001f240000201400 */ /*04b0*/ FFMA R8, R10, R29, R8 ; /* 0x0000001d0a087223 */ /* 0x010fcc0000000008 */ /*04c0*/ F2I.TRUNC.NTZ R8, R8 ; /* 0x0000000800087305 */ /* 0x000e22000020f100 */ /*04d0*/ IADD3 R10, R27, UR4, RZ ; /* 0x000000041b0a7c10 */ /* 0x000fce000fffe0ff */ /*04e0*/ I2F R9, R8 ; /* 0x0000000800097306 */ /* 0x001f620000201400 */ /*04f0*/ ISETP.NE.AND P1, PT, R10, RZ, PT ; /* 0x000000ff0a00720c */ /* 0x000fe20003f25270 */ /*0500*/ FFMA R9, R12, R28, R9 ; /* 0x0000001c0c097223 */ /* 0x020fcc0000000009 */ /*0510*/ F2I.TRUNC.NTZ R21, R9 ; /* 0x0000000900157305 */ /* 0x00006c000020f100 */ /*0520*/ @P1 BRA 0x310 ; /* 0xfffffde000001947 */ /* 0x000fea000383ffff */ /*0530*/ @!P0 BRA 0x6f0 ; /* 0x000001b000008947 */ /* 0x000fea0003800000 */ /*0540*/ HFMA2.MMA R13, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0d7435 */ /* 0x000fe200000001ff */ /*0550*/ IADD3 R4, R18, UR4, RZ ; /* 0x0000000412047c10 */ /* 0x000fe2000fffe0ff */ /*0560*/ IMAD R8, R17, UR4, R14 ; /* 0x0000000411087c24 */ /* 0x001fd0000f8e020e */ /*0570*/ IMAD.WIDE.U32 R6, R8, R13, c[0x0][0x168] ; /* 0x00005a0008067625 */ /* 0x000fc800078e000d */ /*0580*/ IMAD.WIDE R4, R4, R13, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e020d */ /*0590*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea8000c1e1900 */ /*05a0*/ LDG.E R9, [R4.64] ; /* 0x0000000604097981 */ /* 0x000ea2000c1e1900 */ /*05b0*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea20000201400 */ /*05c0*/ ISETP.NE.AND P0, PT, R16, 0x1, PT ; /* 0x000000011000780c */ /* 0x000fe20003f05270 */ /*05d0*/ FFMA R9, R6, R9, R21 ; /* 0x0000000906097223 */ /* 0x004fcc0000000015 */ /*05e0*/ F2I.TRUNC.NTZ R21, R9 ; /* 0x0000000900157305 */ /* 0x00006c000020f100 */ /*05f0*/ @!P0 BRA 0x6f0 ; /* 0x000000f000008947 */ /* 0x000fea0003800000 */ /*0600*/ IMAD.IADD R8, R17, 0x1, R8 ; /* 0x0000000111087824 */ /* 0x000fe200078e0208 */ /*0610*/ LDG.E R11, [R4.64+0x4] ; /* 0x00000406040b7981 */ /* 0x000ea6000c1e1900 */ /*0620*/ IMAD.WIDE.U32 R6, R8, R13, c[0x0][0x168] ; /* 0x00005a0008067625 */ /* 0x000fcc00078e000d */ /*0630*/ LDG.E R6, [R6.64] ; /* 0x0000000606067981 */ /* 0x000ea2000c1e1900 */ /*0640*/ ISETP.NE.AND P0, PT, R16, 0x2, PT ; /* 0x000000021000780c */ /* 0x000fda0003f05270 */ /*0650*/ @P0 IADD3 R8, R17, R8, RZ ; /* 0x0000000811080210 */ /* 0x000fca0007ffe0ff */ /*0660*/ @P0 IMAD.WIDE.U32 R8, R8, R13, c[0x0][0x168] ; /* 0x00005a0008080625 */ /* 0x001fe400078e000d */ /*0670*/ @P0 LDG.E R13, [R4.64+0x8] ; /* 0x00000806040d0981 */ /* 0x000ee8000c1e1900 */ /*0680*/ @P0 LDG.E R8, [R8.64] ; /* 0x0000000608080981 */ /* 0x000ee2000c1e1900 */ /*0690*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002ea40000201400 */ /*06a0*/ FFMA R11, R6, R11, R21 ; /* 0x0000000b060b7223 */ /* 0x004fcc0000000015 */ /*06b0*/ F2I.TRUNC.NTZ R21, R11 ; /* 0x0000000b00157305 */ /* 0x000e30000020f100 */ /*06c0*/ @P0 I2F R10, R21 ; /* 0x00000015000a0306 */ /* 0x001ee40000201400 */ /*06d0*/ @P0 FFMA R13, R8, R13, R10 ; /* 0x0000000d080d0223 */ /* 0x008fcc000000000a */ /*06e0*/ @P0 F2I.TRUNC.NTZ R21, R13 ; /* 0x0000000d00150305 */ /* 0x000064000020f100 */ /*06f0*/ IMAD R4, R17, R15, R14 ; /* 0x0000000f11047224 */ /* 0x001fe200078e020e */ /*0700*/ I2F R21, R21 ; /* 0x0000001500157306 */ /* 0x002e220000201400 */ /*0710*/ IADD3 R15, R15, 0x1, RZ ; /* 0x000000010f0f7810 */ /* 0x000fe40007ffe0ff */ /*0720*/ MOV R5, 0x4 ; /* 0x0000000400057802 */ /* 0x000fe40000000f00 */ /*0730*/ ISETP.GE.U32.AND P0, PT, R15, R0, PT ; /* 0x000000000f00720c */ /* 0x000fc60003f06070 */ /*0740*/ IMAD.WIDE.U32 R4, R4, R5, c[0x0][0x170] ; /* 0x00005c0004047625 */ /* 0x000fca00078e0005 */ /*0750*/ STG.E [R4.64], R21 ; /* 0x0000001504007986 */ /* 0x0011ea000c101906 */ /*0760*/ @!P0 BRA 0x1d0 ; /* 0xfffffa6000008947 */ /* 0x000fea000383ffff */ /*0770*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0780*/ BRA 0x780; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0790*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0800*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0810*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0820*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0830*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0840*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0850*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0860*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0870*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z15MatrixMulDevicePfS_S_Pi .globl _Z15MatrixMulDevicePfS_S_Pi .p2align 8 .type _Z15MatrixMulDevicePfS_S_Pi,@function _Z15MatrixMulDevicePfS_S_Pi: s_clause 0x1 s_load_b32 s4, s[0:1], 0x20 s_load_b64 s[2:3], s[0:1], 0x18 s_waitcnt lgkmcnt(0) v_cvt_f32_u32_e32 v1, s4 s_load_b32 s12, s[2:3], 0x0 s_sub_i32 s3, 0, s4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_iflag_f32_e32 v1, v1 s_waitcnt_depctr 0xfff v_mul_f32_e32 v1, 0x4f7ffffe, v1 v_cvt_u32_f32_e32 v1, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_readfirstlane_b32 s2, v1 s_mul_i32 s3, s3, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_mul_hi_u32 s3, s2, s3 s_add_i32 s2, s2, s3 s_waitcnt lgkmcnt(0) s_mul_hi_u32 s2, s12, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(SALU_CYCLE_1) s_mul_i32 s3, s2, s4 s_add_i32 s5, s2, 1 s_sub_i32 s3, s12, s3 s_sub_i32 s6, s3, s4 s_cmp_ge_u32 s3, s4 s_cselect_b32 s2, s5, s2 s_cselect_b32 s3, s6, s3 s_add_i32 s5, s2, 1 s_cmp_ge_u32 s3, s4 s_cselect_b32 s8, s5, s2 s_add_i32 s2, s15, 1 s_mul_i32 s13, s8, s15 s_mul_i32 s14, s8, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s14, s14, -1 s_cmp_ge_u32 s13, s14 s_cbranch_scc1 .LBB0_6 s_clause 0x1 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[2:3], s[0:1], 0x10 s_cmp_gt_i32 s12, 0 v_mov_b32_e32 v2, 0 s_cselect_b32 s0, -1, 0 s_mul_i32 s1, s8, s15 v_cndmask_b32_e64 v1, 0, 1, s0 s_mul_i32 s8, s1, s12 s_delay_alu instid0(VALU_DEP_1) v_cmp_ne_u32_e64 s0, 1, v1 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: v_mad_u64_u32 v[4:5], null, s13, s12, v[0:1] v_mov_b32_e32 v5, v2 v_cvt_f32_i32_e32 v1, v3 s_add_i32 s13, s13, 1 s_add_i32 s8, s8, s12 s_cmp_lt_u32 s13, s14 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[4:5] s_waitcnt lgkmcnt(0) v_add_co_u32 v3, vcc_lo, s2, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s3, v5, vcc_lo global_store_b32 v[3:4], v1, off s_cbranch_scc0 .LBB0_6 .LBB0_3: v_mov_b32_e32 v3, v2 s_delay_alu instid0(VALU_DEP_2) s_and_b32 vcc_lo, exec_lo, s0 s_cbranch_vccnz .LBB0_2 s_ashr_i32 s9, s8, 31 v_mov_b32_e32 v3, 0 s_lshl_b64 s[10:11], s[8:9], 2 v_mov_b32_e32 v1, v0 s_waitcnt lgkmcnt(0) s_add_u32 s10, s4, s10 s_addc_u32 s11, s5, s11 s_mov_b32 s1, s12 .p2align 6 .LBB0_5: v_lshlrev_b64 v[4:5], 2, v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_cvt_f32_i32_e32 v3, v3 s_add_i32 s1, s1, -1 v_add_nc_u32_e32 v1, s12, v1 v_add_co_u32 v4, vcc_lo, s6, v4 s_delay_alu instid0(VALU_DEP_4) v_add_co_ci_u32_e32 v5, vcc_lo, s7, v5, vcc_lo global_load_b32 v6, v2, s[10:11] global_load_b32 v4, v[4:5], off s_add_u32 s10, s10, 4 s_addc_u32 s11, s11, 0 s_cmp_lg_u32 s1, 0 s_waitcnt vmcnt(0) v_fmac_f32_e32 v3, v6, v4 s_delay_alu instid0(VALU_DEP_1) v_cvt_i32_f32_e32 v3, v3 s_cbranch_scc1 .LBB0_5 s_branch .LBB0_2 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z15MatrixMulDevicePfS_S_Pi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z15MatrixMulDevicePfS_S_Pi, .Lfunc_end0-_Z15MatrixMulDevicePfS_S_Pi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: hidden_block_count_x - .offset: 36 .size: 4 .value_kind: hidden_block_count_y - .offset: 40 .size: 4 .value_kind: hidden_block_count_z - .offset: 44 .size: 2 .value_kind: hidden_group_size_x - .offset: 46 .size: 2 .value_kind: hidden_group_size_y - .offset: 48 .size: 2 .value_kind: hidden_group_size_z - .offset: 50 .size: 2 .value_kind: hidden_remainder_x - .offset: 52 .size: 2 .value_kind: hidden_remainder_y - .offset: 54 .size: 2 .value_kind: hidden_remainder_z - .offset: 72 .size: 8 .value_kind: hidden_global_offset_x - .offset: 80 .size: 8 .value_kind: hidden_global_offset_y - .offset: 88 .size: 8 .value_kind: hidden_global_offset_z - .offset: 96 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z15MatrixMulDevicePfS_S_Pi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z15MatrixMulDevicePfS_S_Pi.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 7 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00003130_00000000-6_MatrixMulDevice.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi .type _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi, @function _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z15MatrixMulDevicePfS_S_Pi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi, .-_Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi .globl _Z15MatrixMulDevicePfS_S_Pi .type _Z15MatrixMulDevicePfS_S_Pi, @function _Z15MatrixMulDevicePfS_S_Pi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z15MatrixMulDevicePfS_S_PiPfS_S_Pi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z15MatrixMulDevicePfS_S_Pi, .-_Z15MatrixMulDevicePfS_S_Pi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z15MatrixMulDevicePfS_S_Pi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z15MatrixMulDevicePfS_S_Pi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "MatrixMulDevice.hip" .globl _Z30__device_stub__MatrixMulDevicePfS_S_Pi # -- Begin function _Z30__device_stub__MatrixMulDevicePfS_S_Pi .p2align 4, 0x90 .type _Z30__device_stub__MatrixMulDevicePfS_S_Pi,@function _Z30__device_stub__MatrixMulDevicePfS_S_Pi: # @_Z30__device_stub__MatrixMulDevicePfS_S_Pi .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movq %rcx, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z15MatrixMulDevicePfS_S_Pi, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z30__device_stub__MatrixMulDevicePfS_S_Pi, .Lfunc_end0-_Z30__device_stub__MatrixMulDevicePfS_S_Pi .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z15MatrixMulDevicePfS_S_Pi, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z15MatrixMulDevicePfS_S_Pi,@object # @_Z15MatrixMulDevicePfS_S_Pi .section .rodata,"a",@progbits .globl _Z15MatrixMulDevicePfS_S_Pi .p2align 3, 0x0 _Z15MatrixMulDevicePfS_S_Pi: .quad _Z30__device_stub__MatrixMulDevicePfS_S_Pi .size _Z15MatrixMulDevicePfS_S_Pi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z15MatrixMulDevicePfS_S_Pi" .size .L__unnamed_1, 28 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z30__device_stub__MatrixMulDevicePfS_S_Pi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z15MatrixMulDevicePfS_S_Pi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); cudaMalloc( (void**)&ad, csize ); cudaMalloc( (void**)&bd, isize ); cudaMemcpy( ad, a, csize, cudaMemcpyHostToDevice ); cudaMemcpy( bd, b, isize, cudaMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); cudaMemcpy( a, ad, csize, cudaMemcpyDeviceToHost ); cudaFree( ad ); cudaFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
code for sm_80 Function : _Z5helloPiS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R4, R4, c[0x0][0x0], R3 ; /* 0x0000000004047a24 */ /* 0x001fca00078e0203 */ /*0060*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x000fcc00078e0205 */ /*0070*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*0080*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fc800078e0205 */ /*0090*/ IMAD R7, R2, R2, RZ ; /* 0x0000000202077224 */ /* 0x004fc800078e02ff */ /*00a0*/ IMAD R7, R2, R7, RZ ; /* 0x0000000702077224 */ /* 0x000fca00078e02ff */ /*00b0*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00d0*/ BRA 0xd0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); cudaMalloc( (void**)&ad, csize ); cudaMalloc( (void**)&bd, isize ); cudaMemcpy( ad, a, csize, cudaMemcpyHostToDevice ); cudaMemcpy( bd, b, isize, cudaMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); cudaMemcpy( a, ad, csize, cudaMemcpyDeviceToHost ); cudaFree( ad ); cudaFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
.file "tmpxft_001a7480_00000000-6_hello.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z5helloPiS_PiS_ .type _Z26__device_stub__Z5helloPiS_PiS_, @function _Z26__device_stub__Z5helloPiS_PiS_: .LFB2082: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z5helloPiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z26__device_stub__Z5helloPiS_PiS_, .-_Z26__device_stub__Z5helloPiS_PiS_ .globl _Z5helloPiS_ .type _Z5helloPiS_, @function _Z5helloPiS_: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z5helloPiS_PiS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5helloPiS_, .-_Z5helloPiS_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "input " .LC1: .string "%d " .LC2: .string "\n" .LC3: .string "blocks = %d cores = %d\n" .LC4: .string "ouput " .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r13 movslq %eax, %r15 leaq 0(,%r15,4), %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %rbp movq %rbx, %rdi call malloc@PLT movq %rax, %r12 testl %r13d, %r13d jle .L12 leal -1(%r13), %r14d movl $0, %ebx .L13: movl $0, 0(%rbp,%rbx,4) call rand@PLT movslq %eax, %rdx imulq $1374389535, %rdx, %rdx sarq $37, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx imull $100, %edx, %edx subl %edx, %eax addl $1, %eax movl %eax, (%r12,%rbx,4) movq %rbx, %rax addq $1, %rbx cmpq %r14, %rax jne .L13 leal 0(,%r15,4), %eax movl %eax, 12(%rsp) leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rbx leaq 4(%r12,%r14,4), %r15 leaq .LC1(%rip), %r14 .L14: movl (%rbx), %edx movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r15, %rbx jne .L14 .L18: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movslq 12(%rsp), %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbx, %rdx movq %r12, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $512, %ecx movl $16, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $256, 44(%rsp) movl $1, 48(%rsp) leal 510(%r13), %eax movl %r13d, %edx addl $255, %edx cmovns %edx, %eax sarl $8, %eax movl %eax, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L15: movl $2, %ecx movq %rbx, %rdx movq 16(%rsp), %rsi movq %rbp, %rdi call cudaMemcpy@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT testl %r13d, %r13d jle .L16 movq %rbp, %rbx leal -1(%r13), %eax leaq 4(%rbp,%rax,4), %r14 leaq .LC1(%rip), %r13 .L17: movl (%rbx), %edx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L17 .L16: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L26 movl $0, %eax addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L25: .cfi_restore_state movq 24(%rsp), %rsi movq 16(%rsp), %rdi call _Z26__device_stub__Z5helloPiS_PiS_ jmp .L15 .L12: leal 0(,%r15,4), %eax movl %eax, 12(%rsp) leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L18 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC5: .string "_Z5helloPiS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z5helloPiS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); cudaMalloc( (void**)&ad, csize ); cudaMalloc( (void**)&bd, isize ); cudaMemcpy( ad, a, csize, cudaMemcpyHostToDevice ); cudaMemcpy( bd, b, isize, cudaMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); cudaMemcpy( a, ad, csize, cudaMemcpyDeviceToHost ); cudaFree( ad ); cudaFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
#include <hip/hip_runtime.h> #include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); hipMalloc( (void**)&ad, csize ); hipMalloc( (void**)&bd, isize ); hipMemcpy( ad, a, csize, hipMemcpyHostToDevice ); hipMemcpy( bd, b, isize, hipMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); hipMemcpy( a, ad, csize, hipMemcpyDeviceToHost ); hipFree( ad ); hipFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); hipMalloc( (void**)&ad, csize ); hipMalloc( (void**)&bd, isize ); hipMemcpy( ad, a, csize, hipMemcpyHostToDevice ); hipMemcpy( bd, b, isize, hipMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); hipMemcpy( a, ad, csize, hipMemcpyDeviceToHost ); hipFree( ad ); hipFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5helloPiS_ .globl _Z5helloPiS_ .p2align 8 .type _Z5helloPiS_,@function _Z5helloPiS_: s_clause 0x1 s_load_b32 s4, s[0:1], 0x1c s_load_b128 s[0:3], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s4, v[0:1] v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] v_add_co_u32 v2, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s3, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_load_b32 v2, v[2:3], off s_waitcnt vmcnt(0) v_mul_lo_u32 v3, v2, v2 s_delay_alu instid0(VALU_DEP_1) v_mul_lo_u32 v2, v3, v2 global_store_b32 v[0:1], v2, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5helloPiS_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 4 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z5helloPiS_, .Lfunc_end0-_Z5helloPiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5helloPiS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5helloPiS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> //const int N = 16; const int blocksize = 16; __global__ void hello(int *a, int *b) { int z = blockDim.x * blockIdx.x + threadIdx.x; a[z] = b[z]* b[z] * b[z]; //a[threadIdx.x] = 5; } int main(int argc, char** argv) { int SIZE = atoi(argv[1]); int *a =(int*) malloc(SIZE * sizeof(int)); int *b =(int*) malloc(SIZE * sizeof(int)); for(int i=0; i<SIZE; i++) { a[i]=0; b[i]=rand()%100+1; } // int a[N] = {}; //int b[N] = {16, 10, 6, 0, -11, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; int *ad; int *bd; const int csize = SIZE*sizeof(int); const int isize = SIZE*sizeof(int); printf("input "); for (int i=0;i<SIZE;i++) printf("%d ", b[i]); printf("\n"); hipMalloc( (void**)&ad, csize ); hipMalloc( (void**)&bd, isize ); hipMemcpy( ad, a, csize, hipMemcpyHostToDevice ); hipMemcpy( bd, b, isize, hipMemcpyHostToDevice ); dim3 dimBlock( blocksize, 1 ); dim3 dimGrid( 1, 1 ); //hello<<<dimGrid, dimBlock>>>(ad, bd); int cores = 512; //int blocks = SIZE/cores + 16; int blocks = 16; int threadsPerBlock = 256; //int blocksPerGrid =(SIZE + threadsPerBlock – 1) / threadsPerBlock; int blocksPerGrid = ( SIZE + threadsPerBlock -1) / threadsPerBlock; printf("blocks = %d cores = %d\n", blocks, cores); hello<<<blocksPerGrid, threadsPerBlock>>>(ad, bd); hipMemcpy( a, ad, csize, hipMemcpyDeviceToHost ); hipFree( ad ); hipFree( bd ); printf("ouput "); for (int i=0;i<SIZE;i++) printf("%d ", a[i]); printf("\n"); //printf("%s\n", a); free(a); free(b); return EXIT_SUCCESS; }
.text .file "hello.hip" .globl _Z20__device_stub__helloPiS_ # -- Begin function _Z20__device_stub__helloPiS_ .p2align 4, 0x90 .type _Z20__device_stub__helloPiS_,@function _Z20__device_stub__helloPiS_: # @_Z20__device_stub__helloPiS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z5helloPiS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z20__device_stub__helloPiS_, .Lfunc_end0-_Z20__device_stub__helloPiS_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $96, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r15 movslq %r15d, %r12 leaq (,%r12,4), %r14 movq %r14, %rdi callq malloc movq %rax, %rbx movq %r14, %rdi callq malloc movq %rax, %r14 testl %r12d, %r12d jle .LBB1_3 # %bb.1: # %.lr.ph.preheader movl %r15d, %r12d xorl %r13d, %r13d .p2align 4, 0x90 .LBB1_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl $0, (%rbx,%r13,4) callq rand cltq imulq $1374389535, %rax, %rcx # imm = 0x51EB851F movq %rcx, %rdx shrq $63, %rdx sarq $37, %rcx addl %edx, %ecx imull $100, %ecx, %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%r14,%r13,4) incq %r13 cmpq %r13, %r12 jne .LBB1_2 .LBB1_3: # %._crit_edge movl $.L.str, %edi xorl %eax, %eax callq printf testl %r15d, %r15d jle .LBB1_6 # %bb.4: # %.lr.ph56.preheader movl %r15d, %r12d xorl %r13d, %r13d .p2align 4, 0x90 .LBB1_5: # %.lr.ph56 # =>This Inner Loop Header: Depth=1 movl (%r14,%r13,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB1_5 .LBB1_6: # %._crit_edge57 movl $10, %edi callq putchar@PLT movq %r15, %r12 shlq $34, %r12 sarq $32, %r12 movq %rsp, %rdi movq %r12, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r12, %rsi callq hipMalloc movq (%rsp), %rdi movq %rbx, %rsi movq %r12, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq %r14, %rsi movq %r12, %rdx movl $1, %ecx callq hipMemcpy leal 255(%r15), %eax leal 510(%r15), %r13d testl %eax, %eax cmovnsl %eax, %r13d sarl $8, %r13d movl $.L.str.3, %edi movl $16, %esi movl $512, %edx # imm = 0x200 xorl %eax, %eax callq printf movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %r13 orq $256, %rdx # imm = 0x100 movq %r13, %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_8 # %bb.7: movq (%rsp), %rax movq 8(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z5helloPiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_8: movq (%rsp), %rsi movq %rbx, %rdi movq %r12, %rdx movl $2, %ecx callq hipMemcpy movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movl $.L.str.4, %edi xorl %eax, %eax callq printf testl %r15d, %r15d jle .LBB1_11 # %bb.9: # %.lr.ph60.preheader movl %r15d, %r15d xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_10: # %.lr.ph60 # =>This Inner Loop Header: Depth=1 movl (%rbx,%r12,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r12 cmpq %r12, %r15 jne .LBB1_10 .LBB1_11: # %._crit_edge61 movl $10, %edi callq putchar@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $96, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5helloPiS_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z5helloPiS_,@object # @_Z5helloPiS_ .section .rodata,"a",@progbits .globl _Z5helloPiS_ .p2align 3, 0x0 _Z5helloPiS_: .quad _Z20__device_stub__helloPiS_ .size _Z5helloPiS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "input " .size .L.str, 7 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d " .size .L.str.1, 4 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "blocks = %d cores = %d\n" .size .L.str.3, 24 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "ouput " .size .L.str.4, 7 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5helloPiS_" .size .L__unnamed_1, 13 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__helloPiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5helloPiS_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z5helloPiS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R4, R4, c[0x0][0x0], R3 ; /* 0x0000000004047a24 */ /* 0x001fca00078e0203 */ /*0060*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x000fcc00078e0205 */ /*0070*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*0080*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fc800078e0205 */ /*0090*/ IMAD R7, R2, R2, RZ ; /* 0x0000000202077224 */ /* 0x004fc800078e02ff */ /*00a0*/ IMAD R7, R2, R7, RZ ; /* 0x0000000702077224 */ /* 0x000fca00078e02ff */ /*00b0*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00d0*/ BRA 0xd0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5helloPiS_ .globl _Z5helloPiS_ .p2align 8 .type _Z5helloPiS_,@function _Z5helloPiS_: s_clause 0x1 s_load_b32 s4, s[0:1], 0x1c s_load_b128 s[0:3], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s4, v[0:1] v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] v_add_co_u32 v2, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s3, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_load_b32 v2, v[2:3], off s_waitcnt vmcnt(0) v_mul_lo_u32 v3, v2, v2 s_delay_alu instid0(VALU_DEP_1) v_mul_lo_u32 v2, v3, v2 global_store_b32 v[0:1], v2, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5helloPiS_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 4 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z5helloPiS_, .Lfunc_end0-_Z5helloPiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5helloPiS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5helloPiS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_001a7480_00000000-6_hello.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z5helloPiS_PiS_ .type _Z26__device_stub__Z5helloPiS_PiS_, @function _Z26__device_stub__Z5helloPiS_PiS_: .LFB2082: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z5helloPiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z26__device_stub__Z5helloPiS_PiS_, .-_Z26__device_stub__Z5helloPiS_PiS_ .globl _Z5helloPiS_ .type _Z5helloPiS_, @function _Z5helloPiS_: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z5helloPiS_PiS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5helloPiS_, .-_Z5helloPiS_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "input " .LC1: .string "%d " .LC2: .string "\n" .LC3: .string "blocks = %d cores = %d\n" .LC4: .string "ouput " .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r13 movslq %eax, %r15 leaq 0(,%r15,4), %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %rbp movq %rbx, %rdi call malloc@PLT movq %rax, %r12 testl %r13d, %r13d jle .L12 leal -1(%r13), %r14d movl $0, %ebx .L13: movl $0, 0(%rbp,%rbx,4) call rand@PLT movslq %eax, %rdx imulq $1374389535, %rdx, %rdx sarq $37, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx imull $100, %edx, %edx subl %edx, %eax addl $1, %eax movl %eax, (%r12,%rbx,4) movq %rbx, %rax addq $1, %rbx cmpq %r14, %rax jne .L13 leal 0(,%r15,4), %eax movl %eax, 12(%rsp) leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r12, %rbx leaq 4(%r12,%r14,4), %r15 leaq .LC1(%rip), %r14 .L14: movl (%rbx), %edx movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r15, %rbx jne .L14 .L18: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movslq 12(%rsp), %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbx, %rdx movq %r12, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $512, %ecx movl $16, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $256, 44(%rsp) movl $1, 48(%rsp) leal 510(%r13), %eax movl %r13d, %edx addl $255, %edx cmovns %edx, %eax sarl $8, %eax movl %eax, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L15: movl $2, %ecx movq %rbx, %rdx movq 16(%rsp), %rsi movq %rbp, %rdi call cudaMemcpy@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT testl %r13d, %r13d jle .L16 movq %rbp, %rbx leal -1(%r13), %eax leaq 4(%rbp,%rax,4), %r14 leaq .LC1(%rip), %r13 .L17: movl (%rbx), %edx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L17 .L16: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L26 movl $0, %eax addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L25: .cfi_restore_state movq 24(%rsp), %rsi movq 16(%rsp), %rdi call _Z26__device_stub__Z5helloPiS_PiS_ jmp .L15 .L12: leal 0(,%r15,4), %eax movl %eax, 12(%rsp) leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L18 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC5: .string "_Z5helloPiS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z5helloPiS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello.hip" .globl _Z20__device_stub__helloPiS_ # -- Begin function _Z20__device_stub__helloPiS_ .p2align 4, 0x90 .type _Z20__device_stub__helloPiS_,@function _Z20__device_stub__helloPiS_: # @_Z20__device_stub__helloPiS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z5helloPiS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z20__device_stub__helloPiS_, .Lfunc_end0-_Z20__device_stub__helloPiS_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $96, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r15 movslq %r15d, %r12 leaq (,%r12,4), %r14 movq %r14, %rdi callq malloc movq %rax, %rbx movq %r14, %rdi callq malloc movq %rax, %r14 testl %r12d, %r12d jle .LBB1_3 # %bb.1: # %.lr.ph.preheader movl %r15d, %r12d xorl %r13d, %r13d .p2align 4, 0x90 .LBB1_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl $0, (%rbx,%r13,4) callq rand cltq imulq $1374389535, %rax, %rcx # imm = 0x51EB851F movq %rcx, %rdx shrq $63, %rdx sarq $37, %rcx addl %edx, %ecx imull $100, %ecx, %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%r14,%r13,4) incq %r13 cmpq %r13, %r12 jne .LBB1_2 .LBB1_3: # %._crit_edge movl $.L.str, %edi xorl %eax, %eax callq printf testl %r15d, %r15d jle .LBB1_6 # %bb.4: # %.lr.ph56.preheader movl %r15d, %r12d xorl %r13d, %r13d .p2align 4, 0x90 .LBB1_5: # %.lr.ph56 # =>This Inner Loop Header: Depth=1 movl (%r14,%r13,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB1_5 .LBB1_6: # %._crit_edge57 movl $10, %edi callq putchar@PLT movq %r15, %r12 shlq $34, %r12 sarq $32, %r12 movq %rsp, %rdi movq %r12, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r12, %rsi callq hipMalloc movq (%rsp), %rdi movq %rbx, %rsi movq %r12, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq %r14, %rsi movq %r12, %rdx movl $1, %ecx callq hipMemcpy leal 255(%r15), %eax leal 510(%r15), %r13d testl %eax, %eax cmovnsl %eax, %r13d sarl $8, %r13d movl $.L.str.3, %edi movl $16, %esi movl $512, %edx # imm = 0x200 xorl %eax, %eax callq printf movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %r13 orq $256, %rdx # imm = 0x100 movq %r13, %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_8 # %bb.7: movq (%rsp), %rax movq 8(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z5helloPiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_8: movq (%rsp), %rsi movq %rbx, %rdi movq %r12, %rdx movl $2, %ecx callq hipMemcpy movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movl $.L.str.4, %edi xorl %eax, %eax callq printf testl %r15d, %r15d jle .LBB1_11 # %bb.9: # %.lr.ph60.preheader movl %r15d, %r15d xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_10: # %.lr.ph60 # =>This Inner Loop Header: Depth=1 movl (%rbx,%r12,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r12 cmpq %r12, %r15 jne .LBB1_10 .LBB1_11: # %._crit_edge61 movl $10, %edi callq putchar@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $96, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5helloPiS_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z5helloPiS_,@object # @_Z5helloPiS_ .section .rodata,"a",@progbits .globl _Z5helloPiS_ .p2align 3, 0x0 _Z5helloPiS_: .quad _Z20__device_stub__helloPiS_ .size _Z5helloPiS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "input " .size .L.str, 7 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d " .size .L.str.1, 4 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "blocks = %d cores = %d\n" .size .L.str.3, 24 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "ouput " .size .L.str.4, 7 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5helloPiS_" .size .L__unnamed_1, 13 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__helloPiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5helloPiS_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/* This is a automatically generated test. Do not modify */ #include <stdio.h> #include <stdlib.h> #include <math.h> __global__ void compute(float comp, int var_1,int var_2,int var_3,float var_4,int var_5,float var_6,float var_7,float var_8,float var_9,float var_10,float var_11,float var_12,float var_13,float var_14,float var_15,float var_16,float var_17,float var_18,float var_19,float var_20,float var_21,float var_22) { for (int i=0; i < var_1; ++i) { for (int i=0; i < var_2; ++i) { for (int i=0; i < var_3; ++i) { if (comp == asinf(+1.6052E-29f * (var_4 - (+1.3418E-44f * -1.1849E-36f)))) { float tmp_1 = +1.4940E-23f; float tmp_2 = -1.6240E35f + -1.2672E-41f + +1.0623E34f + atan2f(logf(var_6 * -0.0f), (-1.4919E-42f * atan2f((-1.9619E-37f * (-1.6127E-37f + acosf(var_7 * -1.0424E-41f / -1.8312E-36f * var_8))), var_9 * (-1.5333E-43f * var_10 / powf(var_11 - coshf((var_12 - coshf(var_13 / -1.9055E20f))), -1.0315E9f))))); comp = tmp_2 * tmp_1 - +1.6798E36f * (+1.1643E34f - (var_14 * +1.4241E36f * (var_15 + var_16))); comp += var_17 / floorf(+0.0f); for (int i=0; i < var_5; ++i) { comp += +1.6689E-43f - asinf(-1.4086E35f / +0.0f * var_18 + var_19); } if (comp == (+1.3292E34f - var_20 * (-1.9025E-42f - var_21))) { comp = +1.9010E19f * var_22 * -1.4243E-5f * tanhf((-1.4812E-35f * +1.5328E36f)); } } } } } printf("%.17g\n", comp); } float* initPointer(float v) { float *ret = (float*) malloc(sizeof(float)*10); for(int i=0; i < 10; ++i) ret[i] = v; return ret; } int main(int argc, char** argv) { /* Program variables */ float tmp_1 = atof(argv[1]); int tmp_2 = atoi(argv[2]); int tmp_3 = atoi(argv[3]); int tmp_4 = atoi(argv[4]); float tmp_5 = atof(argv[5]); int tmp_6 = atoi(argv[6]); float tmp_7 = atof(argv[7]); float tmp_8 = atof(argv[8]); float tmp_9 = atof(argv[9]); float tmp_10 = atof(argv[10]); float tmp_11 = atof(argv[11]); float tmp_12 = atof(argv[12]); float tmp_13 = atof(argv[13]); float tmp_14 = atof(argv[14]); float tmp_15 = atof(argv[15]); float tmp_16 = atof(argv[16]); float tmp_17 = atof(argv[17]); float tmp_18 = atof(argv[18]); float tmp_19 = atof(argv[19]); float tmp_20 = atof(argv[20]); float tmp_21 = atof(argv[21]); float tmp_22 = atof(argv[22]); float tmp_23 = atof(argv[23]); compute<<<1,1>>>(tmp_1,tmp_2,tmp_3,tmp_4,tmp_5,tmp_6,tmp_7,tmp_8,tmp_9,tmp_10,tmp_11,tmp_12,tmp_13,tmp_14,tmp_15,tmp_16,tmp_17,tmp_18,tmp_19,tmp_20,tmp_21,tmp_22,tmp_23); cudaDeviceSynchronize(); return 0; }
.file "tmpxft_000990a0_00000000-6_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z11initPointerf .type _Z11initPointerf, @function _Z11initPointerf: .LFB2057: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movd %xmm0, %ebx movl $40, %edi call malloc@PLT movq %rax, %rdx leaq 40(%rax), %rcx .L4: movl %ebx, (%rdx) addq $4, %rdx cmpq %rcx, %rdx jne .L4 popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2057: .size _Z11initPointerf, .-_Z11initPointerf .globl _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff .type _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff, @function _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff: .LFB2083: .cfi_startproc endbr64 subq $312, %rsp .cfi_def_cfa_offset 320 movss %xmm0, 44(%rsp) movl %edi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movss %xmm1, 28(%rsp) movl %ecx, 24(%rsp) movss %xmm2, 20(%rsp) movss %xmm3, 16(%rsp) movss %xmm4, 12(%rsp) movss %xmm5, 8(%rsp) movss %xmm6, 4(%rsp) movss %xmm7, (%rsp) movq %fs:40, %rax movq %rax, 296(%rsp) xorl %eax, %eax leaq 44(%rsp), %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rax movq %rax, 120(%rsp) leaq 36(%rsp), %rax movq %rax, 128(%rsp) leaq 32(%rsp), %rax movq %rax, 136(%rsp) leaq 28(%rsp), %rax movq %rax, 144(%rsp) leaq 24(%rsp), %rax movq %rax, 152(%rsp) leaq 20(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 12(%rsp), %rax movq %rax, 176(%rsp) leaq 8(%rsp), %rax movq %rax, 184(%rsp) leaq 4(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 320(%rsp), %rax movq %rax, 208(%rsp) leaq 328(%rsp), %rax movq %rax, 216(%rsp) leaq 336(%rsp), %rax movq %rax, 224(%rsp) leaq 344(%rsp), %rax movq %rax, 232(%rsp) leaq 352(%rsp), %rax movq %rax, 240(%rsp) leaq 360(%rsp), %rax movq %rax, 248(%rsp) leaq 368(%rsp), %rax movq %rax, 256(%rsp) leaq 376(%rsp), %rax movq %rax, 264(%rsp) leaq 384(%rsp), %rax movq %rax, 272(%rsp) leaq 392(%rsp), %rax movq %rax, 280(%rsp) leaq 400(%rsp), %rax movq %rax, 288(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 296(%rsp), %rax subq %fs:40, %rax jne .L12 addq $312, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 328 pushq 56(%rsp) .cfi_def_cfa_offset 336 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z7computefiiififffffffffffffffff(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 320 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff, .-_Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff .globl _Z7computefiiififffffffffffffffff .type _Z7computefiiififffffffffffffffff, @function _Z7computefiiififffffffffffffffff: .LFB2084: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movss 192(%rsp), %xmm8 movss %xmm8, 80(%rsp) movss 184(%rsp), %xmm8 movss %xmm8, 72(%rsp) movss 176(%rsp), %xmm8 movss %xmm8, 64(%rsp) movss 168(%rsp), %xmm8 movss %xmm8, 56(%rsp) movss 160(%rsp), %xmm8 movss %xmm8, 48(%rsp) movss 152(%rsp), %xmm8 movss %xmm8, 40(%rsp) movss 144(%rsp), %xmm8 movss %xmm8, 32(%rsp) movss 136(%rsp), %xmm8 movss %xmm8, 24(%rsp) movss 128(%rsp), %xmm8 movss %xmm8, 16(%rsp) movss 120(%rsp), %xmm8 movss %xmm8, 8(%rsp) movss 112(%rsp), %xmm8 movss %xmm8, (%rsp) call _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff addq $104, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7computefiiififffffffffffffffff, .-_Z7computefiiififffffffffffffffff .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r14 .cfi_def_cfa_offset 16 .cfi_offset 14, -16 pushq %r13 .cfi_def_cfa_offset 24 .cfi_offset 13, -24 pushq %r12 .cfi_def_cfa_offset 32 .cfi_offset 12, -32 pushq %rbp .cfi_def_cfa_offset 40 .cfi_offset 6, -40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset 3, -48 subq $192, %rsp .cfi_def_cfa_offset 240 movq %rsi, %rbx movq 8(%rsi), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 152(%rsp) movq 16(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %rbp movq 24(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r12 movq 32(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r13 movq 40(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 144(%rsp) movq 48(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r14 movq 56(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 136(%rsp) movq 64(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 128(%rsp) movq 72(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 120(%rsp) movq 80(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 112(%rsp) movq 88(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 104(%rsp) movq 96(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 96(%rsp) movq 104(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 88(%rsp) movq 112(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 80(%rsp) movq 120(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 72(%rsp) movq 128(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 64(%rsp) movq 136(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 56(%rsp) movq 144(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 48(%rsp) movq 152(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 40(%rsp) movq 160(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 32(%rsp) movq 168(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 24(%rsp) movq 176(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 16(%rsp) movq 184(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 8(%rsp) movl $1, 180(%rsp) movl $1, 184(%rsp) movl $1, 168(%rsp) movl $1, 172(%rsp) movl $0, %r9d movl $0, %r8d movq 180(%rsp), %rdx movl $1, %ecx movq 168(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L18 .L16: call cudaDeviceSynchronize@PLT movl $0, %eax addq $192, %rsp .cfi_remember_state .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .L18: .cfi_restore_state pxor %xmm0, %xmm0 cvtsd2ss 152(%rsp), %xmm0 subq $96, %rsp .cfi_def_cfa_offset 336 pxor %xmm1, %xmm1 cvtsd2ss 104(%rsp), %xmm1 movss %xmm1, 80(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 112(%rsp), %xmm1 movss %xmm1, 72(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 120(%rsp), %xmm1 movss %xmm1, 64(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 128(%rsp), %xmm1 movss %xmm1, 56(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 136(%rsp), %xmm1 movss %xmm1, 48(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 144(%rsp), %xmm1 movss %xmm1, 40(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 152(%rsp), %xmm1 movss %xmm1, 32(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 160(%rsp), %xmm1 movss %xmm1, 24(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 168(%rsp), %xmm1 movss %xmm1, 16(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 176(%rsp), %xmm1 movss %xmm1, 8(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 184(%rsp), %xmm1 movss %xmm1, (%rsp) pxor %xmm7, %xmm7 cvtsd2ss 192(%rsp), %xmm7 pxor %xmm6, %xmm6 cvtsd2ss 200(%rsp), %xmm6 pxor %xmm5, %xmm5 cvtsd2ss 208(%rsp), %xmm5 pxor %xmm4, %xmm4 cvtsd2ss 216(%rsp), %xmm4 pxor %xmm3, %xmm3 cvtsd2ss 224(%rsp), %xmm3 pxor %xmm2, %xmm2 cvtsd2ss 232(%rsp), %xmm2 movl %r14d, %ecx pxor %xmm1, %xmm1 cvtsd2ss 240(%rsp), %xmm1 movl %r13d, %edx movl %r12d, %esi movl %ebp, %edi call _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff addq $96, %rsp .cfi_def_cfa_offset 240 jmp .L16 .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z7computefiiififffffffffffffffff" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z7computefiiififffffffffffffffff(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
/* This is a automatically generated test. Do not modify */ #include <stdio.h> #include <stdlib.h> #include <math.h> __global__ void compute(float comp, int var_1,int var_2,int var_3,float var_4,int var_5,float var_6,float var_7,float var_8,float var_9,float var_10,float var_11,float var_12,float var_13,float var_14,float var_15,float var_16,float var_17,float var_18,float var_19,float var_20,float var_21,float var_22) { for (int i=0; i < var_1; ++i) { for (int i=0; i < var_2; ++i) { for (int i=0; i < var_3; ++i) { if (comp == asinf(+1.6052E-29f * (var_4 - (+1.3418E-44f * -1.1849E-36f)))) { float tmp_1 = +1.4940E-23f; float tmp_2 = -1.6240E35f + -1.2672E-41f + +1.0623E34f + atan2f(logf(var_6 * -0.0f), (-1.4919E-42f * atan2f((-1.9619E-37f * (-1.6127E-37f + acosf(var_7 * -1.0424E-41f / -1.8312E-36f * var_8))), var_9 * (-1.5333E-43f * var_10 / powf(var_11 - coshf((var_12 - coshf(var_13 / -1.9055E20f))), -1.0315E9f))))); comp = tmp_2 * tmp_1 - +1.6798E36f * (+1.1643E34f - (var_14 * +1.4241E36f * (var_15 + var_16))); comp += var_17 / floorf(+0.0f); for (int i=0; i < var_5; ++i) { comp += +1.6689E-43f - asinf(-1.4086E35f / +0.0f * var_18 + var_19); } if (comp == (+1.3292E34f - var_20 * (-1.9025E-42f - var_21))) { comp = +1.9010E19f * var_22 * -1.4243E-5f * tanhf((-1.4812E-35f * +1.5328E36f)); } } } } } printf("%.17g\n", comp); } float* initPointer(float v) { float *ret = (float*) malloc(sizeof(float)*10); for(int i=0; i < 10; ++i) ret[i] = v; return ret; } int main(int argc, char** argv) { /* Program variables */ float tmp_1 = atof(argv[1]); int tmp_2 = atoi(argv[2]); int tmp_3 = atoi(argv[3]); int tmp_4 = atoi(argv[4]); float tmp_5 = atof(argv[5]); int tmp_6 = atoi(argv[6]); float tmp_7 = atof(argv[7]); float tmp_8 = atof(argv[8]); float tmp_9 = atof(argv[9]); float tmp_10 = atof(argv[10]); float tmp_11 = atof(argv[11]); float tmp_12 = atof(argv[12]); float tmp_13 = atof(argv[13]); float tmp_14 = atof(argv[14]); float tmp_15 = atof(argv[15]); float tmp_16 = atof(argv[16]); float tmp_17 = atof(argv[17]); float tmp_18 = atof(argv[18]); float tmp_19 = atof(argv[19]); float tmp_20 = atof(argv[20]); float tmp_21 = atof(argv[21]); float tmp_22 = atof(argv[22]); float tmp_23 = atof(argv[23]); compute<<<1,1>>>(tmp_1,tmp_2,tmp_3,tmp_4,tmp_5,tmp_6,tmp_7,tmp_8,tmp_9,tmp_10,tmp_11,tmp_12,tmp_13,tmp_14,tmp_15,tmp_16,tmp_17,tmp_18,tmp_19,tmp_20,tmp_21,tmp_22,tmp_23); cudaDeviceSynchronize(); return 0; }
/* This is a automatically generated test. Do not modify */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <math.h> __global__ void compute(float comp, int var_1,int var_2,int var_3,float var_4,int var_5,float var_6,float var_7,float var_8,float var_9,float var_10,float var_11,float var_12,float var_13,float var_14,float var_15,float var_16,float var_17,float var_18,float var_19,float var_20,float var_21,float var_22) { for (int i=0; i < var_1; ++i) { for (int i=0; i < var_2; ++i) { for (int i=0; i < var_3; ++i) { if (comp == asinf(+1.6052E-29f * (var_4 - (+1.3418E-44f * -1.1849E-36f)))) { float tmp_1 = +1.4940E-23f; float tmp_2 = -1.6240E35f + -1.2672E-41f + +1.0623E34f + atan2f(logf(var_6 * -0.0f), (-1.4919E-42f * atan2f((-1.9619E-37f * (-1.6127E-37f + acosf(var_7 * -1.0424E-41f / -1.8312E-36f * var_8))), var_9 * (-1.5333E-43f * var_10 / powf(var_11 - coshf((var_12 - coshf(var_13 / -1.9055E20f))), -1.0315E9f))))); comp = tmp_2 * tmp_1 - +1.6798E36f * (+1.1643E34f - (var_14 * +1.4241E36f * (var_15 + var_16))); comp += var_17 / floorf(+0.0f); for (int i=0; i < var_5; ++i) { comp += +1.6689E-43f - asinf(-1.4086E35f / +0.0f * var_18 + var_19); } if (comp == (+1.3292E34f - var_20 * (-1.9025E-42f - var_21))) { comp = +1.9010E19f * var_22 * -1.4243E-5f * tanhf((-1.4812E-35f * +1.5328E36f)); } } } } } printf("%.17g\n", comp); } float* initPointer(float v) { float *ret = (float*) malloc(sizeof(float)*10); for(int i=0; i < 10; ++i) ret[i] = v; return ret; } int main(int argc, char** argv) { /* Program variables */ float tmp_1 = atof(argv[1]); int tmp_2 = atoi(argv[2]); int tmp_3 = atoi(argv[3]); int tmp_4 = atoi(argv[4]); float tmp_5 = atof(argv[5]); int tmp_6 = atoi(argv[6]); float tmp_7 = atof(argv[7]); float tmp_8 = atof(argv[8]); float tmp_9 = atof(argv[9]); float tmp_10 = atof(argv[10]); float tmp_11 = atof(argv[11]); float tmp_12 = atof(argv[12]); float tmp_13 = atof(argv[13]); float tmp_14 = atof(argv[14]); float tmp_15 = atof(argv[15]); float tmp_16 = atof(argv[16]); float tmp_17 = atof(argv[17]); float tmp_18 = atof(argv[18]); float tmp_19 = atof(argv[19]); float tmp_20 = atof(argv[20]); float tmp_21 = atof(argv[21]); float tmp_22 = atof(argv[22]); float tmp_23 = atof(argv[23]); compute<<<1,1>>>(tmp_1,tmp_2,tmp_3,tmp_4,tmp_5,tmp_6,tmp_7,tmp_8,tmp_9,tmp_10,tmp_11,tmp_12,tmp_13,tmp_14,tmp_15,tmp_16,tmp_17,tmp_18,tmp_19,tmp_20,tmp_21,tmp_22,tmp_23); hipDeviceSynchronize(); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
/* This is a automatically generated test. Do not modify */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <math.h> __global__ void compute(float comp, int var_1,int var_2,int var_3,float var_4,int var_5,float var_6,float var_7,float var_8,float var_9,float var_10,float var_11,float var_12,float var_13,float var_14,float var_15,float var_16,float var_17,float var_18,float var_19,float var_20,float var_21,float var_22) { for (int i=0; i < var_1; ++i) { for (int i=0; i < var_2; ++i) { for (int i=0; i < var_3; ++i) { if (comp == asinf(+1.6052E-29f * (var_4 - (+1.3418E-44f * -1.1849E-36f)))) { float tmp_1 = +1.4940E-23f; float tmp_2 = -1.6240E35f + -1.2672E-41f + +1.0623E34f + atan2f(logf(var_6 * -0.0f), (-1.4919E-42f * atan2f((-1.9619E-37f * (-1.6127E-37f + acosf(var_7 * -1.0424E-41f / -1.8312E-36f * var_8))), var_9 * (-1.5333E-43f * var_10 / powf(var_11 - coshf((var_12 - coshf(var_13 / -1.9055E20f))), -1.0315E9f))))); comp = tmp_2 * tmp_1 - +1.6798E36f * (+1.1643E34f - (var_14 * +1.4241E36f * (var_15 + var_16))); comp += var_17 / floorf(+0.0f); for (int i=0; i < var_5; ++i) { comp += +1.6689E-43f - asinf(-1.4086E35f / +0.0f * var_18 + var_19); } if (comp == (+1.3292E34f - var_20 * (-1.9025E-42f - var_21))) { comp = +1.9010E19f * var_22 * -1.4243E-5f * tanhf((-1.4812E-35f * +1.5328E36f)); } } } } } printf("%.17g\n", comp); } float* initPointer(float v) { float *ret = (float*) malloc(sizeof(float)*10); for(int i=0; i < 10; ++i) ret[i] = v; return ret; } int main(int argc, char** argv) { /* Program variables */ float tmp_1 = atof(argv[1]); int tmp_2 = atoi(argv[2]); int tmp_3 = atoi(argv[3]); int tmp_4 = atoi(argv[4]); float tmp_5 = atof(argv[5]); int tmp_6 = atoi(argv[6]); float tmp_7 = atof(argv[7]); float tmp_8 = atof(argv[8]); float tmp_9 = atof(argv[9]); float tmp_10 = atof(argv[10]); float tmp_11 = atof(argv[11]); float tmp_12 = atof(argv[12]); float tmp_13 = atof(argv[13]); float tmp_14 = atof(argv[14]); float tmp_15 = atof(argv[15]); float tmp_16 = atof(argv[16]); float tmp_17 = atof(argv[17]); float tmp_18 = atof(argv[18]); float tmp_19 = atof(argv[19]); float tmp_20 = atof(argv[20]); float tmp_21 = atof(argv[21]); float tmp_22 = atof(argv[22]); float tmp_23 = atof(argv[23]); compute<<<1,1>>>(tmp_1,tmp_2,tmp_3,tmp_4,tmp_5,tmp_6,tmp_7,tmp_8,tmp_9,tmp_10,tmp_11,tmp_12,tmp_13,tmp_14,tmp_15,tmp_16,tmp_17,tmp_18,tmp_19,tmp_20,tmp_21,tmp_22,tmp_23); hipDeviceSynchronize(); return 0; }
.text .file "test.hip" .globl _Z22__device_stub__computefiiififffffffffffffffff # -- Begin function _Z22__device_stub__computefiiififffffffffffffffff .p2align 4, 0x90 .type _Z22__device_stub__computefiiififffffffffffffffff,@function _Z22__device_stub__computefiiififffffffffffffffff: # @_Z22__device_stub__computefiiififffffffffffffffff .cfi_startproc # %bb.0: subq $280, %rsp # imm = 0x118 .cfi_def_cfa_offset 288 movss %xmm0, 44(%rsp) movl %edi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movss %xmm1, 28(%rsp) movl %ecx, 24(%rsp) movss %xmm2, 20(%rsp) movss %xmm3, 16(%rsp) movss %xmm4, 12(%rsp) movss %xmm5, 8(%rsp) movss %xmm6, 4(%rsp) movss %xmm7, (%rsp) leaq 44(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rax movq %rax, 104(%rsp) leaq 36(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 28(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 12(%rsp), %rax movq %rax, 160(%rsp) leaq 8(%rsp), %rax movq %rax, 168(%rsp) leaq 4(%rsp), %rax movq %rax, 176(%rsp) movq %rsp, %rax movq %rax, 184(%rsp) leaq 288(%rsp), %rax movq %rax, 192(%rsp) leaq 296(%rsp), %rax movq %rax, 200(%rsp) leaq 304(%rsp), %rax movq %rax, 208(%rsp) leaq 312(%rsp), %rax movq %rax, 216(%rsp) leaq 320(%rsp), %rax movq %rax, 224(%rsp) leaq 328(%rsp), %rax movq %rax, 232(%rsp) leaq 336(%rsp), %rax movq %rax, 240(%rsp) leaq 344(%rsp), %rax movq %rax, 248(%rsp) leaq 352(%rsp), %rax movq %rax, 256(%rsp) leaq 360(%rsp), %rax movq %rax, 264(%rsp) leaq 368(%rsp), %rax movq %rax, 272(%rsp) leaq 80(%rsp), %rdi leaq 64(%rsp), %rsi leaq 56(%rsp), %rdx leaq 48(%rsp), %rcx callq __hipPopCallConfiguration movq 80(%rsp), %rsi movl 88(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z7computefiiififffffffffffffffff, %edi pushq 48(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $296, %rsp # imm = 0x128 .cfi_adjust_cfa_offset -296 retq .Lfunc_end0: .size _Z22__device_stub__computefiiififffffffffffffffff, .Lfunc_end0-_Z22__device_stub__computefiiififffffffffffffffff .cfi_endproc # -- End function .globl _Z11initPointerf # -- Begin function _Z11initPointerf .p2align 4, 0x90 .type _Z11initPointerf,@function _Z11initPointerf: # @_Z11initPointerf .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movss %xmm0, 4(%rsp) # 4-byte Spill movl $40, %edi callq malloc movss 4(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero xorl %ecx, %ecx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movss %xmm0, (%rax,%rcx,4) incq %rcx cmpq $10, %rcx jne .LBB1_1 # %bb.2: popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z11initPointerf, .Lfunc_end1-_Z11initPointerf .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $240, %rsp .cfi_def_cfa_offset 288 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rsi, %r13 movq 8(%rsi), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 232(%rsp) # 8-byte Spill movq 16(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %rbx movq 24(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r14 movq 32(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r15 movq 40(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 224(%rsp) # 8-byte Spill movq 48(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r12 movq 56(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 216(%rsp) # 8-byte Spill movq 64(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 208(%rsp) # 8-byte Spill movq 72(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 200(%rsp) # 8-byte Spill movq 80(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 104(%rsp) # 8-byte Spill movq 88(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 96(%rsp) # 8-byte Spill movq 96(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 88(%rsp) # 8-byte Spill movq 104(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 192(%rsp) # 8-byte Spill movq 112(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 184(%rsp) # 8-byte Spill movq 120(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 176(%rsp) # 8-byte Spill movq 128(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 168(%rsp) # 8-byte Spill movq 136(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 160(%rsp) # 8-byte Spill movq 144(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 152(%rsp) # 8-byte Spill movq 152(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 144(%rsp) # 8-byte Spill movq 160(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 136(%rsp) # 8-byte Spill movq 168(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 128(%rsp) # 8-byte Spill movq 176(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 120(%rsp) # 8-byte Spill movq 184(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 112(%rsp) # 8-byte Spill movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movsd 112(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm8 movsd 120(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm9 movsd 128(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm10 movsd 136(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm11 movsd 144(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm12 movsd 152(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm13 movsd 160(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm14 movsd 168(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm15 movsd 176(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm2 movsd 184(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm3 movsd 192(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm4 movsd 88(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 88(%rsp) # 4-byte Spill movsd 96(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 96(%rsp) # 4-byte Spill movsd 104(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 104(%rsp) # 4-byte Spill movsd 200(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm7 movsd 208(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm6 movsd 216(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm5 movsd 224(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm1 movsd 232(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm8, 80(%rsp) movss %xmm9, 72(%rsp) movss %xmm10, 64(%rsp) movss %xmm11, 56(%rsp) movss %xmm12, 48(%rsp) movss %xmm13, 40(%rsp) movss %xmm14, 32(%rsp) movss %xmm15, 24(%rsp) movss %xmm2, 16(%rsp) movss %xmm3, 8(%rsp) movss %xmm4, (%rsp) movl %ebx, %edi movl %r14d, %esi movl %r15d, %edx movl %r12d, %ecx movaps %xmm5, %xmm2 movaps %xmm6, %xmm3 movaps %xmm7, %xmm4 movss 104(%rsp), %xmm5 # 4-byte Reload # xmm5 = mem[0],zero,zero,zero movss 96(%rsp), %xmm6 # 4-byte Reload # xmm6 = mem[0],zero,zero,zero movss 88(%rsp), %xmm7 # 4-byte Reload # xmm7 = mem[0],zero,zero,zero callq _Z22__device_stub__computefiiififffffffffffffffff .LBB2_2: callq hipDeviceSynchronize xorl %eax, %eax addq $240, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7computefiiififffffffffffffffff, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z7computefiiififffffffffffffffff,@object # @_Z7computefiiififffffffffffffffff .section .rodata,"a",@progbits .globl _Z7computefiiififffffffffffffffff .p2align 3, 0x0 _Z7computefiiififffffffffffffffff: .quad _Z22__device_stub__computefiiififffffffffffffffff .size _Z7computefiiififffffffffffffffff, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z7computefiiififffffffffffffffff" .size .L__unnamed_1, 34 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__computefiiififffffffffffffffff .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7computefiiififffffffffffffffff .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000990a0_00000000-6_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z11initPointerf .type _Z11initPointerf, @function _Z11initPointerf: .LFB2057: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movd %xmm0, %ebx movl $40, %edi call malloc@PLT movq %rax, %rdx leaq 40(%rax), %rcx .L4: movl %ebx, (%rdx) addq $4, %rdx cmpq %rcx, %rdx jne .L4 popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2057: .size _Z11initPointerf, .-_Z11initPointerf .globl _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff .type _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff, @function _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff: .LFB2083: .cfi_startproc endbr64 subq $312, %rsp .cfi_def_cfa_offset 320 movss %xmm0, 44(%rsp) movl %edi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movss %xmm1, 28(%rsp) movl %ecx, 24(%rsp) movss %xmm2, 20(%rsp) movss %xmm3, 16(%rsp) movss %xmm4, 12(%rsp) movss %xmm5, 8(%rsp) movss %xmm6, 4(%rsp) movss %xmm7, (%rsp) movq %fs:40, %rax movq %rax, 296(%rsp) xorl %eax, %eax leaq 44(%rsp), %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rax movq %rax, 120(%rsp) leaq 36(%rsp), %rax movq %rax, 128(%rsp) leaq 32(%rsp), %rax movq %rax, 136(%rsp) leaq 28(%rsp), %rax movq %rax, 144(%rsp) leaq 24(%rsp), %rax movq %rax, 152(%rsp) leaq 20(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 12(%rsp), %rax movq %rax, 176(%rsp) leaq 8(%rsp), %rax movq %rax, 184(%rsp) leaq 4(%rsp), %rax movq %rax, 192(%rsp) movq %rsp, %rax movq %rax, 200(%rsp) leaq 320(%rsp), %rax movq %rax, 208(%rsp) leaq 328(%rsp), %rax movq %rax, 216(%rsp) leaq 336(%rsp), %rax movq %rax, 224(%rsp) leaq 344(%rsp), %rax movq %rax, 232(%rsp) leaq 352(%rsp), %rax movq %rax, 240(%rsp) leaq 360(%rsp), %rax movq %rax, 248(%rsp) leaq 368(%rsp), %rax movq %rax, 256(%rsp) leaq 376(%rsp), %rax movq %rax, 264(%rsp) leaq 384(%rsp), %rax movq %rax, 272(%rsp) leaq 392(%rsp), %rax movq %rax, 280(%rsp) leaq 400(%rsp), %rax movq %rax, 288(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 296(%rsp), %rax subq %fs:40, %rax jne .L12 addq $312, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 328 pushq 56(%rsp) .cfi_def_cfa_offset 336 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z7computefiiififffffffffffffffff(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 320 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff, .-_Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff .globl _Z7computefiiififffffffffffffffff .type _Z7computefiiififffffffffffffffff, @function _Z7computefiiififffffffffffffffff: .LFB2084: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movss 192(%rsp), %xmm8 movss %xmm8, 80(%rsp) movss 184(%rsp), %xmm8 movss %xmm8, 72(%rsp) movss 176(%rsp), %xmm8 movss %xmm8, 64(%rsp) movss 168(%rsp), %xmm8 movss %xmm8, 56(%rsp) movss 160(%rsp), %xmm8 movss %xmm8, 48(%rsp) movss 152(%rsp), %xmm8 movss %xmm8, 40(%rsp) movss 144(%rsp), %xmm8 movss %xmm8, 32(%rsp) movss 136(%rsp), %xmm8 movss %xmm8, 24(%rsp) movss 128(%rsp), %xmm8 movss %xmm8, 16(%rsp) movss 120(%rsp), %xmm8 movss %xmm8, 8(%rsp) movss 112(%rsp), %xmm8 movss %xmm8, (%rsp) call _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff addq $104, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7computefiiififffffffffffffffff, .-_Z7computefiiififffffffffffffffff .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r14 .cfi_def_cfa_offset 16 .cfi_offset 14, -16 pushq %r13 .cfi_def_cfa_offset 24 .cfi_offset 13, -24 pushq %r12 .cfi_def_cfa_offset 32 .cfi_offset 12, -32 pushq %rbp .cfi_def_cfa_offset 40 .cfi_offset 6, -40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset 3, -48 subq $192, %rsp .cfi_def_cfa_offset 240 movq %rsi, %rbx movq 8(%rsi), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 152(%rsp) movq 16(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %rbp movq 24(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r12 movq 32(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r13 movq 40(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 144(%rsp) movq 48(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r14 movq 56(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 136(%rsp) movq 64(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 128(%rsp) movq 72(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 120(%rsp) movq 80(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 112(%rsp) movq 88(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 104(%rsp) movq 96(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 96(%rsp) movq 104(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 88(%rsp) movq 112(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 80(%rsp) movq 120(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 72(%rsp) movq 128(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 64(%rsp) movq 136(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 56(%rsp) movq 144(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 48(%rsp) movq 152(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 40(%rsp) movq 160(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 32(%rsp) movq 168(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 24(%rsp) movq 176(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 16(%rsp) movq 184(%rbx), %rdi movl $0, %esi call strtod@PLT movsd %xmm0, 8(%rsp) movl $1, 180(%rsp) movl $1, 184(%rsp) movl $1, 168(%rsp) movl $1, 172(%rsp) movl $0, %r9d movl $0, %r8d movq 180(%rsp), %rdx movl $1, %ecx movq 168(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L18 .L16: call cudaDeviceSynchronize@PLT movl $0, %eax addq $192, %rsp .cfi_remember_state .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .L18: .cfi_restore_state pxor %xmm0, %xmm0 cvtsd2ss 152(%rsp), %xmm0 subq $96, %rsp .cfi_def_cfa_offset 336 pxor %xmm1, %xmm1 cvtsd2ss 104(%rsp), %xmm1 movss %xmm1, 80(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 112(%rsp), %xmm1 movss %xmm1, 72(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 120(%rsp), %xmm1 movss %xmm1, 64(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 128(%rsp), %xmm1 movss %xmm1, 56(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 136(%rsp), %xmm1 movss %xmm1, 48(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 144(%rsp), %xmm1 movss %xmm1, 40(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 152(%rsp), %xmm1 movss %xmm1, 32(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 160(%rsp), %xmm1 movss %xmm1, 24(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 168(%rsp), %xmm1 movss %xmm1, 16(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 176(%rsp), %xmm1 movss %xmm1, 8(%rsp) pxor %xmm1, %xmm1 cvtsd2ss 184(%rsp), %xmm1 movss %xmm1, (%rsp) pxor %xmm7, %xmm7 cvtsd2ss 192(%rsp), %xmm7 pxor %xmm6, %xmm6 cvtsd2ss 200(%rsp), %xmm6 pxor %xmm5, %xmm5 cvtsd2ss 208(%rsp), %xmm5 pxor %xmm4, %xmm4 cvtsd2ss 216(%rsp), %xmm4 pxor %xmm3, %xmm3 cvtsd2ss 224(%rsp), %xmm3 pxor %xmm2, %xmm2 cvtsd2ss 232(%rsp), %xmm2 movl %r14d, %ecx pxor %xmm1, %xmm1 cvtsd2ss 240(%rsp), %xmm1 movl %r13d, %edx movl %r12d, %esi movl %ebp, %edi call _Z47__device_stub__Z7computefiiififfffffffffffffffffiiififffffffffffffffff addq $96, %rsp .cfi_def_cfa_offset 240 jmp .L16 .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z7computefiiififffffffffffffffff" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z7computefiiififffffffffffffffff(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "test.hip" .globl _Z22__device_stub__computefiiififffffffffffffffff # -- Begin function _Z22__device_stub__computefiiififffffffffffffffff .p2align 4, 0x90 .type _Z22__device_stub__computefiiififffffffffffffffff,@function _Z22__device_stub__computefiiififffffffffffffffff: # @_Z22__device_stub__computefiiififffffffffffffffff .cfi_startproc # %bb.0: subq $280, %rsp # imm = 0x118 .cfi_def_cfa_offset 288 movss %xmm0, 44(%rsp) movl %edi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movss %xmm1, 28(%rsp) movl %ecx, 24(%rsp) movss %xmm2, 20(%rsp) movss %xmm3, 16(%rsp) movss %xmm4, 12(%rsp) movss %xmm5, 8(%rsp) movss %xmm6, 4(%rsp) movss %xmm7, (%rsp) leaq 44(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rax movq %rax, 104(%rsp) leaq 36(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 28(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 12(%rsp), %rax movq %rax, 160(%rsp) leaq 8(%rsp), %rax movq %rax, 168(%rsp) leaq 4(%rsp), %rax movq %rax, 176(%rsp) movq %rsp, %rax movq %rax, 184(%rsp) leaq 288(%rsp), %rax movq %rax, 192(%rsp) leaq 296(%rsp), %rax movq %rax, 200(%rsp) leaq 304(%rsp), %rax movq %rax, 208(%rsp) leaq 312(%rsp), %rax movq %rax, 216(%rsp) leaq 320(%rsp), %rax movq %rax, 224(%rsp) leaq 328(%rsp), %rax movq %rax, 232(%rsp) leaq 336(%rsp), %rax movq %rax, 240(%rsp) leaq 344(%rsp), %rax movq %rax, 248(%rsp) leaq 352(%rsp), %rax movq %rax, 256(%rsp) leaq 360(%rsp), %rax movq %rax, 264(%rsp) leaq 368(%rsp), %rax movq %rax, 272(%rsp) leaq 80(%rsp), %rdi leaq 64(%rsp), %rsi leaq 56(%rsp), %rdx leaq 48(%rsp), %rcx callq __hipPopCallConfiguration movq 80(%rsp), %rsi movl 88(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z7computefiiififffffffffffffffff, %edi pushq 48(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $296, %rsp # imm = 0x128 .cfi_adjust_cfa_offset -296 retq .Lfunc_end0: .size _Z22__device_stub__computefiiififffffffffffffffff, .Lfunc_end0-_Z22__device_stub__computefiiififffffffffffffffff .cfi_endproc # -- End function .globl _Z11initPointerf # -- Begin function _Z11initPointerf .p2align 4, 0x90 .type _Z11initPointerf,@function _Z11initPointerf: # @_Z11initPointerf .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movss %xmm0, 4(%rsp) # 4-byte Spill movl $40, %edi callq malloc movss 4(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero xorl %ecx, %ecx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movss %xmm0, (%rax,%rcx,4) incq %rcx cmpq $10, %rcx jne .LBB1_1 # %bb.2: popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z11initPointerf, .Lfunc_end1-_Z11initPointerf .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $240, %rsp .cfi_def_cfa_offset 288 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rsi, %r13 movq 8(%rsi), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 232(%rsp) # 8-byte Spill movq 16(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %rbx movq 24(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r14 movq 32(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r15 movq 40(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 224(%rsp) # 8-byte Spill movq 48(%r13), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r12 movq 56(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 216(%rsp) # 8-byte Spill movq 64(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 208(%rsp) # 8-byte Spill movq 72(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 200(%rsp) # 8-byte Spill movq 80(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 104(%rsp) # 8-byte Spill movq 88(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 96(%rsp) # 8-byte Spill movq 96(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 88(%rsp) # 8-byte Spill movq 104(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 192(%rsp) # 8-byte Spill movq 112(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 184(%rsp) # 8-byte Spill movq 120(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 176(%rsp) # 8-byte Spill movq 128(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 168(%rsp) # 8-byte Spill movq 136(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 160(%rsp) # 8-byte Spill movq 144(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 152(%rsp) # 8-byte Spill movq 152(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 144(%rsp) # 8-byte Spill movq 160(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 136(%rsp) # 8-byte Spill movq 168(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 128(%rsp) # 8-byte Spill movq 176(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 120(%rsp) # 8-byte Spill movq 184(%r13), %rdi xorl %esi, %esi callq strtod movsd %xmm0, 112(%rsp) # 8-byte Spill movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movsd 112(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm8 movsd 120(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm9 movsd 128(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm10 movsd 136(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm11 movsd 144(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm12 movsd 152(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm13 movsd 160(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm14 movsd 168(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm15 movsd 176(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm2 movsd 184(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm3 movsd 192(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm4 movsd 88(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 88(%rsp) # 4-byte Spill movsd 96(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 96(%rsp) # 4-byte Spill movsd 104(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm0, 104(%rsp) # 4-byte Spill movsd 200(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm7 movsd 208(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm6 movsd 216(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm5 movsd 224(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm1 movsd 232(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero cvtsd2ss %xmm0, %xmm0 movss %xmm8, 80(%rsp) movss %xmm9, 72(%rsp) movss %xmm10, 64(%rsp) movss %xmm11, 56(%rsp) movss %xmm12, 48(%rsp) movss %xmm13, 40(%rsp) movss %xmm14, 32(%rsp) movss %xmm15, 24(%rsp) movss %xmm2, 16(%rsp) movss %xmm3, 8(%rsp) movss %xmm4, (%rsp) movl %ebx, %edi movl %r14d, %esi movl %r15d, %edx movl %r12d, %ecx movaps %xmm5, %xmm2 movaps %xmm6, %xmm3 movaps %xmm7, %xmm4 movss 104(%rsp), %xmm5 # 4-byte Reload # xmm5 = mem[0],zero,zero,zero movss 96(%rsp), %xmm6 # 4-byte Reload # xmm6 = mem[0],zero,zero,zero movss 88(%rsp), %xmm7 # 4-byte Reload # xmm7 = mem[0],zero,zero,zero callq _Z22__device_stub__computefiiififffffffffffffffff .LBB2_2: callq hipDeviceSynchronize xorl %eax, %eax addq $240, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7computefiiififffffffffffffffff, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end3: .size __hip_module_ctor, .Lfunc_end3-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB4_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB4_2: retq .Lfunc_end4: .size __hip_module_dtor, .Lfunc_end4-__hip_module_dtor .cfi_endproc # -- End function .type _Z7computefiiififffffffffffffffff,@object # @_Z7computefiiififffffffffffffffff .section .rodata,"a",@progbits .globl _Z7computefiiififffffffffffffffff .p2align 3, 0x0 _Z7computefiiififffffffffffffffff: .quad _Z22__device_stub__computefiiififffffffffffffffff .size _Z7computefiiififffffffffffffffff, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z7computefiiififffffffffffffffff" .size .L__unnamed_1, 34 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__computefiiififffffffffffffffff .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z7computefiiififffffffffffffffff .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
//headers #include <stdio.h> #include <cuda.h> #define imin(a, b) ((a < b) ? a : b) #define sum_squares(x) (x * (x + 1) * (2 * x + 1) / 6) //global variables float *hostA = NULL; float *hostB = NULL; float *partial_hostC = NULL; float *deviceA = NULL; float *deviceB = NULL; float *partial_deviceC = NULL; const int iNumberOfArrayElements = 33 * 1024; const int threadsPerBlock = 256; const int blocksPerGrid = imin(32, (iNumberOfArrayElements + threadsPerBlock - 1) / threadsPerBlock); // *** CUDA KERNEL DEFINITION *** __global__ void vecDotProduct(float *input1, float *input2, float *output) { //variable declaration //shared across all threads within block __shared__ float cache[threadsPerBlock]; int tid = blockIdx.x * blockDim.x + threadIdx.x; int cacheIndex = threadIdx.x; float temp = 0; //code while(tid < iNumberOfArrayElements) { temp += input1[tid] * input2[tid]; tid += blockDim.x * gridDim.x; } //set the cache values cache[cacheIndex] = temp; //synchronize threads in the block __syncthreads(); //summation reduction int i = blockDim.x / 2; while(i != 0) { if(cacheIndex < i) cache[cacheIndex] += cache[cacheIndex + i]; __syncthreads(); i /= 2; } //copy to output memory if(cacheIndex == 0) output[blockIdx.x] = cache[0]; } int main(void) { //function declaration void cleanup(void); //code //allocate memory on host hostA = (float *)malloc(iNumberOfArrayElements * sizeof(float)); if(hostA == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 1.\nExiting Now ...\n"); exit(EXIT_FAILURE); } hostB = (float *)malloc(iNumberOfArrayElements * sizeof(float)); if(hostB == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 2.\nExiting Now ...\n"); cleanup(); exit(EXIT_FAILURE); } partial_hostC = (float *)malloc(blocksPerGrid * sizeof(float)); if(partial_hostC == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Output Array.\nExiting Now ...\n"); cleanup(); exit(EXIT_FAILURE); } //allocate memory on device cudaError_t err = cudaSuccess; err = cudaMalloc((void **)&deviceA, iNumberOfArrayElements * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMalloc((void **)&deviceB, iNumberOfArrayElements * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMalloc((void **)&partial_deviceC, blocksPerGrid * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //fill the host input array for(int i = 0; i < iNumberOfArrayElements; i++) { hostA[i] = i; hostB[i] = i * 2; } //copy the host input array to device input err = cudaMemcpy(deviceA, hostA, iNumberOfArrayElements * sizeof(float), cudaMemcpyHostToDevice); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMemcpy(deviceB, hostB, iNumberOfArrayElements * sizeof(float), cudaMemcpyHostToDevice); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //launch the kernel vecDotProduct<<<blocksPerGrid, threadsPerBlock>>>(deviceA, deviceB, partial_deviceC); //copy the device output array back to host err = cudaMemcpy(partial_hostC, partial_deviceC, blocksPerGrid * sizeof(float), cudaMemcpyDeviceToHost); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //calculate final result on host float finalC = 0.0f; for(int i = 0; i < blocksPerGrid; i++) { finalC += partial_hostC[i]; } //check if the final value is correct if(finalC == (2 * sum_squares((float)(iNumberOfArrayElements - 1)))) printf("Dot Product Calculated On Device Is Accurate.\n"); else printf("Dot Product Calculated On Device Is Not Accurate.\n"); printf("Dot Product = %0.6.\n", finalC); printf("Expected Product = %0.6f.\n", 2 * sum_squares((float)(iNumberOfArrayElements - 1))); //total cleanup cleanup(); return (0); } void cleanup(void) { //code //free device memory if(partial_deviceC) { cudaFree(partial_deviceC); partial_deviceC = NULL; } if(deviceB) { cudaFree(deviceB); deviceB = NULL; } if(deviceA) { cudaFree(deviceA); deviceA = NULL; } //free host memory if(partial_hostC) { free(partial_hostC); partial_hostC = NULL; } if(hostB) { free(hostB); hostB = NULL; } if(hostA) { free(hostA); hostA = NULL; } }
code for sm_80 Function : _Z13vecDotProductPfS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R7, SR_CTAID.X ; /* 0x0000000000077919 */ /* 0x000e220000002500 */ /*0020*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe20000000a00 */ /*0030*/ BSSY B0, 0x150 ; /* 0x0000011000007945 */ /* 0x000fe20003800000 */ /*0040*/ HFMA2.MMA R6, -RZ, RZ, 0, 0 ; /* 0x00000000ff067435 */ /* 0x000fe200000001ff */ /*0050*/ S2R R8, SR_TID.X ; /* 0x0000000000087919 */ /* 0x000e240000002100 */ /*0060*/ IMAD R0, R7, c[0x0][0x0], R8 ; /* 0x0000000007007a24 */ /* 0x001fca00078e0208 */ /*0070*/ ISETP.GT.AND P0, PT, R0, 0x83ff, PT ; /* 0x000083ff0000780c */ /* 0x000fda0003f04270 */ /*0080*/ @P0 BRA 0x140 ; /* 0x000000b000000947 */ /* 0x000fea0003800000 */ /*0090*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fe400078e00ff */ /*00a0*/ MOV R5, 0x4 ; /* 0x0000000400057802 */ /* 0x000fca0000000f00 */ /*00b0*/ IMAD.WIDE R2, R0, R5, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x000fc800078e0205 */ /*00c0*/ IMAD.WIDE R4, R0, R5, c[0x0][0x168] ; /* 0x00005a0000047625 */ /* 0x000fe400078e0205 */ /*00d0*/ LDG.E R2, [R2.64] ; /* 0x0000000602027981 */ /* 0x000ea8000c1e1900 */ /*00e0*/ LDG.E R5, [R4.64] ; /* 0x0000000604057981 */ /* 0x000ea2000c1e1900 */ /*00f0*/ IMAD.MOV.U32 R9, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff097624 */ /* 0x000fc800078e00ff */ /*0100*/ IMAD R0, R9, c[0x0][0xc], R0 ; /* 0x0000030009007a24 */ /* 0x000fca00078e0200 */ /*0110*/ ISETP.GE.AND P0, PT, R0, 0x8400, PT ; /* 0x000084000000780c */ /* 0x000fe20003f06270 */ /*0120*/ FFMA R6, R5, R2, R6 ; /* 0x0000000205067223 */ /* 0x004fd80000000006 */ /*0130*/ @!P0 BRA 0xa0 ; /* 0xffffff6000008947 */ /* 0x000fea000383ffff */ /*0140*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0150*/ ULDC UR4, c[0x0][0x0] ; /* 0x0000000000047ab9 */ /* 0x000fe20000000800 */ /*0160*/ STS [R8.X4], R6 ; /* 0x0000000608007388 */ /* 0x0001e20000004800 */ /*0170*/ USHF.R.U32.HI UR4, URZ, 0x1, UR4 ; /* 0x000000013f047899 */ /* 0x000fc60008011604 */ /*0180*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0190*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe40003f05270 */ /*01a0*/ ISETP.NE.AND P1, PT, RZ, UR4, PT ; /* 0x00000004ff007c0c */ /* 0x000fda000bf25270 */ /*01b0*/ @!P1 BRA 0x2a0 ; /* 0x000000e000009947 */ /* 0x000fea0003800000 */ /*01c0*/ SHF.L.U32 R0, R8, 0x2, RZ ; /* 0x0000000208007819 */ /* 0x001fe200000006ff */ /*01d0*/ IMAD.U32 R3, RZ, RZ, UR4 ; /* 0x00000004ff037e24 */ /* 0x000fca000f8e00ff */ /*01e0*/ ISETP.GE.AND P1, PT, R8, R3, PT ; /* 0x000000030800720c */ /* 0x000fda0003f26270 */ /*01f0*/ @!P1 LEA R2, R3.reuse, R0, 0x2 ; /* 0x0000000003029211 */ /* 0x040fe200078e10ff */ /*0200*/ @!P1 LDS R4, [R8.X4] ; /* 0x0000000008049984 */ /* 0x000fe80000004800 */ /*0210*/ @!P1 LDS R5, [R2] ; /* 0x0000000002059984 */ /* 0x000e240000000800 */ /*0220*/ @!P1 FADD R4, R4, R5 ; /* 0x0000000504049221 */ /* 0x001fe20000000000 */ /*0230*/ IADD3 R5, R3.reuse, 0x1, RZ ; /* 0x0000000103057810 */ /* 0x040fe40007ffe0ff */ /*0240*/ LEA.HI R3, R3, R3, RZ, 0x1 ; /* 0x0000000303037211 */ /* 0x000fc400078f08ff */ /*0250*/ @!P1 STS [R8.X4], R4 ; /* 0x0000000408009388 */ /* 0x0001e80000004800 */ /*0260*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0270*/ ISETP.GT.U32.AND P1, PT, R5, 0x2, PT ; /* 0x000000020500780c */ /* 0x000fe40003f24070 */ /*0280*/ SHF.R.S32.HI R3, RZ, 0x1, R3 ; /* 0x00000001ff037819 */ /* 0x000fd60000011403 */ /*0290*/ @P1 BRA 0x1e0 ; /* 0xffffff4000001947 */ /* 0x001fea000383ffff */ /*02a0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x001fea0003800000 */ /*02b0*/ LDS R5, [RZ] ; /* 0x00000000ff057984 */ /* 0x000e220000000800 */ /*02c0*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fc800078e00ff */ /*02d0*/ IMAD.WIDE.U32 R2, R7, R2, c[0x0][0x170] ; /* 0x00005c0007027625 */ /* 0x000fca00078e0002 */ /*02e0*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x001fe2000c101906 */ /*02f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0300*/ BRA 0x300; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0310*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0320*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0330*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0340*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0350*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0360*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0370*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0380*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0390*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
//headers #include <stdio.h> #include <cuda.h> #define imin(a, b) ((a < b) ? a : b) #define sum_squares(x) (x * (x + 1) * (2 * x + 1) / 6) //global variables float *hostA = NULL; float *hostB = NULL; float *partial_hostC = NULL; float *deviceA = NULL; float *deviceB = NULL; float *partial_deviceC = NULL; const int iNumberOfArrayElements = 33 * 1024; const int threadsPerBlock = 256; const int blocksPerGrid = imin(32, (iNumberOfArrayElements + threadsPerBlock - 1) / threadsPerBlock); // *** CUDA KERNEL DEFINITION *** __global__ void vecDotProduct(float *input1, float *input2, float *output) { //variable declaration //shared across all threads within block __shared__ float cache[threadsPerBlock]; int tid = blockIdx.x * blockDim.x + threadIdx.x; int cacheIndex = threadIdx.x; float temp = 0; //code while(tid < iNumberOfArrayElements) { temp += input1[tid] * input2[tid]; tid += blockDim.x * gridDim.x; } //set the cache values cache[cacheIndex] = temp; //synchronize threads in the block __syncthreads(); //summation reduction int i = blockDim.x / 2; while(i != 0) { if(cacheIndex < i) cache[cacheIndex] += cache[cacheIndex + i]; __syncthreads(); i /= 2; } //copy to output memory if(cacheIndex == 0) output[blockIdx.x] = cache[0]; } int main(void) { //function declaration void cleanup(void); //code //allocate memory on host hostA = (float *)malloc(iNumberOfArrayElements * sizeof(float)); if(hostA == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 1.\nExiting Now ...\n"); exit(EXIT_FAILURE); } hostB = (float *)malloc(iNumberOfArrayElements * sizeof(float)); if(hostB == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 2.\nExiting Now ...\n"); cleanup(); exit(EXIT_FAILURE); } partial_hostC = (float *)malloc(blocksPerGrid * sizeof(float)); if(partial_hostC == NULL) { printf("CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Output Array.\nExiting Now ...\n"); cleanup(); exit(EXIT_FAILURE); } //allocate memory on device cudaError_t err = cudaSuccess; err = cudaMalloc((void **)&deviceA, iNumberOfArrayElements * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMalloc((void **)&deviceB, iNumberOfArrayElements * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMalloc((void **)&partial_deviceC, blocksPerGrid * sizeof(float)); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //fill the host input array for(int i = 0; i < iNumberOfArrayElements; i++) { hostA[i] = i; hostB[i] = i * 2; } //copy the host input array to device input err = cudaMemcpy(deviceA, hostA, iNumberOfArrayElements * sizeof(float), cudaMemcpyHostToDevice); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } err = cudaMemcpy(deviceB, hostB, iNumberOfArrayElements * sizeof(float), cudaMemcpyHostToDevice); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //launch the kernel vecDotProduct<<<blocksPerGrid, threadsPerBlock>>>(deviceA, deviceB, partial_deviceC); //copy the device output array back to host err = cudaMemcpy(partial_hostC, partial_deviceC, blocksPerGrid * sizeof(float), cudaMemcpyDeviceToHost); if(err != cudaSuccess) { printf("GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n", cudaGetErrorString(err), __FILE__, __LINE__); cleanup(); exit(EXIT_FAILURE); } //calculate final result on host float finalC = 0.0f; for(int i = 0; i < blocksPerGrid; i++) { finalC += partial_hostC[i]; } //check if the final value is correct if(finalC == (2 * sum_squares((float)(iNumberOfArrayElements - 1)))) printf("Dot Product Calculated On Device Is Accurate.\n"); else printf("Dot Product Calculated On Device Is Not Accurate.\n"); printf("Dot Product = %0.6.\n", finalC); printf("Expected Product = %0.6f.\n", 2 * sum_squares((float)(iNumberOfArrayElements - 1))); //total cleanup cleanup(); return (0); } void cleanup(void) { //code //free device memory if(partial_deviceC) { cudaFree(partial_deviceC); partial_deviceC = NULL; } if(deviceB) { cudaFree(deviceB); deviceB = NULL; } if(deviceA) { cudaFree(deviceA); deviceA = NULL; } //free host memory if(partial_hostC) { free(partial_hostC); partial_hostC = NULL; } if(hostB) { free(hostB); hostB = NULL; } if(hostA) { free(hostA); hostA = NULL; } }
.file "tmpxft_000ce620_00000000-6_VectorDotProduct.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z7cleanupv .type _Z7cleanupv, @function _Z7cleanupv: .LFB2058: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq partial_deviceC(%rip), %rdi testq %rdi, %rdi je .L4 call cudaFree@PLT movq $0, partial_deviceC(%rip) .L4: movq deviceB(%rip), %rdi testq %rdi, %rdi je .L5 call cudaFree@PLT movq $0, deviceB(%rip) .L5: movq deviceA(%rip), %rdi testq %rdi, %rdi je .L6 call cudaFree@PLT movq $0, deviceA(%rip) .L6: movq partial_hostC(%rip), %rdi testq %rdi, %rdi je .L7 call free@PLT movq $0, partial_hostC(%rip) .L7: movq hostB(%rip), %rdi testq %rdi, %rdi je .L8 call free@PLT movq $0, hostB(%rip) .L8: movq hostA(%rip), %rdi testq %rdi, %rdi je .L3 call free@PLT movq $0, hostA(%rip) .L3: addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z7cleanupv, .-_Z7cleanupv .globl _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_ .type _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_, @function _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_: .LFB2083: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 120(%rsp), %rax subq %fs:40, %rax jne .L16 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z13vecDotProductPfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_, .-_Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_ .globl _Z13vecDotProductPfS_S_ .type _Z13vecDotProductPfS_S_, @function _Z13vecDotProductPfS_S_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z13vecDotProductPfS_S_, .-_Z13vecDotProductPfS_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC1: .string "CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 1.\nExiting Now ...\n" .align 8 .LC2: .string "CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Input Array 2.\nExiting Now ...\n" .align 8 .LC3: .string "CPU Memory Fatal Error - Can Not Allocate Enough Memory For Host Output Array.\nExiting Now ...\n" .align 8 .LC4: .string "/home/ubuntu/Datasets/stackv2/train-structured/yashPat98/HPP/master/CUDA/CUDA-by-Example/12-VectorDotProduct/VectorDotProduct.cu" .align 8 .LC5: .string "GPU Memory Fatal Error - %s In The File Name %s At Line No %d.\nExiting Now ...\n" .align 8 .LC7: .string "Dot Product Calculated On Device Is Accurate.\n" .align 8 .LC8: .string "Dot Product Calculated On Device Is Not Accurate.\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC9: .string "Dot Product = %0.6.\n" .LC11: .string "Expected Product = %0.6f.\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $56, %rsp .cfi_def_cfa_offset 64 movl $135168, %edi call malloc@PLT movq %rax, hostA(%rip) testq %rax, %rax je .L39 movl $135168, %edi call malloc@PLT movq %rax, hostB(%rip) testq %rax, %rax je .L40 movl $128, %edi call malloc@PLT movq %rax, partial_hostC(%rip) testq %rax, %rax je .L41 movl $135168, %esi leaq deviceA(%rip), %rdi call cudaMalloc@PLT testl %eax, %eax jne .L42 movl $135168, %esi leaq deviceB(%rip), %rdi call cudaMalloc@PLT testl %eax, %eax jne .L43 movl $128, %esi leaq partial_deviceC(%rip), %rdi call cudaMalloc@PLT testl %eax, %eax jne .L44 movl $0, %eax .L25: pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 movq hostA(%rip), %rdx movss %xmm0, (%rdx,%rax,4) leal (%rax,%rax), %edx pxor %xmm0, %xmm0 cvtsi2ssl %edx, %xmm0 movq hostB(%rip), %rdx movss %xmm0, (%rdx,%rax,4) addq $1, %rax cmpq $33792, %rax jne .L25 movl $1, %ecx movl $135168, %edx movq hostA(%rip), %rsi movq deviceA(%rip), %rdi call cudaMemcpy@PLT testl %eax, %eax jne .L45 movl $1, %ecx movl $135168, %edx movq hostB(%rip), %rsi movq deviceB(%rip), %rdi call cudaMemcpy@PLT testl %eax, %eax jne .L46 movl $256, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $32, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 36(%rsp), %rdx movl $1, %ecx movq 24(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L47 .L28: movl $2, %ecx movl $128, %edx movq partial_deviceC(%rip), %rsi movq partial_hostC(%rip), %rdi call cudaMemcpy@PLT testl %eax, %eax jne .L48 movq partial_hostC(%rip), %rax leaq 128(%rax), %rdx movl $0x00000000, 12(%rsp) .L30: movss 12(%rsp), %xmm1 addss (%rax), %xmm1 movss %xmm1, 12(%rsp) addq $4, %rax cmpq %rdx, %rax jne .L30 ucomiss .LC6(%rip), %xmm1 jp .L31 jne .L31 leaq .LC7(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L33: pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 leaq .LC9(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movsd .LC10(%rip), %xmm0 leaq .LC11(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT call _Z7cleanupv movl $0, %eax addq $56, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L39: .cfi_restore_state leaq .LC1(%rip), %rsi movl $2, %edi call __printf_chk@PLT movl $1, %edi call exit@PLT .L40: leaq .LC2(%rip), %rsi movl $2, %edi call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L41: leaq .LC3(%rip), %rsi movl $2, %edi call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L42: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $96, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L43: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $104, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L44: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $112, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L45: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $128, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L46: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $136, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L47: movq partial_deviceC(%rip), %rdx movq deviceB(%rip), %rsi movq deviceA(%rip), %rdi call _Z37__device_stub__Z13vecDotProductPfS_S_PfS_S_ jmp .L28 .L48: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $148, %r8d leaq .LC4(%rip), %rcx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT call _Z7cleanupv movl $1, %edi call exit@PLT .L31: leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L33 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC12: .string "_Z13vecDotProductPfS_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _Z13vecDotProductPfS_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .globl partial_deviceC .bss .align 8 .type partial_deviceC, @object .size partial_deviceC, 8 partial_deviceC: .zero 8 .globl deviceB .align 8 .type deviceB, @object .size deviceB, 8 deviceB: .zero 8 .globl deviceA .align 8 .type deviceA, @object .size deviceA, 8 deviceA: .zero 8 .globl partial_hostC .align 8 .type partial_hostC, @object .size partial_hostC, 8 partial_hostC: .zero 8 .globl hostB .align 8 .type hostB, @object .size hostB, 8 hostB: .zero 8 .globl hostA .align 8 .type hostA, @object .size hostA, 8 hostA: .zero 8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC6: .long 1438329312 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC10: .long 0 .long 1119315260 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4: