system
stringclasses
7 values
input
stringlengths
10
94.5k
output
stringlengths
14
55.5k
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000c06c5_00000000-6_find_all_sums_hub_kernel.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 _Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_ .type _Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_, @function _Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 56(%rsp) movl %esi, 52(%rsp) movq %rdx, 40(%rsp) movq %rcx, 32(%rsp) movq %r8, 24(%rsp) movq %r9, 16(%rsp) movq 208(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 52(%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) 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 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 72(%rsp) .cfi_def_cfa_offset 216 pushq 72(%rsp) .cfi_def_cfa_offset 224 leaq 144(%rsp), %r9 movq 108(%rsp), %rcx movl 116(%rsp), %r8d movq 96(%rsp), %rsi movl 104(%rsp), %edx leaq _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_(%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 _Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_, .-_Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_ .globl _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .type _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_, @function _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z57__device_stub__Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_PiiPfS_S_S0_S0_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_, .-_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_" .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 _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_(%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 "find_all_sums_hub_kernel.hip" .globl _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ # -- Begin function _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .p2align 4, 0x90 .type _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_,@function _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_: # @_Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movl %esi, 4(%rsp) movq %rdx, 80(%rsp) movq %rcx, 72(%rsp) movq %r8, 64(%rsp) movq %r9, 56(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 4(%rsp), %rax movq %rax, 104(%rsp) leaq 80(%rsp), %rax movq %rax, 112(%rsp) leaq 72(%rsp), %rax movq %rax, 120(%rsp) leaq 64(%rsp), %rax movq %rax, 128(%rsp) leaq 56(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%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 $_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_, %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_end0: .size _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_, .Lfunc_end0-_Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .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 $_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_, %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 _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_,@object # @_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .section .rodata,"a",@progbits .globl _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .p2align 3, 0x0 _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_: .quad _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .size _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_" .size .L__unnamed_1, 44 .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 _Z39__device_stub__find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z24find_all_sums_hub_kernelPiiPfS_S_S0_S0_ .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.
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
code for sm_80 Function : _Z19markFilterEdges_gpuPiS_S_S_i .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*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0070*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fc800078e00ff */ /*0080*/ IMAD.WIDE R2, R0, R7, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x001fcc00078e0207 */ /*0090*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea4000c1e1900 */ /*00a0*/ IMAD.WIDE R4, R2, R7, c[0x0][0x170] ; /* 0x00005c0002047625 */ /* 0x004fcc00078e0207 */ /*00b0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ SHF.R.S32.HI R11, RZ, 0x1f, R0 ; /* 0x0000001fff0b7819 */ /* 0x000fe40000011400 */ /*00d0*/ ISETP.NE.AND P0, PT, R4, -0x1, PT ; /* 0xffffffff0400780c */ /* 0x004fda0003f05270 */ /*00e0*/ @!P0 LEA R8, P1, R0, c[0x0][0x168], 0x2 ; /* 0x00005a0000088a11 */ /* 0x000fc800078210ff */ /*00f0*/ @!P0 LEA.HI.X R9, R0, c[0x0][0x16c], R11, 0x2, P1 ; /* 0x00005b0000098a11 */ /* 0x000fca00008f140b */ /*0100*/ @!P0 LDG.E R6, [R8.64] ; /* 0x0000000408068981 */ /* 0x000ea4000c1e1900 */ /*0110*/ @!P0 IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006068625 */ /* 0x004fcc00078e0207 */ /*0120*/ @!P0 LDG.E R6, [R6.64] ; /* 0x0000000406068981 */ /* 0x000ea2000c1e1900 */ /*0130*/ LEA R10, P1, R0, c[0x0][0x178], 0x2 ; /* 0x00005e00000a7a11 */ /* 0x000fe200078210ff */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff057624 */ /* 0x000fc600078e00ff */ /*0150*/ LEA.HI.X R11, R0, c[0x0][0x17c], R11, 0x2, P1 ; /* 0x00005f00000b7a11 */ /* 0x000fe200008f140b */ /*0160*/ IMAD R0, R5, c[0x0][0xc], R0 ; /* 0x0000030005007a24 */ /* 0x000fe200078e0200 */ /*0170*/ ISETP.EQ.AND P0, PT, R6, -0x1, !P0 ; /* 0xffffffff0600780c */ /* 0x004fc80004702270 */ /*0180*/ SEL R3, RZ, 0x1, !P0 ; /* 0x00000001ff037807 */ /* 0x000fe40004000000 */ /*0190*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fc60003f06270 */ /*01a0*/ STG.E [R10.64], R3 ; /* 0x000000030a007986 */ /* 0x0001f4000c101904 */ /*01b0*/ @!P0 BRA 0x70 ; /* 0xfffffeb000008947 */ /* 0x000fea000383ffff */ /*01c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*01d0*/ BRA 0x1d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
.file "tmpxft_000d8b08_00000000-6_markFilterEdges.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .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 .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i .type _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i, @function _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i: .LFB2081: .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 .L7 .L3: movq 152(%rsp), %rax subq %fs:40, %rax jne .L8 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .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 _Z19markFilterEdges_gpuPiS_S_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i, .-_Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i .globl _Z19markFilterEdges_gpuPiS_S_S_i .type _Z19markFilterEdges_gpuPiS_S_S_i, @function _Z19markFilterEdges_gpuPiS_S_S_i: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size _Z19markFilterEdges_gpuPiS_S_S_i, .-_Z19markFilterEdges_gpuPiS_S_S_i .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z19markFilterEdges_gpuPiS_S_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .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 _Z19markFilterEdges_gpuPiS_S_S_i(%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 .LFE2084: .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.
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z19markFilterEdges_gpuPiS_S_S_i .globl _Z19markFilterEdges_gpuPiS_S_S_i .p2align 8 .type _Z19markFilterEdges_gpuPiS_S_S_i,@function _Z19markFilterEdges_gpuPiS_S_S_i: s_clause 0x1 s_load_b32 s4, s[0:1], 0x34 s_load_b32 s12, s[0:1], 0x20 s_add_u32 s2, s0, 40 s_addc_u32 s3, s1, 0 s_waitcnt lgkmcnt(0) s_and_b32 s13, s4, 0xffff s_mov_b32 s4, exec_lo v_mad_u64_u32 v[1:2], null, s15, s13, v[0:1] s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s12, v1 s_cbranch_execz .LBB0_7 s_load_b32 s2, s[2:3], 0x0 s_load_b256 s[4:11], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_mul_i32 s1, s2, s13 s_mov_b32 s2, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s3 v_add_nc_u32_e32 v1, s1, v1 v_add_co_u32 v2, s0, s10, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_ci_u32_e64 v3, s0, s11, v3, s0 v_cmp_le_i32_e32 vcc_lo, s12, v1 global_store_b32 v[2:3], v0, off s_or_b32 s2, vcc_lo, s2 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s2 s_cbranch_execz .LBB0_7 .LBB0_3: v_ashrrev_i32_e32 v2, 31, v1 s_mov_b32 s3, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[2:3], 2, v[1:2] v_add_co_u32 v4, vcc_lo, s4, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v5, vcc_lo, s5, v3, vcc_lo global_load_b32 v4, v[4:5], off s_waitcnt vmcnt(0) v_ashrrev_i32_e32 v5, 31, v4 v_lshlrev_b64 v[4:5], 2, v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v4, vcc_lo, s8, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s9, v5, vcc_lo global_load_b32 v0, v[4:5], off s_waitcnt vmcnt(0) v_cmp_ne_u32_e64 s0, -1, v0 v_cmpx_eq_u32_e32 -1, v0 s_cbranch_execz .LBB0_5 v_add_co_u32 v4, vcc_lo, s6, v2 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v3, vcc_lo s_and_not1_b32 s0, s0, exec_lo s_mov_b32 s13, 1 global_load_b32 v4, v[4:5], off s_waitcnt vmcnt(0) 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], 2, v[4:5] v_add_co_u32 v4, vcc_lo, s8, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_4) | instid1(SALU_CYCLE_1) v_add_co_ci_u32_e32 v5, vcc_lo, s9, v5, vcc_lo global_load_b32 v0, v[4:5], off s_waitcnt vmcnt(0) v_cmp_ne_u32_e32 vcc_lo, -1, v0 s_and_b32 s14, vcc_lo, exec_lo s_or_b32 s0, s0, s14 .LBB0_5: s_or_b32 exec_lo, exec_lo, s3 v_mov_b32_e32 v0, s13 s_and_saveexec_b32 s3, s0 s_cbranch_execz .LBB0_2 v_mov_b32_e32 v0, 0 s_branch .LBB0_2 .LBB0_7: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19markFilterEdges_gpuPiS_S_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .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 _Z19markFilterEdges_gpuPiS_S_S_i, .Lfunc_end0-_Z19markFilterEdges_gpuPiS_S_S_i .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: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z19markFilterEdges_gpuPiS_S_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19markFilterEdges_gpuPiS_S_S_i.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.
/* ********************************************** * CS314 Principles of Programming Languages * * Spring 2020 * ********************************************** */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void markFilterEdges_gpu(int * src, int * dst, int * matches, int * keepEdges, int numEdges) { /*YOUR CODE HERE*/ //basically check if source or dst is matched, then set to 1 //if not matched then I can set to keepEdges[tid] = 0 int tid = blockIdx.x * blockDim.x + threadIdx.x;//this is the thread id and will be used as index while( tid < numEdges ) { if( matches[src[tid]] == -1 && matches[dst[tid]] == -1 )//either source or destination has an unmatched node, meaning we want to keep this edge { keepEdges[tid] = 1;//if we want to keep the edge, set it to 1 } else { keepEdges[tid] = 0;//otherwise set it to 0 } tid += ( blockDim.x * gridDim.x ); } return; }
.text .file "markFilterEdges.hip" .globl _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i # -- Begin function _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .p2align 4, 0x90 .type _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i,@function _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i: # @_Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .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 $_Z19markFilterEdges_gpuPiS_S_S_i, %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_end0: .size _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i, .Lfunc_end0-_Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .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 $_Z19markFilterEdges_gpuPiS_S_S_i, %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 _Z19markFilterEdges_gpuPiS_S_S_i,@object # @_Z19markFilterEdges_gpuPiS_S_S_i .section .rodata,"a",@progbits .globl _Z19markFilterEdges_gpuPiS_S_S_i .p2align 3, 0x0 _Z19markFilterEdges_gpuPiS_S_S_i: .quad _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .size _Z19markFilterEdges_gpuPiS_S_S_i, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z19markFilterEdges_gpuPiS_S_S_i" .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 _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19markFilterEdges_gpuPiS_S_S_i .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 : _Z19markFilterEdges_gpuPiS_S_S_i .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*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0070*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fc800078e00ff */ /*0080*/ IMAD.WIDE R2, R0, R7, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x001fcc00078e0207 */ /*0090*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea4000c1e1900 */ /*00a0*/ IMAD.WIDE R4, R2, R7, c[0x0][0x170] ; /* 0x00005c0002047625 */ /* 0x004fcc00078e0207 */ /*00b0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ SHF.R.S32.HI R11, RZ, 0x1f, R0 ; /* 0x0000001fff0b7819 */ /* 0x000fe40000011400 */ /*00d0*/ ISETP.NE.AND P0, PT, R4, -0x1, PT ; /* 0xffffffff0400780c */ /* 0x004fda0003f05270 */ /*00e0*/ @!P0 LEA R8, P1, R0, c[0x0][0x168], 0x2 ; /* 0x00005a0000088a11 */ /* 0x000fc800078210ff */ /*00f0*/ @!P0 LEA.HI.X R9, R0, c[0x0][0x16c], R11, 0x2, P1 ; /* 0x00005b0000098a11 */ /* 0x000fca00008f140b */ /*0100*/ @!P0 LDG.E R6, [R8.64] ; /* 0x0000000408068981 */ /* 0x000ea4000c1e1900 */ /*0110*/ @!P0 IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006068625 */ /* 0x004fcc00078e0207 */ /*0120*/ @!P0 LDG.E R6, [R6.64] ; /* 0x0000000406068981 */ /* 0x000ea2000c1e1900 */ /*0130*/ LEA R10, P1, R0, c[0x0][0x178], 0x2 ; /* 0x00005e00000a7a11 */ /* 0x000fe200078210ff */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff057624 */ /* 0x000fc600078e00ff */ /*0150*/ LEA.HI.X R11, R0, c[0x0][0x17c], R11, 0x2, P1 ; /* 0x00005f00000b7a11 */ /* 0x000fe200008f140b */ /*0160*/ IMAD R0, R5, c[0x0][0xc], R0 ; /* 0x0000030005007a24 */ /* 0x000fe200078e0200 */ /*0170*/ ISETP.EQ.AND P0, PT, R6, -0x1, !P0 ; /* 0xffffffff0600780c */ /* 0x004fc80004702270 */ /*0180*/ SEL R3, RZ, 0x1, !P0 ; /* 0x00000001ff037807 */ /* 0x000fe40004000000 */ /*0190*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fc60003f06270 */ /*01a0*/ STG.E [R10.64], R3 ; /* 0x000000030a007986 */ /* 0x0001f4000c101904 */ /*01b0*/ @!P0 BRA 0x70 ; /* 0xfffffeb000008947 */ /* 0x000fea000383ffff */ /*01c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*01d0*/ BRA 0x1d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z19markFilterEdges_gpuPiS_S_S_i .globl _Z19markFilterEdges_gpuPiS_S_S_i .p2align 8 .type _Z19markFilterEdges_gpuPiS_S_S_i,@function _Z19markFilterEdges_gpuPiS_S_S_i: s_clause 0x1 s_load_b32 s4, s[0:1], 0x34 s_load_b32 s12, s[0:1], 0x20 s_add_u32 s2, s0, 40 s_addc_u32 s3, s1, 0 s_waitcnt lgkmcnt(0) s_and_b32 s13, s4, 0xffff s_mov_b32 s4, exec_lo v_mad_u64_u32 v[1:2], null, s15, s13, v[0:1] s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s12, v1 s_cbranch_execz .LBB0_7 s_load_b32 s2, s[2:3], 0x0 s_load_b256 s[4:11], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_mul_i32 s1, s2, s13 s_mov_b32 s2, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s3 v_add_nc_u32_e32 v1, s1, v1 v_add_co_u32 v2, s0, s10, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_ci_u32_e64 v3, s0, s11, v3, s0 v_cmp_le_i32_e32 vcc_lo, s12, v1 global_store_b32 v[2:3], v0, off s_or_b32 s2, vcc_lo, s2 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s2 s_cbranch_execz .LBB0_7 .LBB0_3: v_ashrrev_i32_e32 v2, 31, v1 s_mov_b32 s3, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[2:3], 2, v[1:2] v_add_co_u32 v4, vcc_lo, s4, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v5, vcc_lo, s5, v3, vcc_lo global_load_b32 v4, v[4:5], off s_waitcnt vmcnt(0) v_ashrrev_i32_e32 v5, 31, v4 v_lshlrev_b64 v[4:5], 2, v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v4, vcc_lo, s8, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s9, v5, vcc_lo global_load_b32 v0, v[4:5], off s_waitcnt vmcnt(0) v_cmp_ne_u32_e64 s0, -1, v0 v_cmpx_eq_u32_e32 -1, v0 s_cbranch_execz .LBB0_5 v_add_co_u32 v4, vcc_lo, s6, v2 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v3, vcc_lo s_and_not1_b32 s0, s0, exec_lo s_mov_b32 s13, 1 global_load_b32 v4, v[4:5], off s_waitcnt vmcnt(0) 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], 2, v[4:5] v_add_co_u32 v4, vcc_lo, s8, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_4) | instid1(SALU_CYCLE_1) v_add_co_ci_u32_e32 v5, vcc_lo, s9, v5, vcc_lo global_load_b32 v0, v[4:5], off s_waitcnt vmcnt(0) v_cmp_ne_u32_e32 vcc_lo, -1, v0 s_and_b32 s14, vcc_lo, exec_lo s_or_b32 s0, s0, s14 .LBB0_5: s_or_b32 exec_lo, exec_lo, s3 v_mov_b32_e32 v0, s13 s_and_saveexec_b32 s3, s0 s_cbranch_execz .LBB0_2 v_mov_b32_e32 v0, 0 s_branch .LBB0_2 .LBB0_7: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19markFilterEdges_gpuPiS_S_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .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 _Z19markFilterEdges_gpuPiS_S_S_i, .Lfunc_end0-_Z19markFilterEdges_gpuPiS_S_S_i .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: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z19markFilterEdges_gpuPiS_S_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19markFilterEdges_gpuPiS_S_S_i.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_000d8b08_00000000-6_markFilterEdges.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .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 .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i .type _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i, @function _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i: .LFB2081: .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 .L7 .L3: movq 152(%rsp), %rax subq %fs:40, %rax jne .L8 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .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 _Z19markFilterEdges_gpuPiS_S_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i, .-_Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i .globl _Z19markFilterEdges_gpuPiS_S_S_i .type _Z19markFilterEdges_gpuPiS_S_S_i, @function _Z19markFilterEdges_gpuPiS_S_S_i: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z46__device_stub__Z19markFilterEdges_gpuPiS_S_S_iPiS_S_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size _Z19markFilterEdges_gpuPiS_S_S_i, .-_Z19markFilterEdges_gpuPiS_S_S_i .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z19markFilterEdges_gpuPiS_S_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .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 _Z19markFilterEdges_gpuPiS_S_S_i(%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 .LFE2084: .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 "markFilterEdges.hip" .globl _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i # -- Begin function _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .p2align 4, 0x90 .type _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i,@function _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i: # @_Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .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 $_Z19markFilterEdges_gpuPiS_S_S_i, %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_end0: .size _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i, .Lfunc_end0-_Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .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 $_Z19markFilterEdges_gpuPiS_S_S_i, %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 _Z19markFilterEdges_gpuPiS_S_S_i,@object # @_Z19markFilterEdges_gpuPiS_S_S_i .section .rodata,"a",@progbits .globl _Z19markFilterEdges_gpuPiS_S_S_i .p2align 3, 0x0 _Z19markFilterEdges_gpuPiS_S_S_i: .quad _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .size _Z19markFilterEdges_gpuPiS_S_S_i, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z19markFilterEdges_gpuPiS_S_S_i" .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 _Z34__device_stub__markFilterEdges_gpuPiS_S_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19markFilterEdges_gpuPiS_S_S_i .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> extern "C" { __global__ void helloWorldKernel(int x); } __global__ void helloWorldKernel(int x) { printf("Hello World %d\n",x); return; }
code for sm_80 Function : helloWorldKernel .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 */ /* 0x000fca00078e00ff */ /*0010*/ IADD3 R1, R1, -0x8, RZ ; /* 0xfffffff801017810 */ /* 0x000fe20007ffe0ff */ /*0020*/ IMAD.MOV.U32 R8, RZ, RZ, c[0x0][0x160] ; /* 0x00005800ff087624 */ /* 0x000fe200078e00ff */ /*0030*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe20000000f00 */ /*0040*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff047624 */ /* 0x000fe200078e00ff */ /*0050*/ IADD3 R6, P0, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fe20007f1e0ff */ /*0060*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff057624 */ /* 0x000fe200078e00ff */ /*0070*/ STL [R1], R8 ; /* 0x0000000801007387 */ /* 0x0001e20000100800 */ /*0080*/ LDC.64 R2, c[0x4][R0] ; /* 0x0100000000027b82 */ /* 0x0000640000000a00 */ /*0090*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P0 ; /* 0x00000900ff077624 */ /* 0x000fcc00000e06ff */ /*00a0*/ LEPC R8 ; /* 0x000000000008734e */ /* 0x001fe40000000000 */ /*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 */ /* 0x002fea0003c00000 */ /*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> extern "C" { __global__ void helloWorldKernel(int x); } __global__ void helloWorldKernel(int x) { printf("Hello World %d\n",x); return; }
.file "tmpxft_0012f284_00000000-6_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .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 .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z35__device_stub__Z16helloWorldKernelii .type _Z35__device_stub__Z16helloWorldKernelii, @function _Z35__device_stub__Z16helloWorldKernelii: .LFB2081: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movl %edi, 12(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 12(%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 .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .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 helloWorldKernel(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z35__device_stub__Z16helloWorldKernelii, .-_Z35__device_stub__Z16helloWorldKernelii .globl helloWorldKernel .type helloWorldKernel, @function helloWorldKernel: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z16helloWorldKernelii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size helloWorldKernel, .-helloWorldKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "helloWorldKernel" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .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 helloWorldKernel(%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 .LFE2084: .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> extern "C" { __global__ void helloWorldKernel(int x); } __global__ void helloWorldKernel(int x) { printf("Hello World %d\n",x); return; }
#include <hip/hip_runtime.h> #include <stdio.h> extern "C" { __global__ void helloWorldKernel(int x); } __global__ void helloWorldKernel(int x) { printf("Hello World %d\n",x); return; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> extern "C" { __global__ void helloWorldKernel(int x); } __global__ void helloWorldKernel(int x) { printf("Hello World %d\n",x); return; }
.text .file "kernel.hip" .globl __device_stub__helloWorldKernel # -- Begin function __device_stub__helloWorldKernel .p2align 4, 0x90 .type __device_stub__helloWorldKernel,@function __device_stub__helloWorldKernel: # @__device_stub__helloWorldKernel .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movl %edi, 12(%rsp) leaq 12(%rsp), %rax movq %rax, 16(%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 16(%rsp), %r9 movl $helloWorldKernel, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size __device_stub__helloWorldKernel, .Lfunc_end0-__device_stub__helloWorldKernel .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 $helloWorldKernel, %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 helloWorldKernel,@object # @helloWorldKernel .section .rodata,"a",@progbits .globl helloWorldKernel .p2align 3, 0x0 helloWorldKernel: .quad __device_stub__helloWorldKernel .size helloWorldKernel, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "helloWorldKernel" .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 __device_stub__helloWorldKernel .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym helloWorldKernel .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_0012f284_00000000-6_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .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 .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z35__device_stub__Z16helloWorldKernelii .type _Z35__device_stub__Z16helloWorldKernelii, @function _Z35__device_stub__Z16helloWorldKernelii: .LFB2081: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movl %edi, 12(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 12(%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 .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .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 helloWorldKernel(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z35__device_stub__Z16helloWorldKernelii, .-_Z35__device_stub__Z16helloWorldKernelii .globl helloWorldKernel .type helloWorldKernel, @function helloWorldKernel: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z16helloWorldKernelii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size helloWorldKernel, .-helloWorldKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "helloWorldKernel" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .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 helloWorldKernel(%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 .LFE2084: .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 "kernel.hip" .globl __device_stub__helloWorldKernel # -- Begin function __device_stub__helloWorldKernel .p2align 4, 0x90 .type __device_stub__helloWorldKernel,@function __device_stub__helloWorldKernel: # @__device_stub__helloWorldKernel .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movl %edi, 12(%rsp) leaq 12(%rsp), %rax movq %rax, 16(%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 16(%rsp), %r9 movl $helloWorldKernel, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size __device_stub__helloWorldKernel, .Lfunc_end0-__device_stub__helloWorldKernel .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 $helloWorldKernel, %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 helloWorldKernel,@object # @helloWorldKernel .section .rodata,"a",@progbits .globl helloWorldKernel .p2align 3, 0x0 helloWorldKernel: .quad __device_stub__helloWorldKernel .size helloWorldKernel, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "helloWorldKernel" .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 __device_stub__helloWorldKernel .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym helloWorldKernel .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 HIP/AMD source code.
#include<iostream> #include<string> #include<malloc.h> #include<fstream> #include<sstream> #include<vector> #include<cmath> #include<cstdio> #include<stdlib.h> #include<cuda.h> #include<cuda_runtime.h> #define MAX_CLUSTER_SIZE 6 #define BLOCK_SIZE 16 using namespace std; typedef vector<double> record_t; typedef vector<record_t> data_t; long long int max_atoms=0; long long int max_energy_states=0; int k_clus,cls; double cluster_lambda=0.0; int f=0; istream& operator >>(istream& ins, record_t &record) { record.clear(); string line; getline(ins,line); stringstream ss(line); string field; while(getline(ss,field,',')) { stringstream fs(field); double d=0.0; fs>>d; record.push_back(d); } return ins; } istream& operator >> (istream& ins, data_t& data) { data.clear(); record_t record; while(ins>>record) { data.push_back(record); } return ins; } __global__ void findMeans(const float *cluster, const float *instance, int *rnk, const int num, const int clus_size, const int maxAtoms) { int i = blockIdx.x * blockDim.x + threadIdx.x; // int i = threadIdx.x; if(i<num) { float min=0.0; for(int j=0;j<maxAtoms;j++) { float d=0.0; float s=0.0; for(int k=0;k<3;k++) { d=instance[(j*5)+k+2]-cluster[((0*clus_size)+j)*maxAtoms +k]; s=s+(d*d); } s=sqrt(s); min+=s; } int cls=-1; for(int clus=0;clus<clus_size;clus++) { float val=0.0; for(int p=0;p<maxAtoms;p++) { float d=0.0; float s=0.0; for(int l=0;l<3;l++) { d=instance[((i*maxAtoms+p)*5)+l+2]-cluster[((clus*clus_size)+p)*maxAtoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[i]=cls+1; } } int main(void) { cudaError_t err=cudaSuccess; data_t data; ifstream infile("alanine.csv"); infile>>data; if(!infile.eof()) { cout<<"Error! Please check file.\n"; return 1; } infile.close(); cout<<"Number of rows in file : "<<data.size()<<endl; for(long long int i=0;i<data.size();i++) { if(i>0) { if(data[i][1]<data[i-1][1]) { max_atoms=i; break; } } } cout<<"Number of atoms : "<<max_atoms<<endl; max_energy_states=data[data.size()-1][0]; // max_energy_states=32; cout<<"Total number of conformations : "<<max_energy_states<<endl; //allocate host cluster int s = MAX_CLUSTER_SIZE*max_atoms*3; size_t sz = s*sizeof(float); float *host_cluster=(float *)malloc(sz); if(host_cluster==NULL) cout<<"Not able to allocate host cluster"<<endl; //Initializing host cluster for(int i=0;i<MAX_CLUSTER_SIZE;i++) { int clus_num=rand()%max_energy_states + 0; for(long long int ii=0;ii<max_atoms;ii++) { for(int jj=0;jj<3;jj++) { host_cluster[((i*MAX_CLUSTER_SIZE)+ii)*max_atoms+jj]=data[clus_num*max_atoms+ii][jj+2]; } } } //allocate device cluster float *d_clus = NULL; cudaMalloc((void **)&d_clus, sz); if(err!=cudaSuccess) { cout<<"Failed to allocate cluster on device!\n"; exit(EXIT_FAILURE); } //initializing device cluster cudaMemcpy(d_clus, host_cluster, sz, cudaMemcpyHostToDevice); if(err!=cudaSuccess) { cout<<"Failed to copy cluster to device\n"; exit(EXIT_FAILURE); } //allocate host instance float *host_instances; long int row=data.size(); int col=5; long int size=row*col; host_instances=new float[size]; for(long long int i=0;i<row;i++) { for(int j=0;j<col;j++) { host_instances[i*col+j]=data[i][j]; } } //allocate device instance float *device_instances; err=cudaMalloc((void **)&device_instances,size*sizeof(float)); if(err!=cudaSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } err=cudaMemcpy(device_instances, host_instances, size*sizeof(float), cudaMemcpyHostToDevice); if(err!=cudaSuccess) { cout<<"error copying device instance"<<endl; exit(EXIT_FAILURE); } int *rnk_host; int rnk[max_energy_states]; rnk_host=new int[max_energy_states]; for(long long int i=0;i<max_energy_states;i++) { rnk_host[i]=0; rnk[i]=0; } int *rnk_device; err=cudaMalloc((void **)&rnk_device,max_energy_states*sizeof(int)); if(err!=cudaSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } err=cudaMemcpy(rnk_device, rnk_host, max_energy_states*sizeof(int), cudaMemcpyHostToDevice); if(err!=cudaSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } // dim3 dimBlock(BLOCK_SIZE,BLOCK_SIZE); int num_blocks=(max_energy_states+511)/512; //while(1) //{ findMeans<<<48,256>>>(d_clus, device_instances, rnk_device, max_energy_states, MAX_CLUSTER_SIZE, max_atoms); err=cudaMemcpy(rnk_host, rnk_device, max_energy_states*sizeof(int), cudaMemcpyDeviceToHost); if(err!=cudaSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } //while(1) //{ //E - Step for(long long int j=0;j<max_energy_states;j++) { double min=0.0; for(long long int y=0;y<max_atoms;y++) { double d=0.0,s=0.0; for(int z=0;z<3;z++) { d=host_instances[(y*5)+z+2]-host_cluster[((0*MAX_CLUSTER_SIZE)+y)*max_atoms +z]; s=s+(d*d); } s=sqrt(s); min+=s; } cls=-1; for(int clus=0;clus<MAX_CLUSTER_SIZE;clus++) { double val=0.0; for(long long int p=0;p<max_atoms;p++) { double d=0.0,s=0.0; for(int l=0;l<3;l++) { d=data[j*max_atoms+p][l+2]-host_cluster[((clus*MAX_CLUSTER_SIZE)+p)*max_atoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[j]=cls+1; } int flag=0; for(long int i=0;i<100;i++) { if(rnk[i]!=rnk_host[i]) flag=1; cout<<rnk[i]<<"\t"<<rnk_host[i]<<endl; } if(flag==1) cout<<"Not Matching"<<endl; /* //M - Step int sizes[MAX_CLUSTER_SIZE]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) sizes[jj]=0; double means[MAX_CLUSTER_SIZE][max_atoms][3]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { means[jj][kk][ll]=0; } } } for(long long int jj=0;jj<max_energy_states;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { double d=data[jj*max_atoms+kk][ll+2]; means[rnk[jj]-1][kk][ll]+=d; sizes[rnk[jj]-1]++; } } } for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { if(cluster[jj][kk][ll]!=means[jj][kk][ll]/sizes[jj]) { f=1; } if(sizes[jj]!=0) cluster[jj][kk][ll]=means[jj][kk][ll]/sizes[jj]; else cluster[jj][kk][ll]=0; } } } if(f==0) break; f=0; }*/ /*for(int i=0;i<max_energy_states;i++) { cout<<rnk_host[i]<<" "; } cout<<endl;*/ return 0; }
#include<iostream> #include<string> #include<malloc.h> #include<fstream> #include<sstream> #include<vector> #include<cmath> #include<cstdio> #include<stdlib.h> #include<hip/hip_runtime.h> #define MAX_CLUSTER_SIZE 6 #define BLOCK_SIZE 16 using namespace std; typedef vector<double> record_t; typedef vector<record_t> data_t; long long int max_atoms=0; long long int max_energy_states=0; int k_clus,cls; double cluster_lambda=0.0; int f=0; istream& operator >>(istream& ins, record_t &record) { record.clear(); string line; getline(ins,line); stringstream ss(line); string field; while(getline(ss,field,',')) { stringstream fs(field); double d=0.0; fs>>d; record.push_back(d); } return ins; } istream& operator >> (istream& ins, data_t& data) { data.clear(); record_t record; while(ins>>record) { data.push_back(record); } return ins; } __global__ void findMeans(const float *cluster, const float *instance, int *rnk, const int num, const int clus_size, const int maxAtoms) { int i = blockIdx.x * blockDim.x + threadIdx.x; // int i = threadIdx.x; if(i<num) { float min=0.0; for(int j=0;j<maxAtoms;j++) { float d=0.0; float s=0.0; for(int k=0;k<3;k++) { d=instance[(j*5)+k+2]-cluster[((0*clus_size)+j)*maxAtoms +k]; s=s+(d*d); } s=sqrt(s); min+=s; } int cls=-1; for(int clus=0;clus<clus_size;clus++) { float val=0.0; for(int p=0;p<maxAtoms;p++) { float d=0.0; float s=0.0; for(int l=0;l<3;l++) { d=instance[((i*maxAtoms+p)*5)+l+2]-cluster[((clus*clus_size)+p)*maxAtoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[i]=cls+1; } } int main(void) { hipError_t err=hipSuccess; data_t data; ifstream infile("alanine.csv"); infile>>data; if(!infile.eof()) { cout<<"Error! Please check file.\n"; return 1; } infile.close(); cout<<"Number of rows in file : "<<data.size()<<endl; for(long long int i=0;i<data.size();i++) { if(i>0) { if(data[i][1]<data[i-1][1]) { max_atoms=i; break; } } } cout<<"Number of atoms : "<<max_atoms<<endl; max_energy_states=data[data.size()-1][0]; // max_energy_states=32; cout<<"Total number of conformations : "<<max_energy_states<<endl; //allocate host cluster int s = MAX_CLUSTER_SIZE*max_atoms*3; size_t sz = s*sizeof(float); float *host_cluster=(float *)malloc(sz); if(host_cluster==NULL) cout<<"Not able to allocate host cluster"<<endl; //Initializing host cluster for(int i=0;i<MAX_CLUSTER_SIZE;i++) { int clus_num=rand()%max_energy_states + 0; for(long long int ii=0;ii<max_atoms;ii++) { for(int jj=0;jj<3;jj++) { host_cluster[((i*MAX_CLUSTER_SIZE)+ii)*max_atoms+jj]=data[clus_num*max_atoms+ii][jj+2]; } } } //allocate device cluster float *d_clus = NULL; hipMalloc((void **)&d_clus, sz); if(err!=hipSuccess) { cout<<"Failed to allocate cluster on device!\n"; exit(EXIT_FAILURE); } //initializing device cluster hipMemcpy(d_clus, host_cluster, sz, hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"Failed to copy cluster to device\n"; exit(EXIT_FAILURE); } //allocate host instance float *host_instances; long int row=data.size(); int col=5; long int size=row*col; host_instances=new float[size]; for(long long int i=0;i<row;i++) { for(int j=0;j<col;j++) { host_instances[i*col+j]=data[i][j]; } } //allocate device instance float *device_instances; err=hipMalloc((void **)&device_instances,size*sizeof(float)); if(err!=hipSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } err=hipMemcpy(device_instances, host_instances, size*sizeof(float), hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"error copying device instance"<<endl; exit(EXIT_FAILURE); } int *rnk_host; int rnk[max_energy_states]; rnk_host=new int[max_energy_states]; for(long long int i=0;i<max_energy_states;i++) { rnk_host[i]=0; rnk[i]=0; } int *rnk_device; err=hipMalloc((void **)&rnk_device,max_energy_states*sizeof(int)); if(err!=hipSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } err=hipMemcpy(rnk_device, rnk_host, max_energy_states*sizeof(int), hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } // dim3 dimBlock(BLOCK_SIZE,BLOCK_SIZE); int num_blocks=(max_energy_states+511)/512; //while(1) //{ findMeans<<<48,256>>>(d_clus, device_instances, rnk_device, max_energy_states, MAX_CLUSTER_SIZE, max_atoms); err=hipMemcpy(rnk_host, rnk_device, max_energy_states*sizeof(int), hipMemcpyDeviceToHost); if(err!=hipSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } //while(1) //{ //E - Step for(long long int j=0;j<max_energy_states;j++) { double min=0.0; for(long long int y=0;y<max_atoms;y++) { double d=0.0,s=0.0; for(int z=0;z<3;z++) { d=host_instances[(y*5)+z+2]-host_cluster[((0*MAX_CLUSTER_SIZE)+y)*max_atoms +z]; s=s+(d*d); } s=sqrt(s); min+=s; } cls=-1; for(int clus=0;clus<MAX_CLUSTER_SIZE;clus++) { double val=0.0; for(long long int p=0;p<max_atoms;p++) { double d=0.0,s=0.0; for(int l=0;l<3;l++) { d=data[j*max_atoms+p][l+2]-host_cluster[((clus*MAX_CLUSTER_SIZE)+p)*max_atoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[j]=cls+1; } int flag=0; for(long int i=0;i<100;i++) { if(rnk[i]!=rnk_host[i]) flag=1; cout<<rnk[i]<<"\t"<<rnk_host[i]<<endl; } if(flag==1) cout<<"Not Matching"<<endl; /* //M - Step int sizes[MAX_CLUSTER_SIZE]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) sizes[jj]=0; double means[MAX_CLUSTER_SIZE][max_atoms][3]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { means[jj][kk][ll]=0; } } } for(long long int jj=0;jj<max_energy_states;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { double d=data[jj*max_atoms+kk][ll+2]; means[rnk[jj]-1][kk][ll]+=d; sizes[rnk[jj]-1]++; } } } for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { if(cluster[jj][kk][ll]!=means[jj][kk][ll]/sizes[jj]) { f=1; } if(sizes[jj]!=0) cluster[jj][kk][ll]=means[jj][kk][ll]/sizes[jj]; else cluster[jj][kk][ll]=0; } } } if(f==0) break; f=0; }*/ /*for(int i=0;i<max_energy_states;i++) { cout<<rnk_host[i]<<" "; } cout<<endl;*/ return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include<iostream> #include<string> #include<malloc.h> #include<fstream> #include<sstream> #include<vector> #include<cmath> #include<cstdio> #include<stdlib.h> #include<hip/hip_runtime.h> #define MAX_CLUSTER_SIZE 6 #define BLOCK_SIZE 16 using namespace std; typedef vector<double> record_t; typedef vector<record_t> data_t; long long int max_atoms=0; long long int max_energy_states=0; int k_clus,cls; double cluster_lambda=0.0; int f=0; istream& operator >>(istream& ins, record_t &record) { record.clear(); string line; getline(ins,line); stringstream ss(line); string field; while(getline(ss,field,',')) { stringstream fs(field); double d=0.0; fs>>d; record.push_back(d); } return ins; } istream& operator >> (istream& ins, data_t& data) { data.clear(); record_t record; while(ins>>record) { data.push_back(record); } return ins; } __global__ void findMeans(const float *cluster, const float *instance, int *rnk, const int num, const int clus_size, const int maxAtoms) { int i = blockIdx.x * blockDim.x + threadIdx.x; // int i = threadIdx.x; if(i<num) { float min=0.0; for(int j=0;j<maxAtoms;j++) { float d=0.0; float s=0.0; for(int k=0;k<3;k++) { d=instance[(j*5)+k+2]-cluster[((0*clus_size)+j)*maxAtoms +k]; s=s+(d*d); } s=sqrt(s); min+=s; } int cls=-1; for(int clus=0;clus<clus_size;clus++) { float val=0.0; for(int p=0;p<maxAtoms;p++) { float d=0.0; float s=0.0; for(int l=0;l<3;l++) { d=instance[((i*maxAtoms+p)*5)+l+2]-cluster[((clus*clus_size)+p)*maxAtoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[i]=cls+1; } } int main(void) { hipError_t err=hipSuccess; data_t data; ifstream infile("alanine.csv"); infile>>data; if(!infile.eof()) { cout<<"Error! Please check file.\n"; return 1; } infile.close(); cout<<"Number of rows in file : "<<data.size()<<endl; for(long long int i=0;i<data.size();i++) { if(i>0) { if(data[i][1]<data[i-1][1]) { max_atoms=i; break; } } } cout<<"Number of atoms : "<<max_atoms<<endl; max_energy_states=data[data.size()-1][0]; // max_energy_states=32; cout<<"Total number of conformations : "<<max_energy_states<<endl; //allocate host cluster int s = MAX_CLUSTER_SIZE*max_atoms*3; size_t sz = s*sizeof(float); float *host_cluster=(float *)malloc(sz); if(host_cluster==NULL) cout<<"Not able to allocate host cluster"<<endl; //Initializing host cluster for(int i=0;i<MAX_CLUSTER_SIZE;i++) { int clus_num=rand()%max_energy_states + 0; for(long long int ii=0;ii<max_atoms;ii++) { for(int jj=0;jj<3;jj++) { host_cluster[((i*MAX_CLUSTER_SIZE)+ii)*max_atoms+jj]=data[clus_num*max_atoms+ii][jj+2]; } } } //allocate device cluster float *d_clus = NULL; hipMalloc((void **)&d_clus, sz); if(err!=hipSuccess) { cout<<"Failed to allocate cluster on device!\n"; exit(EXIT_FAILURE); } //initializing device cluster hipMemcpy(d_clus, host_cluster, sz, hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"Failed to copy cluster to device\n"; exit(EXIT_FAILURE); } //allocate host instance float *host_instances; long int row=data.size(); int col=5; long int size=row*col; host_instances=new float[size]; for(long long int i=0;i<row;i++) { for(int j=0;j<col;j++) { host_instances[i*col+j]=data[i][j]; } } //allocate device instance float *device_instances; err=hipMalloc((void **)&device_instances,size*sizeof(float)); if(err!=hipSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } err=hipMemcpy(device_instances, host_instances, size*sizeof(float), hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"error copying device instance"<<endl; exit(EXIT_FAILURE); } int *rnk_host; int rnk[max_energy_states]; rnk_host=new int[max_energy_states]; for(long long int i=0;i<max_energy_states;i++) { rnk_host[i]=0; rnk[i]=0; } int *rnk_device; err=hipMalloc((void **)&rnk_device,max_energy_states*sizeof(int)); if(err!=hipSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } err=hipMemcpy(rnk_device, rnk_host, max_energy_states*sizeof(int), hipMemcpyHostToDevice); if(err!=hipSuccess) { cout<<"error allocating device instance"<<endl; exit(EXIT_FAILURE); } // dim3 dimBlock(BLOCK_SIZE,BLOCK_SIZE); int num_blocks=(max_energy_states+511)/512; //while(1) //{ findMeans<<<48,256>>>(d_clus, device_instances, rnk_device, max_energy_states, MAX_CLUSTER_SIZE, max_atoms); err=hipMemcpy(rnk_host, rnk_device, max_energy_states*sizeof(int), hipMemcpyDeviceToHost); if(err!=hipSuccess) { cout<<"error allocating device rnk"<<endl; exit(EXIT_FAILURE); } //while(1) //{ //E - Step for(long long int j=0;j<max_energy_states;j++) { double min=0.0; for(long long int y=0;y<max_atoms;y++) { double d=0.0,s=0.0; for(int z=0;z<3;z++) { d=host_instances[(y*5)+z+2]-host_cluster[((0*MAX_CLUSTER_SIZE)+y)*max_atoms +z]; s=s+(d*d); } s=sqrt(s); min+=s; } cls=-1; for(int clus=0;clus<MAX_CLUSTER_SIZE;clus++) { double val=0.0; for(long long int p=0;p<max_atoms;p++) { double d=0.0,s=0.0; for(int l=0;l<3;l++) { d=data[j*max_atoms+p][l+2]-host_cluster[((clus*MAX_CLUSTER_SIZE)+p)*max_atoms +l]; s=s+(d*d); } s=sqrt(s); val+=s; } if(val<=min) { cls=clus; min=val; } } rnk[j]=cls+1; } int flag=0; for(long int i=0;i<100;i++) { if(rnk[i]!=rnk_host[i]) flag=1; cout<<rnk[i]<<"\t"<<rnk_host[i]<<endl; } if(flag==1) cout<<"Not Matching"<<endl; /* //M - Step int sizes[MAX_CLUSTER_SIZE]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) sizes[jj]=0; double means[MAX_CLUSTER_SIZE][max_atoms][3]; for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { means[jj][kk][ll]=0; } } } for(long long int jj=0;jj<max_energy_states;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { double d=data[jj*max_atoms+kk][ll+2]; means[rnk[jj]-1][kk][ll]+=d; sizes[rnk[jj]-1]++; } } } for(int jj=0;jj<MAX_CLUSTER_SIZE;jj++) { for(long long int kk=0;kk<max_atoms;kk++) { for(int ll=0;ll<3;ll++) { if(cluster[jj][kk][ll]!=means[jj][kk][ll]/sizes[jj]) { f=1; } if(sizes[jj]!=0) cluster[jj][kk][ll]=means[jj][kk][ll]/sizes[jj]; else cluster[jj][kk][ll]=0; } } } if(f==0) break; f=0; }*/ /*for(int i=0;i<max_energy_states;i++) { cout<<rnk_host[i]<<" "; } cout<<endl;*/ return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z9findMeansPKfS0_Piiii .globl _Z9findMeansPKfS0_Piiii .p2align 8 .type _Z9findMeansPKfS0_Piiii,@function _Z9findMeansPKfS0_Piiii: s_clause 0x1 s_load_b32 s2, s[0:1], 0x34 s_load_b32 s3, s[0:1], 0x18 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_i32_e64 s3, v1 s_cbranch_execz .LBB0_18 s_clause 0x1 s_load_b32 s10, s[0:1], 0x20 s_load_b128 s[4:7], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s10, 1 s_cbranch_scc1 .LBB0_6 v_mov_b32_e32 v0, 0 s_mov_b32 s9, 0 s_add_u32 s11, s6, 8 s_addc_u32 s12, s7, 0 s_mov_b32 s8, s9 s_mov_b32 s13, s9 .LBB0_3: s_lshl_b64 s[2:3], s[8:9], 2 v_mov_b32_e32 v2, 0 s_add_u32 s14, s4, s2 s_addc_u32 s15, s5, s3 s_mov_b64 s[2:3], 0 .LBB0_4: s_delay_alu instid0(SALU_CYCLE_1) s_add_u32 s16, s11, s2 s_addc_u32 s17, s12, s3 s_add_u32 s18, s14, s2 s_addc_u32 s19, s15, s3 s_load_b32 s16, s[16:17], 0x0 s_load_b32 s17, s[18:19], 0x0 s_add_u32 s2, s2, 4 s_addc_u32 s3, s3, 0 s_cmp_eq_u32 s2, 12 s_waitcnt lgkmcnt(0) v_sub_f32_e64 v3, s16, s17 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v2, v3, v3 s_cbranch_scc0 .LBB0_4 s_delay_alu instid0(VALU_DEP_1) v_mul_f32_e32 v3, 0x4f800000, v2 v_cmp_gt_f32_e32 vcc_lo, 0xf800000, v2 s_add_i32 s13, s13, 1 s_add_i32 s8, s8, s10 s_add_u32 s11, s11, 20 s_addc_u32 s12, s12, 0 v_cndmask_b32_e32 v2, v2, v3, vcc_lo s_cmp_eq_u32 s13, s10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_sqrt_f32_e32 v3, v2 s_waitcnt_depctr 0xfff v_add_nc_u32_e32 v4, -1, v3 v_add_nc_u32_e32 v5, 1, v3 v_fma_f32 v6, -v4, v3, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_fma_f32 v7, -v5, v3, v2 v_cmp_ge_f32_e64 s2, 0, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v3, v3, v4, s2 v_cmp_lt_f32_e64 s2, 0, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e64 v3, v3, v5, s2 v_mul_f32_e32 v4, 0x37800000, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v3, v3, v4, vcc_lo v_cmp_class_f32_e64 vcc_lo, v2, 0x260 v_cndmask_b32_e32 v2, v3, v2, vcc_lo s_delay_alu instid0(VALU_DEP_1) v_add_f32_e32 v0, v0, v2 s_cbranch_scc0 .LBB0_3 s_branch .LBB0_7 .LBB0_6: v_mov_b32_e32 v0, 0 .LBB0_7: s_load_b32 s11, s[0:1], 0x1c s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s11, 1 s_cbranch_scc1 .LBB0_16 v_mul_lo_u32 v2, s10, v1 s_cmp_gt_i32 s10, 0 v_mov_b32_e32 v4, -1 s_cselect_b32 s13, -1, 0 s_add_u32 s6, s6, 8 s_mov_b32 s9, 0 s_mul_i32 s12, s10, s11 s_addc_u32 s7, s7, 0 v_lshl_add_u32 v5, v2, 2, v2 s_mov_b32 s14, 0 s_mov_b32 s15, 0 s_branch .LBB0_10 .LBB0_9: s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(SALU_CYCLE_1) v_cmp_nle_f32_e32 vcc_lo, v6, v0 s_add_i32 s14, s14, s12 v_cndmask_b32_e32 v0, v6, v0, vcc_lo v_cndmask_b32_e32 v4, s15, v4, vcc_lo s_add_i32 s15, s15, 1 s_cmp_eq_u32 s15, s11 s_cbranch_scc1 .LBB0_15 .LBB0_10: v_mov_b32_e32 v6, 0 s_and_not1_b32 vcc_lo, exec_lo, s13 s_cbranch_vccnz .LBB0_9 v_mov_b32_e32 v2, v5 s_mov_b32 s16, 0 s_mov_b32 s8, s14 .LBB0_12: s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_ashrrev_i32_e32 v3, 31, v2 s_lshl_b64 s[2:3], s[8:9], 2 s_add_u32 s17, s4, s2 s_addc_u32 s18, s5, s3 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_lshlrev_b64 v[7:8], 2, v[2:3] v_mov_b32_e32 v3, 0 s_mov_b64 s[2:3], 0 v_add_co_u32 v7, vcc_lo, s6, v7 s_delay_alu instid0(VALU_DEP_3) v_add_co_ci_u32_e32 v8, vcc_lo, s7, v8, vcc_lo .LBB0_13: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v9, vcc_lo, v7, s2 v_add_co_ci_u32_e32 v10, vcc_lo, s3, v8, vcc_lo s_add_u32 s20, s17, s2 s_addc_u32 s21, s18, s3 s_add_u32 s2, s2, 4 global_load_b32 v9, v[9:10], off s_load_b32 s19, s[20:21], 0x0 s_addc_u32 s3, s3, 0 s_cmp_eq_u32 s2, 12 s_waitcnt vmcnt(0) lgkmcnt(0) v_subrev_f32_e32 v9, s19, v9 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v3, v9, v9 s_cbranch_scc0 .LBB0_13 s_delay_alu instid0(VALU_DEP_1) v_dual_mul_f32 v7, 0x4f800000, v3 :: v_dual_add_nc_u32 v2, 5, v2 v_cmp_gt_f32_e32 vcc_lo, 0xf800000, v3 s_add_i32 s16, s16, 1 s_add_i32 s8, s8, s10 s_cmp_eq_u32 s16, s10 v_cndmask_b32_e32 v3, v3, v7, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_sqrt_f32_e32 v7, v3 s_waitcnt_depctr 0xfff v_add_nc_u32_e32 v8, -1, v7 v_add_nc_u32_e32 v9, 1, v7 v_fma_f32 v10, -v8, v7, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_fma_f32 v11, -v9, v7, v3 v_cmp_ge_f32_e64 s2, 0, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v7, v7, v8, s2 v_cmp_lt_f32_e64 s2, 0, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e64 v7, v7, v9, s2 v_mul_f32_e32 v8, 0x37800000, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v7, v7, v8, vcc_lo v_cmp_class_f32_e64 vcc_lo, v3, 0x260 v_cndmask_b32_e32 v3, v7, v3, vcc_lo s_delay_alu instid0(VALU_DEP_1) v_add_f32_e32 v6, v6, v3 s_cbranch_scc0 .LBB0_12 s_branch .LBB0_9 .LBB0_15: s_delay_alu instid0(VALU_DEP_1) v_add_nc_u32_e32 v0, 1, v4 s_branch .LBB0_17 .LBB0_16: v_mov_b32_e32 v0, 0 .LBB0_17: s_load_b64 s[0:1], s[0:1], 0x10 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[1:2], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v1, vcc_lo, s0, v1 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v2, vcc_lo, s1, v2, vcc_lo global_store_b32 v[1:2], v0, off .LBB0_18: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z9findMeansPKfS0_Piiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .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 22 .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 _Z9findMeansPKfS0_Piiii, .Lfunc_end0-_Z9findMeansPKfS0_Piiii .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: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z9findMeansPKfS0_Piiii .private_segment_fixed_size: 0 .sgpr_count: 24 .sgpr_spill_count: 0 .symbol: _Z9findMeansPKfS0_Piiii.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 CUDA source code to CUDA device assembly.
#include<stdio.h> #include<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<cudaGetErrorString(cudaGetLastError())<<std::endl; return 0; }
code for sm_80 Function : _Z3foov .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 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include<stdio.h> #include<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<cudaGetErrorString(cudaGetLastError())<<std::endl; return 0; }
.file "tmpxft_00121c8f_00000000-6_test.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .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 .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z21__device_stub__Z3foovv .type _Z21__device_stub__Z3foovv, @function _Z21__device_stub__Z3foovv: .LFB3694: .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 _Z3foov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z21__device_stub__Z3foovv, .-_Z21__device_stub__Z3foovv .globl _Z3foov .type _Z3foov, @function _Z3foov: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z21__device_stub__Z3foovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z3foov, .-_Z3foov .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "The Result is " .text .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $32, %rsp .cfi_def_cfa_offset 48 movl $1, 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: leaq .LC0(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rbx call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %rbx, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $0, %eax addq $32, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z21__device_stub__Z3foovv jmp .L12 .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z3foov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .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 _Z3foov(%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 .LFE3697: .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<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<cudaGetErrorString(cudaGetLastError())<<std::endl; return 0; }
#include <hip/hip_runtime.h> #include<stdio.h> #include<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<hipGetErrorString(hipGetLastError())<<std::endl; 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<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<hipGetErrorString(hipGetLastError())<<std::endl; return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3foov .globl _Z3foov .p2align 8 .type _Z3foov,@function _Z3foov: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3foov .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .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_end0: .size _Z3foov, .Lfunc_end0-_Z3foov .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: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3foov .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z3foov.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.
#include <hip/hip_runtime.h> #include<stdio.h> #include<iostream> __global__ void foo() {} int main() { foo<<<1,1>>>(); std::cout<<"The Result is " <<hipGetErrorString(hipGetLastError())<<std::endl; return 0; }
.text .file "test.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__foov # -- Begin function _Z18__device_stub__foov .p2align 4, 0x90 .type _Z18__device_stub__foov,@function _Z18__device_stub__foov: # @_Z18__device_stub__foov .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 $_Z3foov, %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 _Z18__device_stub__foov, .Lfunc_end0-_Z18__device_stub__foov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $64, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -16 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: 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 $_Z3foov, %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: movl $_ZSt4cout, %edi movl $.L.str, %esi movl $14, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l callq hipGetLastError movl %eax, %edi callq hipGetErrorString testq %rax, %rax je .LBB1_3 # %bb.4: movq %rax, %rdi movq %rax, %rbx callq strlen movl $_ZSt4cout, %edi movq %rbx, %rsi movq %rax, %rdx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l jmp .LBB1_5 .LBB1_3: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax leaq _ZSt4cout(%rax), %rdi movl _ZSt4cout+32(%rax), %esi orl $1, %esi callq _ZNSt9basic_iosIcSt11char_traitsIcEE5clearESt12_Ios_Iostate .LBB1_5: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %rbx testq %rbx, %rbx je .LBB1_10 # %bb.6: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB1_8 # %bb.7: movzbl 67(%rbx), %eax jmp .LBB1_9 .LBB1_8: movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB1_9: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %eax, %eax addq $64, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_10: .cfi_def_cfa_offset 80 callq _ZSt16__throw_bad_castv .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 $_Z3foov, %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 _Z3foov,@object # @_Z3foov .section .rodata,"a",@progbits .globl _Z3foov .p2align 3, 0x0 _Z3foov: .quad _Z18__device_stub__foov .size _Z3foov, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "The Result is " .size .L.str, 15 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z3foov" .size .L__unnamed_1, 8 .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 _Z18__device_stub__foov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3foov .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 : _Z3foov .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 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3foov .globl _Z3foov .p2align 8 .type _Z3foov,@function _Z3foov: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3foov .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .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_end0: .size _Z3foov, .Lfunc_end0-_Z3foov .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: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3foov .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z3foov.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_00121c8f_00000000-6_test.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .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 .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z21__device_stub__Z3foovv .type _Z21__device_stub__Z3foovv, @function _Z21__device_stub__Z3foovv: .LFB3694: .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 _Z3foov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z21__device_stub__Z3foovv, .-_Z21__device_stub__Z3foovv .globl _Z3foov .type _Z3foov, @function _Z3foov: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z21__device_stub__Z3foovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z3foov, .-_Z3foov .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "The Result is " .text .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $32, %rsp .cfi_def_cfa_offset 48 movl $1, 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: leaq .LC0(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rbx call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %rbx, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $0, %eax addq $32, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z21__device_stub__Z3foovv jmp .L12 .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z3foov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .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 _Z3foov(%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 .LFE3697: .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" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__foov # -- Begin function _Z18__device_stub__foov .p2align 4, 0x90 .type _Z18__device_stub__foov,@function _Z18__device_stub__foov: # @_Z18__device_stub__foov .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 $_Z3foov, %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 _Z18__device_stub__foov, .Lfunc_end0-_Z18__device_stub__foov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $64, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -16 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: 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 $_Z3foov, %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: movl $_ZSt4cout, %edi movl $.L.str, %esi movl $14, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l callq hipGetLastError movl %eax, %edi callq hipGetErrorString testq %rax, %rax je .LBB1_3 # %bb.4: movq %rax, %rdi movq %rax, %rbx callq strlen movl $_ZSt4cout, %edi movq %rbx, %rsi movq %rax, %rdx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l jmp .LBB1_5 .LBB1_3: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax leaq _ZSt4cout(%rax), %rdi movl _ZSt4cout+32(%rax), %esi orl $1, %esi callq _ZNSt9basic_iosIcSt11char_traitsIcEE5clearESt12_Ios_Iostate .LBB1_5: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %rbx testq %rbx, %rbx je .LBB1_10 # %bb.6: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB1_8 # %bb.7: movzbl 67(%rbx), %eax jmp .LBB1_9 .LBB1_8: movq %rbx, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) .LBB1_9: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %eax, %eax addq $64, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_10: .cfi_def_cfa_offset 80 callq _ZSt16__throw_bad_castv .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 $_Z3foov, %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 _Z3foov,@object # @_Z3foov .section .rodata,"a",@progbits .globl _Z3foov .p2align 3, 0x0 _Z3foov: .quad _Z18__device_stub__foov .size _Z3foov, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "The Result is " .size .L.str, 15 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z3foov" .size .L__unnamed_1, 8 .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 _Z18__device_stub__foov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3foov .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.
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
.file "tmpxft_00078055_00000000-6_list.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2037: .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 .LFE2037: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2060: .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 .LFE2060: .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.
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
.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.
#ifndef _LIST_CU_CUDA #define _LIST_CU_CUDA template <typename T> struct Node { T data; Node *next; }; template <typename T> class List { private: Node<T> *head; int count = 0; public: __device__ __host__ List() { head = NULL; } __device__ __host__ bool isEmpty() { return (count == 0); } __device__ __host__ void push(T val) { Node<T> *n = new Node<T>(); n->data = val; n->next = head; head = n; count++; } __device__ __host__ T pop() { if (isEmpty()) { return NULL; } else { if (head) { T p = head->data; head = head->next; count--; return p; } } } __device__ __host__ bool search(T val) { Node<T> *temp = head; while (temp->next) { if (temp->data == val) return true; else temp = temp->next; } delete temp; return false; } __device__ __host__ int size() { return count; } __device__ __host__ T operator[](int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } __device__ __host__ T get_index(int index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; } /*__device__ __host__ T operator[](int & index) { int itr = 0; Node<T> *temp = head; while (itr != index && temp->next) { temp = temp->next; itr++; } return temp->data; }*/ }; #endif //_LIST_CU_CUDA
.text .file "list.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_00078055_00000000-6_list.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2037: .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 .LFE2037: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2060: .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 .LFE2060: .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 "list.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 device assembly.
#include <stdio.h> #include <iostream> #include <cuda.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) cudaMallocPitch(&d_a,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... cudaMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione cudaMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding cudaMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,cudaMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); cudaFree(d_a); cudaFree(d_b); cudaFree(d_c); }
code for sm_80 Function : _Z19calcolaProdPuntualePiS_S_iii .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 R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e280000002600 */ /*0020*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e280000002200 */ /*0030*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e680000002100 */ /*0040*/ S2R R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e620000002500 */ /*0050*/ IMAD R3, R3, c[0x0][0x4], R2 ; /* 0x0000010003037a24 */ /* 0x001fca00078e0202 */ /*0060*/ ISETP.GE.AND P0, PT, R3, c[0x0][0x17c], PT ; /* 0x00005f0003007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R0, R5, c[0x0][0x0], R0 ; /* 0x0000000005007a24 */ /* 0x002fca00078e0200 */ /*0080*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x178], P0 ; /* 0x00005e0000007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*00b0*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*00c0*/ USHF.R.S32.HI UR4, URZ, 0x2, UR4 ; /* 0x000000023f047899 */ /* 0x000fcc0008011404 */ /*00d0*/ IMAD R0, R0, UR4, R3 ; /* 0x0000000400007c24 */ /* 0x000fe2000f8e0203 */ /*00e0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*00f0*/ IMAD.WIDE R4, R0, R7, c[0x0][0x168] ; /* 0x00005a0000047625 */ /* 0x000fc800078e0207 */ /*0100*/ IMAD.WIDE R2, R0.reuse, R7.reuse, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x0c0fe400078e0207 */ /*0110*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0130*/ IMAD.WIDE R6, R0, R7, c[0x0][0x170] ; /* 0x00005c0000067625 */ /* 0x000fc800078e0207 */ /*0140*/ IMAD R9, R4, R3, RZ ; /* 0x0000000304097224 */ /* 0x004fca00078e02ff */ /*0150*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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> #include <iostream> #include <cuda.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) cudaMallocPitch(&d_a,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... cudaMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione cudaMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding cudaMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,cudaMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); cudaFree(d_a); cudaFree(d_b); cudaFree(d_c); }
.file "tmpxft_00003f3a_00000000-6_eserc10.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3674: .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 .LFE3674: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z14inizializzaCPUPiii .type _Z14inizializzaCPUPiii, @function _Z14inizializzaCPUPiii: .LFB3669: .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 $24, %rsp .cfi_def_cfa_offset 80 movq %rdi, %r15 movl %esi, %ebx movl %esi, 4(%rsp) movl %edx, %r14d movl $0, %edi call time@PLT movl %eax, %edi call srand@PLT testl %ebx, %ebx jle .L3 movl $0, %r13d movl $0, %r12d movslq %r14d, %rax movq %rax, 8(%rsp) jmp .L5 .L7: movslq %r13d, %rax leaq (%r15,%rax,4), %rbx movq 8(%rsp), %rsi addq %rsi, %rax leaq (%r15,%rax,4), %rbp .L6: call rand@PLT movslq %eax, %rdx imulq $1717986919, %rdx, %rdx sarq $34, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx leal (%rdx,%rdx,4), %edx addl %edx, %edx subl %edx, %eax addl $1, %eax movl %eax, (%rbx) addq $4, %rbx cmpq %rbp, %rbx jne .L6 .L8: addl $1, %r12d addl %r14d, %r13d cmpl %r12d, 4(%rsp) je .L3 .L5: testl %r14d, %r14d jg .L7 jmp .L8 .L3: addq $24, %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 .cfi_endproc .LFE3669: .size _Z14inizializzaCPUPiii, .-_Z14inizializzaCPUPiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "--------------------------------" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string " " .text .globl _Z9stampaCPUPiii .type _Z9stampaCPUPiii, @function _Z9stampaCPUPiii: .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 $40, %rsp .cfi_def_cfa_offset 96 movq %rdi, 16(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movl $32, %edx leaq .LC0(%rip), %rsi leaq _ZSt4cout(%rip), %rbx movq %rbx, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %rbx testq %rbx, %rbx je .L25 cmpb $0, 56(%rbx) je .L13 movzbl 67(%rbx), %eax .L14: movsbl %al, %esi leaq _ZSt4cout(%rip), %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT cmpl $0, 12(%rsp) jle .L11 movl $0, %r15d movl $0, %r14d movslq 8(%rsp), %rax movq %rax, 24(%rsp) leaq _ZSt4cout(%rip), %rbp leaq .LC1(%rip), %r13 jmp .L16 .L25: call _ZSt16__throw_bad_castv@PLT .L13: movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) jmp .L14 .L26: call _ZSt16__throw_bad_castv@PLT .L27: movzbl 67(%rbx), %esi .L20: movsbl %sil, %esi movq %rbp, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addl $1, %r14d movl 8(%rsp), %eax addl %eax, %r15d cmpl %r14d, 12(%rsp) je .L11 .L16: cmpl $0, 8(%rsp) jle .L22 movslq %r15d, %rax movq 16(%rsp), %rcx leaq (%rcx,%rax,4), %rbx movq 24(%rsp), %rdx addq %rdx, %rax leaq (%rcx,%rax,4), %r12 .L17: movl (%rbx), %esi movq %rbp, %rdi call _ZNSolsEi@PLT movq %rax, %rdi movl $1, %edx movq %r13, %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $4, %rbx cmpq %r12, %rbx jne .L17 .L22: movq 0(%rbp), %rax movq -24(%rax), %rax movq 240(%rbp,%rax), %rbx testq %rbx, %rbx je .L26 cmpb $0, 56(%rbx) jne .L27 movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) movl %eax, %esi jmp .L20 .L11: addq $40, %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 .cfi_endproc .LFE3670: .size _Z9stampaCPUPiii, .-_Z9stampaCPUPiii .globl _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii .type _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii, @function _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii: .LFB3696: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%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 20(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%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 .L32 .L28: movq 168(%rsp), %rax subq %fs:40, %rax jne .L33 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L32: .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 _Z19calcolaProdPuntualePiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L28 .L33: call __stack_chk_fail@PLT .cfi_endproc .LFE3696: .size _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii, .-_Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii .globl _Z19calcolaProdPuntualePiS_S_iii .type _Z19calcolaProdPuntualePiS_S_iii, @function _Z19calcolaProdPuntualePiS_S_iii: .LFB3697: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3697: .size _Z19calcolaProdPuntualePiS_S_iii, .-_Z19calcolaProdPuntualePiS_S_iii .section .rodata.str1.1 .LC2: .string "%d" .text .globl main .type main, @function main: .LFB3671: .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 $80, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 68(%rsp) cmpl $5, %edi je .L37 movl $5, 8(%rsp) movl $5, 12(%rsp) movl $2, 52(%rsp) movl $2, 60(%rsp) movl $2, %ecx movl $5, %r14d .L38: addl $1, %ecx movl %ecx, 48(%rsp) .L39: movl 12(%rsp), %r13d movl 52(%rsp), %ecx movl %r13d, %eax movl $0, %edx divl %ecx movl %eax, 64(%rsp) testl %edx, %edx je .L40 addl $1, %ecx movl %ecx, 52(%rsp) .L40: movl %r13d, %ebx imull %r14d, %ebx movslq %ebx, %rbx salq $2, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r12 movq %rbx, %rdi call malloc@PLT movq %rax, %rbp movl %r13d, %edx movl %r14d, %esi movq %r12, %rdi call _Z14inizializzaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %rbp, %rdi call _Z14inizializzaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %r12, %rdi call _Z9stampaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %rbp, %rdi call _Z9stampaCPUPiii movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 40(%rsp), %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 32(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 12(%rsp), %rcx salq $2, %rcx subq $8, %rsp .cfi_def_cfa_offset 136 pushq $1 .cfi_def_cfa_offset 144 movslq 24(%rsp), %r9 movq %rcx, %r8 movq %r12, %rdx movq 56(%rsp), %rsi movq 32(%rsp), %rdi call cudaMemcpy2D@PLT movslq 28(%rsp), %rcx salq $2, %rcx movl $1, (%rsp) movslq 24(%rsp), %r9 movq %rcx, %r8 movq %rbp, %rdx movq 56(%rsp), %rsi movq 40(%rsp), %rdi call cudaMemcpy2D@PLT addq $16, %rsp .cfi_def_cfa_offset 128 movl 56(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 48(%rsp), %rdx movq 60(%rsp), %rdi movl 68(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L44 .L41: movl 12(%rsp), %ebx movl 8(%rsp), %r14d movl %ebx, %edi imull %r14d, %edi movslq %edi, %rdi salq $2, %rdi call malloc@PLT movq %rax, %r13 movslq %ebx, %rsi salq $2, %rsi subq $8, %rsp .cfi_def_cfa_offset 136 pushq $2 .cfi_def_cfa_offset 144 movslq %r14d, %r9 movq %rsi, %r8 movq 56(%rsp), %rcx movq 48(%rsp), %rdx movq %rax, %rdi call cudaMemcpy2D@PLT addq $16, %rsp .cfi_def_cfa_offset 128 movl 12(%rsp), %edx movl 8(%rsp), %esi movq %r13, %rdi call _Z9stampaCPUPiii movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@PLT movq 72(%rsp), %rax subq %fs:40, %rax jne .L45 movl $0, %eax addq $80, %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 .L37: .cfi_restore_state movq %rsi, %rbx leaq 8(%rsp), %rdx movq 8(%rsi), %rdi leaq .LC2(%rip), %rbp movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 12(%rsp), %rdx movq 16(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 48(%rsp), %rdx movq 24(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 52(%rsp), %rdx movq 32(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT movl 8(%rsp), %r14d movl 48(%rsp), %ecx movl %r14d, %eax movl $0, %edx divl %ecx movl %eax, 60(%rsp) testl %edx, %edx je .L39 jmp .L38 .L44: movl 40(%rsp), %r9d movl 12(%rsp), %r8d movl 8(%rsp), %ecx movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq 16(%rsp), %rdi call _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii jmp .L41 .L45: call __stack_chk_fail@PLT .cfi_endproc .LFE3671: .size main, .-main .section .rodata.str1.8 .align 8 .LC3: .string "_Z19calcolaProdPuntualePiS_S_iii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3699: .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z19calcolaProdPuntualePiS_S_iii(%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 .LFE3699: .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 <iostream> #include <cuda.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) cudaMallocPitch(&d_a,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); cudaMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... cudaMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione cudaMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,cudaMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding cudaMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,cudaMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); cudaFree(d_a); cudaFree(d_b); cudaFree(d_c); }
#include <stdio.h> #include <iostream> #include <hip/hip_runtime.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) hipMallocPitch(&d_a,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... hipMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione hipMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding hipMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,hipMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); 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 <stdio.h> #include <iostream> #include <hip/hip_runtime.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) hipMallocPitch(&d_a,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... hipMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione hipMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding hipMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,hipMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); hipFree(d_a); hipFree(d_b); hipFree(d_c); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z19calcolaProdPuntualePiS_S_iii .globl _Z19calcolaProdPuntualePiS_S_iii .p2align 8 .type _Z19calcolaProdPuntualePiS_S_iii,@function _Z19calcolaProdPuntualePiS_S_iii: s_clause 0x1 s_load_b32 s4, s[0:1], 0x34 s_load_b64 s[2:3], s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s5, s4, 0xffff s_lshr_b32 s4, s4, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s14, s5, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s4, v[3:4] v_cmp_gt_i32_e32 vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_gt_i32_e64 s2, s3, v1 s_and_b32 s2, vcc_lo, s2 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_2 s_clause 0x2 s_load_b32 s2, s[0:1], 0x20 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_ashr_i32 s2, s2, 2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, v0, s2, v[1:2] v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[2:3] v_add_co_u32 v2, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(0) v_mul_lo_u32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19calcolaProdPuntualePiS_S_iii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .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 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 _Z19calcolaProdPuntualePiS_S_iii, .Lfunc_end0-_Z19calcolaProdPuntualePiS_S_iii .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: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z19calcolaProdPuntualePiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19calcolaProdPuntualePiS_S_iii.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 <stdio.h> #include <iostream> #include <hip/hip_runtime.h> #include <time.h> using namespace std; //prodotto puntuale tra 2 matrici, usando il padding di memoria (pitch) //input size matrici (m,n) , dimensioni blocchi (righe di thread, colonne di thread) __host__ void inizializzaCPU(int *a,int m,int n){ srand((unsigned int)time(NULL)); for(int i=0;i<m;i++) for(int j=0;j<n;j++) a[i*n+j]=1+rand()%10; } __host__ void stampaCPU(int *a,int m,int n){ cout<<"--------------------------------"<<endl; for(int i=0;i<m;i++){ for(int j=0;j<n;j++) cout<<a[i*n+j]<<" "; cout<<endl; } } __global__ void calcolaProdPuntuale(int *a,int *b,int *c,int m,int n,int pitch){ //attenzione a come si accede alle matrici allocate in memoria gpu perchè c'e' il pitch int nColonneEffettive = pitch/sizeof(int); //perchè il pitch è la lunghezza effettiva delle righe matrici allocate in memoria gpu, ma in bytes int iRiga = threadIdx.x + blockIdx.x * blockDim.x; //indice riga del thread GLOBALMENTE alla griglia int iCol = threadIdx.y + blockIdx.y * blockDim.y ; //indice colonna del thread GLOBALMENTE alla griglia if(iRiga >= m || iCol >=n) return; c[iRiga * nColonneEffettive + iCol] = a[iRiga * nColonneEffettive + iCol] * b[iRiga * nColonneEffettive + iCol]; } int main(int argc,char *argv[]){ int m,n; dim3 dimBlocco; //default constructor, qui salviamo il numero di righe e di colonne (di thread) in un blocco dim3 dimGriglia; //qui salviamo il numero di righe e di colonne (di blocchi) della griglia if(argc!=5){ m=5; n=5; dimBlocco.x=2; //2 righe di thread in un blocco dimBlocco.y=2; //2 colonne di thread in un blocco //quindi 4 thread in totale per un blocco } else{ sscanf(argv[1],"%d",&m); sscanf(argv[2],"%d",&n); sscanf(argv[3],"%d",& dimBlocco.x); //n righe di thread in blocco sscanf(argv[4],"%d",& dimBlocco.y); //n di colonne di thread in blocco } dimGriglia.x = m / dimBlocco.x; if(m % dimBlocco.x != 0) dimBlocco.x ++; dimGriglia.y = n / dimBlocco.y; if(n % dimBlocco.y != 0) dimBlocco.y ++; //strutture dati su host int *h_a,*h_b,*h_c; //alloco h_a=(int *)malloc(n*m*sizeof(int)); h_b=(int *)malloc(n*m*sizeof(int)); h_c=(int *)malloc(n*m*sizeof(int)); //inizializzo le matrici inizializzaCPU(h_a,m,n); inizializzaCPU(h_b,m,n); stampaCPU(h_a,m,n); stampaCPU(h_b,m,n); //per le strutture dati sulla memoria gpu int *d_a,*d_b,*d_c; size_t pitch; //qui verrà salvato dalla cudaMallocPitch la lunghezza effettiva (paddata) in bytes //ATTENZIONE CHE IL PITCH DEVE ESSERE UNSIGNED LONG //alloco memoria sull'heap gpu usando il pitch (padding) hipMallocPitch(&d_a,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_b,&pitch,n*sizeof(int),m); hipMallocPitch((void **)&d_c,&pitch,n*sizeof(int),m); //ora in pitch c'e' la lunghezza effettiva (in bytes) delle righe //ora devo copiare il contenuto delle matrici dalla memoria host a memoria device //pero' la memoria device è paddata (pitch) quindi devo usare la cudaMemcpy2D per evitare //di inserire dati in quelle che sono le celle di padding, quindi... hipMemcpy2D(d_a,pitch,h_a,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //i parametri sono : destinazione, lunghezza effettiva in bytes delle righe in memoria device (quindi quest'informazione sta in pitch) //origine, lunghezza in bytes delle righe nella memoria host (potrebbe esserci il pitch anche qui) //numero di bytes da copiare nella direzione orizzontale (lunghezza bytes riga) //numero di righe da copiare //flag destinazione hipMemcpy2D(d_b,pitch,h_b,n*sizeof(int),n*sizeof(int),m,hipMemcpyHostToDevice); //lancio il kernel calcolaProdPuntuale<<<dimGriglia, dimBlocco>>>(d_a,d_b,d_c,m,n,pitch); //ricopio dalla memoria device alla memoria host int *copyFromGPU=(int *)malloc(n*m*sizeof(int)); //usando sempre la cudaMemcpy2D perchè nell'origine ci sono celle di padding hipMemcpy2D(copyFromGPU,n*sizeof(int),d_c,pitch,n*sizeof(int),m,hipMemcpyDeviceToHost); //stampo stampaCPU(copyFromGPU,m,n); //libero memoria free(h_a); free(h_b); free(h_c); hipFree(d_a); hipFree(d_b); hipFree(d_c); }
.text .file "eserc10.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z14inizializzaCPUPiii # -- Begin function _Z14inizializzaCPUPiii .p2align 4, 0x90 .type _Z14inizializzaCPUPiii,@function _Z14inizializzaCPUPiii: # @_Z14inizializzaCPUPiii .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 $24, %rsp .cfi_def_cfa_offset 80 .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 %edx, %ebx movl %esi, %ebp movq %rdi, 8(%rsp) # 8-byte Spill xorl %r15d, %r15d xorl %edi, %edi callq time movl %eax, %edi callq srand testl %ebp, %ebp jle .LBB0_6 # %bb.1: # %.preheader.lr.ph movl %ebp, %eax movq %rax, 16(%rsp) # 8-byte Spill movl %ebx, %r13d xorl %ebp, %ebp jmp .LBB0_2 .p2align 4, 0x90 .LBB0_5: # %._crit_edge # in Loop: Header=BB0_2 Depth=1 incq %rbp addl %ebx, %r15d cmpq 16(%rsp), %rbp # 8-byte Folded Reload je .LBB0_6 .LBB0_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_4 Depth 2 testl %ebx, %ebx jle .LBB0_5 # %bb.3: # %.lr.ph # in Loop: Header=BB0_2 Depth=1 movl %r15d, %eax movq 8(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r14 xorl %r12d, %r12d .p2align 4, 0x90 .LBB0_4: # Parent Loop BB0_2 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%r14,%r12,4) incq %r12 cmpq %r12, %r13 jne .LBB0_4 jmp .LBB0_5 .LBB0_6: # %._crit_edge13 addq $24, %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_end0: .size _Z14inizializzaCPUPiii, .Lfunc_end0-_Z14inizializzaCPUPiii .cfi_endproc # -- End function .globl _Z9stampaCPUPiii # -- Begin function _Z9stampaCPUPiii .p2align 4, 0x90 .type _Z9stampaCPUPiii,@function _Z9stampaCPUPiii: # @_Z9stampaCPUPiii .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 $24, %rsp .cfi_def_cfa_offset 80 .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 %edx, %ebx movl %esi, %ebp movq %rdi, 8(%rsp) # 8-byte Spill movl $_ZSt4cout, %edi movl $.L.str, %esi movl $32, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r15 testq %r15, %r15 je .LBB1_15 # %bb.1: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r15) je .LBB1_3 # %bb.2: movzbl 67(%r15), %eax jmp .LBB1_4 .LBB1_3: movq %r15, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) .LBB1_4: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv testl %ebp, %ebp jle .LBB1_14 # %bb.5: # %.preheader.lr.ph movl %ebp, %eax movq %rax, 16(%rsp) # 8-byte Spill movl %ebx, %r13d xorl %ebp, %ebp xorl %r14d, %r14d jmp .LBB1_6 .p2align 4, 0x90 .LBB1_12: # in Loop: Header=BB1_6 Depth=1 movq %r15, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) .LBB1_13: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit15 # in Loop: Header=BB1_6 Depth=1 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r14 addl %ebx, %ebp cmpq 16(%rsp), %r14 # 8-byte Folded Reload je .LBB1_14 .LBB1_6: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_8 Depth 2 testl %ebx, %ebx jle .LBB1_9 # %bb.7: # %.lr.ph # in Loop: Header=BB1_6 Depth=1 movl %ebp, %eax movq 8(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r15 xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_8: # Parent Loop BB1_6 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r12,4), %esi movl $_ZSt4cout, %edi callq _ZNSolsEi movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r12 cmpq %r12, %r13 jne .LBB1_8 .LBB1_9: # %._crit_edge # in Loop: Header=BB1_6 Depth=1 movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r15 testq %r15, %r15 je .LBB1_15 # %bb.10: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i12 # in Loop: Header=BB1_6 Depth=1 cmpb $0, 56(%r15) je .LBB1_12 # %bb.11: # in Loop: Header=BB1_6 Depth=1 movzbl 67(%r15), %eax jmp .LBB1_13 .LBB1_14: # %._crit_edge18 addq $24, %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 .LBB1_15: .cfi_def_cfa_offset 80 callq _ZSt16__throw_bad_castv .Lfunc_end1: .size _Z9stampaCPUPiii, .Lfunc_end1-_Z9stampaCPUPiii .cfi_endproc # -- End function .globl _Z34__device_stub__calcolaProdPuntualePiS_S_iii # -- Begin function _Z34__device_stub__calcolaProdPuntualePiS_S_iii .p2align 4, 0x90 .type _Z34__device_stub__calcolaProdPuntualePiS_S_iii,@function _Z34__device_stub__calcolaProdPuntualePiS_S_iii: # @_Z34__device_stub__calcolaProdPuntualePiS_S_iii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 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 20(%rsp), %rax movq %rax, 120(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%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 $_Z19calcolaProdPuntualePiS_S_iii, %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_end2: .size _Z34__device_stub__calcolaProdPuntualePiS_S_iii, .Lfunc_end2-_Z34__device_stub__calcolaProdPuntualePiS_S_iii .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 $248, %rsp .cfi_def_cfa_offset 304 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movabsq $4294967297, %rax # imm = 0x100000001 movq %rax, 24(%rsp) movl $1, 32(%rsp) cmpl $5, %edi jne .LBB3_1 # %bb.2: leaq 28(%rsp), %rbx movq 8(%rsi), %rdi leaq 20(%rsp), %rdx movq %rsi, %r14 movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 16(%r14), %rdi leaq 16(%rsp), %rdx movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 24(%r14), %rdi leaq 24(%rsp), %rdx movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 32(%r14), %rdi movl $.L.str.2, %esi movq %rbx, %rdx xorl %eax, %eax callq __isoc23_sscanf jmp .LBB3_3 .LBB3_1: movl $5, 20(%rsp) movl $5, 16(%rsp) movabsq $8589934594, %rax # imm = 0x200000002 movq %rax, 24(%rsp) .LBB3_3: movl 20(%rsp), %r12d movl 24(%rsp), %ecx movl %r12d, %eax xorl %edx, %edx divl %ecx # kill: def $eax killed $eax def $rax movq %rax, 112(%rsp) # 8-byte Spill testl %edx, %edx je .LBB3_5 # %bb.4: incl %ecx movl %ecx, 24(%rsp) .LBB3_5: movl 16(%rsp), %ebp movl 28(%rsp), %ecx movl %ebp, %eax xorl %edx, %edx divl %ecx # kill: def $eax killed $eax def $rax testl %edx, %edx je .LBB3_7 # %bb.6: incl %ecx movl %ecx, 28(%rsp) .LBB3_7: shlq $32, %rax movq %rax, 88(%rsp) # 8-byte Spill movl %ebp, %eax imull %r12d, %eax movslq %eax, %r14 shlq $2, %r14 movq %r14, %rdi callq malloc movq %rax, 104(%rsp) # 8-byte Spill movq %r14, %rdi callq malloc movq %rax, 96(%rsp) # 8-byte Spill xorl %r13d, %r13d xorl %edi, %edi callq time movl %eax, %edi callq srand testl %r12d, %r12d jle .LBB3_13 # %bb.8: # %.preheader.lr.ph.i xorl %r15d, %r15d jmp .LBB3_9 .p2align 4, 0x90 .LBB3_12: # %._crit_edge.i # in Loop: Header=BB3_9 Depth=1 incq %r15 addl %ebp, %r13d cmpq %r12, %r15 je .LBB3_13 .LBB3_9: # %.preheader.i # =>This Loop Header: Depth=1 # Child Loop BB3_11 Depth 2 testl %ebp, %ebp jle .LBB3_12 # %bb.10: # %.lr.ph.i # in Loop: Header=BB3_9 Depth=1 movl %r13d, %eax movq 104(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %rbx xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_11: # Parent Loop BB3_9 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%rbx,%r14,4) incq %r14 cmpq %r14, %rbp jne .LBB3_11 jmp .LBB3_12 .LBB3_13: # %_Z14inizializzaCPUPiii.exit movq 112(%rsp), %rax # 8-byte Reload addq %rax, 88(%rsp) # 8-byte Folded Spill movl 20(%rsp), %r12d movl 16(%rsp), %r13d xorl %ebp, %ebp xorl %edi, %edi callq time movl %eax, %edi callq srand testl %r12d, %r12d jle .LBB3_19 # %bb.14: # %.preheader.lr.ph.i25 xorl %r15d, %r15d jmp .LBB3_15 .p2align 4, 0x90 .LBB3_18: # %._crit_edge.i29 # in Loop: Header=BB3_15 Depth=1 incq %r15 addl %r13d, %ebp cmpq %r12, %r15 je .LBB3_19 .LBB3_15: # %.preheader.i27 # =>This Loop Header: Depth=1 # Child Loop BB3_17 Depth 2 testl %r13d, %r13d jle .LBB3_18 # %bb.16: # %.lr.ph.i32 # in Loop: Header=BB3_15 Depth=1 movl %ebp, %eax movq 96(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %rbx xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_17: # Parent Loop BB3_15 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%rbx,%r14,4) incq %r14 cmpq %r14, %r13 jne .LBB3_17 jmp .LBB3_18 .LBB3_19: # %_Z14inizializzaCPUPiii.exit37 movl 20(%rsp), %esi movl 16(%rsp), %edx movq 104(%rsp), %rbx # 8-byte Reload movq %rbx, %rdi callq _Z9stampaCPUPiii movl 20(%rsp), %esi movl 16(%rsp), %edx movq 96(%rsp), %r14 # 8-byte Reload movq %r14, %rdi callq _Z9stampaCPUPiii movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 64(%rsp), %rdi leaq 40(%rsp), %r12 movq %r12, %rsi callq hipMallocPitch movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 56(%rsp), %rdi movq %r12, %rsi callq hipMallocPitch movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 48(%rsp), %rdi movq %r12, %rsi callq hipMallocPitch movq 64(%rsp), %rdi movq 40(%rsp), %rsi movslq 16(%rsp), %rcx shlq $2, %rcx movslq 20(%rsp), %r9 movl $1, (%rsp) movq %rbx, %rdx movq %rcx, %r8 callq hipMemcpy2D movq 56(%rsp), %rdi movq 40(%rsp), %rsi movslq 16(%rsp), %rcx shlq $2, %rcx movslq 20(%rsp), %r9 movl $1, (%rsp) movq %r14, %rdx movq %rcx, %r8 callq hipMemcpy2D movq 24(%rsp), %rdx movl 32(%rsp), %ecx movq 88(%rsp), %rdi # 8-byte Reload movl $1, %esi xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_21 # %bb.20: movq 64(%rsp), %rax movq 56(%rsp), %rcx movq 48(%rsp), %rdx movl 20(%rsp), %esi movl 16(%rsp), %edi movl 40(%rsp), %r8d movq %rax, 184(%rsp) movq %rcx, 176(%rsp) movq %rdx, 168(%rsp) movl %esi, 84(%rsp) movl %edi, 80(%rsp) movl %r8d, 76(%rsp) leaq 184(%rsp), %rax movq %rax, 192(%rsp) leaq 176(%rsp), %rax movq %rax, 200(%rsp) leaq 168(%rsp), %rax movq %rax, 208(%rsp) leaq 84(%rsp), %rax movq %rax, 216(%rsp) leaq 80(%rsp), %rax movq %rax, 224(%rsp) leaq 76(%rsp), %rax movq %rax, 232(%rsp) leaq 152(%rsp), %rdi leaq 136(%rsp), %rsi leaq 128(%rsp), %rdx leaq 120(%rsp), %rcx callq __hipPopCallConfiguration movq 128(%rsp), %rax movq 120(%rsp), %rdi movq 152(%rsp), %rsi movl 160(%rsp), %edx movq 136(%rsp), %rcx movl 144(%rsp), %r8d movq %rdi, 8(%rsp) movq %rax, (%rsp) leaq 192(%rsp), %r9 movl $_Z19calcolaProdPuntualePiS_S_iii, %edi callq hipLaunchKernel .LBB3_21: movslq 16(%rsp), %r15 movslq 20(%rsp), %r12 movq %r12, %rdi imulq %r15, %rdi shlq $2, %rdi callq malloc movq %rax, %r13 shlq $2, %r15 movq 48(%rsp), %rdx movq 40(%rsp), %rcx movl $2, (%rsp) movq %rax, %rdi movq %r15, %rsi movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D movl 20(%rsp), %esi movl 16(%rsp), %edx movq %r13, %rdi callq _Z9stampaCPUPiii movq %rbx, %rdi callq free movq %r14, %rdi callq free movq 64(%rsp), %rdi callq hipFree movq 56(%rsp), %rdi callq hipFree movq 48(%rsp), %rdi callq hipFree xorl %eax, %eax addq $248, %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_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 $_Z19calcolaProdPuntualePiS_S_iii, %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 .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "--------------------------------" .size .L.str, 33 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz " " .size .L.str.1, 2 .type _Z19calcolaProdPuntualePiS_S_iii,@object # @_Z19calcolaProdPuntualePiS_S_iii .section .rodata,"a",@progbits .globl _Z19calcolaProdPuntualePiS_S_iii .p2align 3, 0x0 _Z19calcolaProdPuntualePiS_S_iii: .quad _Z34__device_stub__calcolaProdPuntualePiS_S_iii .size _Z19calcolaProdPuntualePiS_S_iii, 8 .type .L.str.2,@object # @.str.2 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.2: .asciz "%d" .size .L.str.2, 3 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19calcolaProdPuntualePiS_S_iii" .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 _Z34__device_stub__calcolaProdPuntualePiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _ZSt4cout .addrsig_sym _Z19calcolaProdPuntualePiS_S_iii .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 : _Z19calcolaProdPuntualePiS_S_iii .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 R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e280000002600 */ /*0020*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e280000002200 */ /*0030*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e680000002100 */ /*0040*/ S2R R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e620000002500 */ /*0050*/ IMAD R3, R3, c[0x0][0x4], R2 ; /* 0x0000010003037a24 */ /* 0x001fca00078e0202 */ /*0060*/ ISETP.GE.AND P0, PT, R3, c[0x0][0x17c], PT ; /* 0x00005f0003007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R0, R5, c[0x0][0x0], R0 ; /* 0x0000000005007a24 */ /* 0x002fca00078e0200 */ /*0080*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x178], P0 ; /* 0x00005e0000007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*00b0*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*00c0*/ USHF.R.S32.HI UR4, URZ, 0x2, UR4 ; /* 0x000000023f047899 */ /* 0x000fcc0008011404 */ /*00d0*/ IMAD R0, R0, UR4, R3 ; /* 0x0000000400007c24 */ /* 0x000fe2000f8e0203 */ /*00e0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*00f0*/ IMAD.WIDE R4, R0, R7, c[0x0][0x168] ; /* 0x00005a0000047625 */ /* 0x000fc800078e0207 */ /*0100*/ IMAD.WIDE R2, R0.reuse, R7.reuse, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x0c0fe400078e0207 */ /*0110*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0130*/ IMAD.WIDE R6, R0, R7, c[0x0][0x170] ; /* 0x00005c0000067625 */ /* 0x000fc800078e0207 */ /*0140*/ IMAD R9, R4, R3, RZ ; /* 0x0000000304097224 */ /* 0x004fca00078e02ff */ /*0150*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 _Z19calcolaProdPuntualePiS_S_iii .globl _Z19calcolaProdPuntualePiS_S_iii .p2align 8 .type _Z19calcolaProdPuntualePiS_S_iii,@function _Z19calcolaProdPuntualePiS_S_iii: s_clause 0x1 s_load_b32 s4, s[0:1], 0x34 s_load_b64 s[2:3], s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s5, s4, 0xffff s_lshr_b32 s4, s4, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s14, s5, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s4, v[3:4] v_cmp_gt_i32_e32 vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_gt_i32_e64 s2, s3, v1 s_and_b32 s2, vcc_lo, s2 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_2 s_clause 0x2 s_load_b32 s2, s[0:1], 0x20 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_ashr_i32 s2, s2, 2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, v0, s2, v[1:2] v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[2:3] v_add_co_u32 v2, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(0) v_mul_lo_u32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19calcolaProdPuntualePiS_S_iii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .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 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 _Z19calcolaProdPuntualePiS_S_iii, .Lfunc_end0-_Z19calcolaProdPuntualePiS_S_iii .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: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z19calcolaProdPuntualePiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19calcolaProdPuntualePiS_S_iii.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_00003f3a_00000000-6_eserc10.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3674: .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 .LFE3674: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z14inizializzaCPUPiii .type _Z14inizializzaCPUPiii, @function _Z14inizializzaCPUPiii: .LFB3669: .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 $24, %rsp .cfi_def_cfa_offset 80 movq %rdi, %r15 movl %esi, %ebx movl %esi, 4(%rsp) movl %edx, %r14d movl $0, %edi call time@PLT movl %eax, %edi call srand@PLT testl %ebx, %ebx jle .L3 movl $0, %r13d movl $0, %r12d movslq %r14d, %rax movq %rax, 8(%rsp) jmp .L5 .L7: movslq %r13d, %rax leaq (%r15,%rax,4), %rbx movq 8(%rsp), %rsi addq %rsi, %rax leaq (%r15,%rax,4), %rbp .L6: call rand@PLT movslq %eax, %rdx imulq $1717986919, %rdx, %rdx sarq $34, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx leal (%rdx,%rdx,4), %edx addl %edx, %edx subl %edx, %eax addl $1, %eax movl %eax, (%rbx) addq $4, %rbx cmpq %rbp, %rbx jne .L6 .L8: addl $1, %r12d addl %r14d, %r13d cmpl %r12d, 4(%rsp) je .L3 .L5: testl %r14d, %r14d jg .L7 jmp .L8 .L3: addq $24, %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 .cfi_endproc .LFE3669: .size _Z14inizializzaCPUPiii, .-_Z14inizializzaCPUPiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "--------------------------------" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string " " .text .globl _Z9stampaCPUPiii .type _Z9stampaCPUPiii, @function _Z9stampaCPUPiii: .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 $40, %rsp .cfi_def_cfa_offset 96 movq %rdi, 16(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movl $32, %edx leaq .LC0(%rip), %rsi leaq _ZSt4cout(%rip), %rbx movq %rbx, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %rbx testq %rbx, %rbx je .L25 cmpb $0, 56(%rbx) je .L13 movzbl 67(%rbx), %eax .L14: movsbl %al, %esi leaq _ZSt4cout(%rip), %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT cmpl $0, 12(%rsp) jle .L11 movl $0, %r15d movl $0, %r14d movslq 8(%rsp), %rax movq %rax, 24(%rsp) leaq _ZSt4cout(%rip), %rbp leaq .LC1(%rip), %r13 jmp .L16 .L25: call _ZSt16__throw_bad_castv@PLT .L13: movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) jmp .L14 .L26: call _ZSt16__throw_bad_castv@PLT .L27: movzbl 67(%rbx), %esi .L20: movsbl %sil, %esi movq %rbp, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addl $1, %r14d movl 8(%rsp), %eax addl %eax, %r15d cmpl %r14d, 12(%rsp) je .L11 .L16: cmpl $0, 8(%rsp) jle .L22 movslq %r15d, %rax movq 16(%rsp), %rcx leaq (%rcx,%rax,4), %rbx movq 24(%rsp), %rdx addq %rdx, %rax leaq (%rcx,%rax,4), %r12 .L17: movl (%rbx), %esi movq %rbp, %rdi call _ZNSolsEi@PLT movq %rax, %rdi movl $1, %edx movq %r13, %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $4, %rbx cmpq %r12, %rbx jne .L17 .L22: movq 0(%rbp), %rax movq -24(%rax), %rax movq 240(%rbp,%rax), %rbx testq %rbx, %rbx je .L26 cmpb $0, 56(%rbx) jne .L27 movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) movl %eax, %esi jmp .L20 .L11: addq $40, %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 .cfi_endproc .LFE3670: .size _Z9stampaCPUPiii, .-_Z9stampaCPUPiii .globl _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii .type _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii, @function _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii: .LFB3696: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%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 20(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%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 .L32 .L28: movq 168(%rsp), %rax subq %fs:40, %rax jne .L33 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L32: .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 _Z19calcolaProdPuntualePiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L28 .L33: call __stack_chk_fail@PLT .cfi_endproc .LFE3696: .size _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii, .-_Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii .globl _Z19calcolaProdPuntualePiS_S_iii .type _Z19calcolaProdPuntualePiS_S_iii, @function _Z19calcolaProdPuntualePiS_S_iii: .LFB3697: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3697: .size _Z19calcolaProdPuntualePiS_S_iii, .-_Z19calcolaProdPuntualePiS_S_iii .section .rodata.str1.1 .LC2: .string "%d" .text .globl main .type main, @function main: .LFB3671: .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 $80, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 68(%rsp) cmpl $5, %edi je .L37 movl $5, 8(%rsp) movl $5, 12(%rsp) movl $2, 52(%rsp) movl $2, 60(%rsp) movl $2, %ecx movl $5, %r14d .L38: addl $1, %ecx movl %ecx, 48(%rsp) .L39: movl 12(%rsp), %r13d movl 52(%rsp), %ecx movl %r13d, %eax movl $0, %edx divl %ecx movl %eax, 64(%rsp) testl %edx, %edx je .L40 addl $1, %ecx movl %ecx, 52(%rsp) .L40: movl %r13d, %ebx imull %r14d, %ebx movslq %ebx, %rbx salq $2, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r12 movq %rbx, %rdi call malloc@PLT movq %rax, %rbp movl %r13d, %edx movl %r14d, %esi movq %r12, %rdi call _Z14inizializzaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %rbp, %rdi call _Z14inizializzaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %r12, %rdi call _Z9stampaCPUPiii movl 12(%rsp), %edx movl 8(%rsp), %esi movq %rbp, %rdi call _Z9stampaCPUPiii movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 40(%rsp), %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 8(%rsp), %rcx movslq 12(%rsp), %rdx salq $2, %rdx leaq 32(%rsp), %rdi movq %rbx, %rsi call cudaMallocPitch@PLT movslq 12(%rsp), %rcx salq $2, %rcx subq $8, %rsp .cfi_def_cfa_offset 136 pushq $1 .cfi_def_cfa_offset 144 movslq 24(%rsp), %r9 movq %rcx, %r8 movq %r12, %rdx movq 56(%rsp), %rsi movq 32(%rsp), %rdi call cudaMemcpy2D@PLT movslq 28(%rsp), %rcx salq $2, %rcx movl $1, (%rsp) movslq 24(%rsp), %r9 movq %rcx, %r8 movq %rbp, %rdx movq 56(%rsp), %rsi movq 40(%rsp), %rdi call cudaMemcpy2D@PLT addq $16, %rsp .cfi_def_cfa_offset 128 movl 56(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 48(%rsp), %rdx movq 60(%rsp), %rdi movl 68(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L44 .L41: movl 12(%rsp), %ebx movl 8(%rsp), %r14d movl %ebx, %edi imull %r14d, %edi movslq %edi, %rdi salq $2, %rdi call malloc@PLT movq %rax, %r13 movslq %ebx, %rsi salq $2, %rsi subq $8, %rsp .cfi_def_cfa_offset 136 pushq $2 .cfi_def_cfa_offset 144 movslq %r14d, %r9 movq %rsi, %r8 movq 56(%rsp), %rcx movq 48(%rsp), %rdx movq %rax, %rdi call cudaMemcpy2D@PLT addq $16, %rsp .cfi_def_cfa_offset 128 movl 12(%rsp), %edx movl 8(%rsp), %esi movq %r13, %rdi call _Z9stampaCPUPiii movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@PLT movq 72(%rsp), %rax subq %fs:40, %rax jne .L45 movl $0, %eax addq $80, %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 .L37: .cfi_restore_state movq %rsi, %rbx leaq 8(%rsp), %rdx movq 8(%rsi), %rdi leaq .LC2(%rip), %rbp movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 12(%rsp), %rdx movq 16(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 48(%rsp), %rdx movq 24(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT leaq 52(%rsp), %rdx movq 32(%rbx), %rdi movq %rbp, %rsi movl $0, %eax call __isoc23_sscanf@PLT movl 8(%rsp), %r14d movl 48(%rsp), %ecx movl %r14d, %eax movl $0, %edx divl %ecx movl %eax, 60(%rsp) testl %edx, %edx je .L39 jmp .L38 .L44: movl 40(%rsp), %r9d movl 12(%rsp), %r8d movl 8(%rsp), %ecx movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq 16(%rsp), %rdi call _Z46__device_stub__Z19calcolaProdPuntualePiS_S_iiiPiS_S_iii jmp .L41 .L45: call __stack_chk_fail@PLT .cfi_endproc .LFE3671: .size main, .-main .section .rodata.str1.8 .align 8 .LC3: .string "_Z19calcolaProdPuntualePiS_S_iii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3699: .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z19calcolaProdPuntualePiS_S_iii(%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 .LFE3699: .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 "eserc10.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z14inizializzaCPUPiii # -- Begin function _Z14inizializzaCPUPiii .p2align 4, 0x90 .type _Z14inizializzaCPUPiii,@function _Z14inizializzaCPUPiii: # @_Z14inizializzaCPUPiii .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 $24, %rsp .cfi_def_cfa_offset 80 .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 %edx, %ebx movl %esi, %ebp movq %rdi, 8(%rsp) # 8-byte Spill xorl %r15d, %r15d xorl %edi, %edi callq time movl %eax, %edi callq srand testl %ebp, %ebp jle .LBB0_6 # %bb.1: # %.preheader.lr.ph movl %ebp, %eax movq %rax, 16(%rsp) # 8-byte Spill movl %ebx, %r13d xorl %ebp, %ebp jmp .LBB0_2 .p2align 4, 0x90 .LBB0_5: # %._crit_edge # in Loop: Header=BB0_2 Depth=1 incq %rbp addl %ebx, %r15d cmpq 16(%rsp), %rbp # 8-byte Folded Reload je .LBB0_6 .LBB0_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_4 Depth 2 testl %ebx, %ebx jle .LBB0_5 # %bb.3: # %.lr.ph # in Loop: Header=BB0_2 Depth=1 movl %r15d, %eax movq 8(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r14 xorl %r12d, %r12d .p2align 4, 0x90 .LBB0_4: # Parent Loop BB0_2 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%r14,%r12,4) incq %r12 cmpq %r12, %r13 jne .LBB0_4 jmp .LBB0_5 .LBB0_6: # %._crit_edge13 addq $24, %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_end0: .size _Z14inizializzaCPUPiii, .Lfunc_end0-_Z14inizializzaCPUPiii .cfi_endproc # -- End function .globl _Z9stampaCPUPiii # -- Begin function _Z9stampaCPUPiii .p2align 4, 0x90 .type _Z9stampaCPUPiii,@function _Z9stampaCPUPiii: # @_Z9stampaCPUPiii .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 $24, %rsp .cfi_def_cfa_offset 80 .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 %edx, %ebx movl %esi, %ebp movq %rdi, 8(%rsp) # 8-byte Spill movl $_ZSt4cout, %edi movl $.L.str, %esi movl $32, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r15 testq %r15, %r15 je .LBB1_15 # %bb.1: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r15) je .LBB1_3 # %bb.2: movzbl 67(%r15), %eax jmp .LBB1_4 .LBB1_3: movq %r15, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) .LBB1_4: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv testl %ebp, %ebp jle .LBB1_14 # %bb.5: # %.preheader.lr.ph movl %ebp, %eax movq %rax, 16(%rsp) # 8-byte Spill movl %ebx, %r13d xorl %ebp, %ebp xorl %r14d, %r14d jmp .LBB1_6 .p2align 4, 0x90 .LBB1_12: # in Loop: Header=BB1_6 Depth=1 movq %r15, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) .LBB1_13: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit15 # in Loop: Header=BB1_6 Depth=1 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r14 addl %ebx, %ebp cmpq 16(%rsp), %r14 # 8-byte Folded Reload je .LBB1_14 .LBB1_6: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_8 Depth 2 testl %ebx, %ebx jle .LBB1_9 # %bb.7: # %.lr.ph # in Loop: Header=BB1_6 Depth=1 movl %ebp, %eax movq 8(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r15 xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_8: # Parent Loop BB1_6 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r12,4), %esi movl $_ZSt4cout, %edi callq _ZNSolsEi movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r12 cmpq %r12, %r13 jne .LBB1_8 .LBB1_9: # %._crit_edge # in Loop: Header=BB1_6 Depth=1 movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r15 testq %r15, %r15 je .LBB1_15 # %bb.10: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i12 # in Loop: Header=BB1_6 Depth=1 cmpb $0, 56(%r15) je .LBB1_12 # %bb.11: # in Loop: Header=BB1_6 Depth=1 movzbl 67(%r15), %eax jmp .LBB1_13 .LBB1_14: # %._crit_edge18 addq $24, %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 .LBB1_15: .cfi_def_cfa_offset 80 callq _ZSt16__throw_bad_castv .Lfunc_end1: .size _Z9stampaCPUPiii, .Lfunc_end1-_Z9stampaCPUPiii .cfi_endproc # -- End function .globl _Z34__device_stub__calcolaProdPuntualePiS_S_iii # -- Begin function _Z34__device_stub__calcolaProdPuntualePiS_S_iii .p2align 4, 0x90 .type _Z34__device_stub__calcolaProdPuntualePiS_S_iii,@function _Z34__device_stub__calcolaProdPuntualePiS_S_iii: # @_Z34__device_stub__calcolaProdPuntualePiS_S_iii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 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 20(%rsp), %rax movq %rax, 120(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%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 $_Z19calcolaProdPuntualePiS_S_iii, %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_end2: .size _Z34__device_stub__calcolaProdPuntualePiS_S_iii, .Lfunc_end2-_Z34__device_stub__calcolaProdPuntualePiS_S_iii .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 $248, %rsp .cfi_def_cfa_offset 304 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movabsq $4294967297, %rax # imm = 0x100000001 movq %rax, 24(%rsp) movl $1, 32(%rsp) cmpl $5, %edi jne .LBB3_1 # %bb.2: leaq 28(%rsp), %rbx movq 8(%rsi), %rdi leaq 20(%rsp), %rdx movq %rsi, %r14 movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 16(%r14), %rdi leaq 16(%rsp), %rdx movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 24(%r14), %rdi leaq 24(%rsp), %rdx movl $.L.str.2, %esi xorl %eax, %eax callq __isoc23_sscanf movq 32(%r14), %rdi movl $.L.str.2, %esi movq %rbx, %rdx xorl %eax, %eax callq __isoc23_sscanf jmp .LBB3_3 .LBB3_1: movl $5, 20(%rsp) movl $5, 16(%rsp) movabsq $8589934594, %rax # imm = 0x200000002 movq %rax, 24(%rsp) .LBB3_3: movl 20(%rsp), %r12d movl 24(%rsp), %ecx movl %r12d, %eax xorl %edx, %edx divl %ecx # kill: def $eax killed $eax def $rax movq %rax, 112(%rsp) # 8-byte Spill testl %edx, %edx je .LBB3_5 # %bb.4: incl %ecx movl %ecx, 24(%rsp) .LBB3_5: movl 16(%rsp), %ebp movl 28(%rsp), %ecx movl %ebp, %eax xorl %edx, %edx divl %ecx # kill: def $eax killed $eax def $rax testl %edx, %edx je .LBB3_7 # %bb.6: incl %ecx movl %ecx, 28(%rsp) .LBB3_7: shlq $32, %rax movq %rax, 88(%rsp) # 8-byte Spill movl %ebp, %eax imull %r12d, %eax movslq %eax, %r14 shlq $2, %r14 movq %r14, %rdi callq malloc movq %rax, 104(%rsp) # 8-byte Spill movq %r14, %rdi callq malloc movq %rax, 96(%rsp) # 8-byte Spill xorl %r13d, %r13d xorl %edi, %edi callq time movl %eax, %edi callq srand testl %r12d, %r12d jle .LBB3_13 # %bb.8: # %.preheader.lr.ph.i xorl %r15d, %r15d jmp .LBB3_9 .p2align 4, 0x90 .LBB3_12: # %._crit_edge.i # in Loop: Header=BB3_9 Depth=1 incq %r15 addl %ebp, %r13d cmpq %r12, %r15 je .LBB3_13 .LBB3_9: # %.preheader.i # =>This Loop Header: Depth=1 # Child Loop BB3_11 Depth 2 testl %ebp, %ebp jle .LBB3_12 # %bb.10: # %.lr.ph.i # in Loop: Header=BB3_9 Depth=1 movl %r13d, %eax movq 104(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %rbx xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_11: # Parent Loop BB3_9 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%rbx,%r14,4) incq %r14 cmpq %r14, %rbp jne .LBB3_11 jmp .LBB3_12 .LBB3_13: # %_Z14inizializzaCPUPiii.exit movq 112(%rsp), %rax # 8-byte Reload addq %rax, 88(%rsp) # 8-byte Folded Spill movl 20(%rsp), %r12d movl 16(%rsp), %r13d xorl %ebp, %ebp xorl %edi, %edi callq time movl %eax, %edi callq srand testl %r12d, %r12d jle .LBB3_19 # %bb.14: # %.preheader.lr.ph.i25 xorl %r15d, %r15d jmp .LBB3_15 .p2align 4, 0x90 .LBB3_18: # %._crit_edge.i29 # in Loop: Header=BB3_15 Depth=1 incq %r15 addl %r13d, %ebp cmpq %r12, %r15 je .LBB3_19 .LBB3_15: # %.preheader.i27 # =>This Loop Header: Depth=1 # Child Loop BB3_17 Depth 2 testl %r13d, %r13d jle .LBB3_18 # %bb.16: # %.lr.ph.i32 # in Loop: Header=BB3_15 Depth=1 movl %ebp, %eax movq 96(%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %rbx xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_17: # Parent Loop BB3_15 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx negl %ecx addl %ecx, %eax incl %eax movl %eax, (%rbx,%r14,4) incq %r14 cmpq %r14, %r13 jne .LBB3_17 jmp .LBB3_18 .LBB3_19: # %_Z14inizializzaCPUPiii.exit37 movl 20(%rsp), %esi movl 16(%rsp), %edx movq 104(%rsp), %rbx # 8-byte Reload movq %rbx, %rdi callq _Z9stampaCPUPiii movl 20(%rsp), %esi movl 16(%rsp), %edx movq 96(%rsp), %r14 # 8-byte Reload movq %r14, %rdi callq _Z9stampaCPUPiii movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 64(%rsp), %rdi leaq 40(%rsp), %r12 movq %r12, %rsi callq hipMallocPitch movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 56(%rsp), %rdi movq %r12, %rsi callq hipMallocPitch movslq 16(%rsp), %rdx shlq $2, %rdx movslq 20(%rsp), %rcx leaq 48(%rsp), %rdi movq %r12, %rsi callq hipMallocPitch movq 64(%rsp), %rdi movq 40(%rsp), %rsi movslq 16(%rsp), %rcx shlq $2, %rcx movslq 20(%rsp), %r9 movl $1, (%rsp) movq %rbx, %rdx movq %rcx, %r8 callq hipMemcpy2D movq 56(%rsp), %rdi movq 40(%rsp), %rsi movslq 16(%rsp), %rcx shlq $2, %rcx movslq 20(%rsp), %r9 movl $1, (%rsp) movq %r14, %rdx movq %rcx, %r8 callq hipMemcpy2D movq 24(%rsp), %rdx movl 32(%rsp), %ecx movq 88(%rsp), %rdi # 8-byte Reload movl $1, %esi xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_21 # %bb.20: movq 64(%rsp), %rax movq 56(%rsp), %rcx movq 48(%rsp), %rdx movl 20(%rsp), %esi movl 16(%rsp), %edi movl 40(%rsp), %r8d movq %rax, 184(%rsp) movq %rcx, 176(%rsp) movq %rdx, 168(%rsp) movl %esi, 84(%rsp) movl %edi, 80(%rsp) movl %r8d, 76(%rsp) leaq 184(%rsp), %rax movq %rax, 192(%rsp) leaq 176(%rsp), %rax movq %rax, 200(%rsp) leaq 168(%rsp), %rax movq %rax, 208(%rsp) leaq 84(%rsp), %rax movq %rax, 216(%rsp) leaq 80(%rsp), %rax movq %rax, 224(%rsp) leaq 76(%rsp), %rax movq %rax, 232(%rsp) leaq 152(%rsp), %rdi leaq 136(%rsp), %rsi leaq 128(%rsp), %rdx leaq 120(%rsp), %rcx callq __hipPopCallConfiguration movq 128(%rsp), %rax movq 120(%rsp), %rdi movq 152(%rsp), %rsi movl 160(%rsp), %edx movq 136(%rsp), %rcx movl 144(%rsp), %r8d movq %rdi, 8(%rsp) movq %rax, (%rsp) leaq 192(%rsp), %r9 movl $_Z19calcolaProdPuntualePiS_S_iii, %edi callq hipLaunchKernel .LBB3_21: movslq 16(%rsp), %r15 movslq 20(%rsp), %r12 movq %r12, %rdi imulq %r15, %rdi shlq $2, %rdi callq malloc movq %rax, %r13 shlq $2, %r15 movq 48(%rsp), %rdx movq 40(%rsp), %rcx movl $2, (%rsp) movq %rax, %rdi movq %r15, %rsi movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D movl 20(%rsp), %esi movl 16(%rsp), %edx movq %r13, %rdi callq _Z9stampaCPUPiii movq %rbx, %rdi callq free movq %r14, %rdi callq free movq 64(%rsp), %rdi callq hipFree movq 56(%rsp), %rdi callq hipFree movq 48(%rsp), %rdi callq hipFree xorl %eax, %eax addq $248, %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_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 $_Z19calcolaProdPuntualePiS_S_iii, %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 .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "--------------------------------" .size .L.str, 33 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz " " .size .L.str.1, 2 .type _Z19calcolaProdPuntualePiS_S_iii,@object # @_Z19calcolaProdPuntualePiS_S_iii .section .rodata,"a",@progbits .globl _Z19calcolaProdPuntualePiS_S_iii .p2align 3, 0x0 _Z19calcolaProdPuntualePiS_S_iii: .quad _Z34__device_stub__calcolaProdPuntualePiS_S_iii .size _Z19calcolaProdPuntualePiS_S_iii, 8 .type .L.str.2,@object # @.str.2 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.2: .asciz "%d" .size .L.str.2, 3 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19calcolaProdPuntualePiS_S_iii" .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 _Z34__device_stub__calcolaProdPuntualePiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _ZSt4cout .addrsig_sym _Z19calcolaProdPuntualePiS_S_iii .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.
//////////////////////////////////////////////////////////////////////// // define kernel block size for //////////////////////////////////////////////////////////////////////// #define BLOCK_X 32 #define BLOCK_Y 8 // device code __global__ void GPU_adi_rhs(int NX, int NY, int NZ, float lam, const float* __restrict__ d_u, float* __restrict__ d_du, float* __restrict__ d_ax, float* __restrict__ d_bx, float* __restrict__ d_cx, float* __restrict__ d_ay, float* __restrict__ d_by, float* __restrict__ d_cy, float* __restrict__ d_az, float* __restrict__ d_bz, float* __restrict__ d_cz) { int i, j, k, indg, active; float du, a, b, c; int NXM1 = NX - 1; int NYM1 = NY - 1; int NZM1 = NZ - 1; #define IOFF 1 #define JOFF NX #define KOFF NX*NY // // set up indices for main block // i = threadIdx.x + blockIdx.x*BLOCK_X; j = threadIdx.y + blockIdx.y*BLOCK_Y; indg = i + j*NX; active = (i<NX) && (j<NY); // // loop over k-planes // for (k = 0; k<NZ; k++) { // // calculate r.h.s. and set a,b,c, coefficients // if (active) { if (i == 0 || i == NXM1 || j == 0 || j == NYM1 || k == 0 || k == NZM1) { du = 0.0f; // Dirichlet b.c.'s a = 0.0f; b = 1.0f; c = 0.0f; } else { du = lam * (d_u[indg - IOFF] + d_u[indg + IOFF] + d_u[indg - JOFF] + d_u[indg + JOFF] + d_u[indg - KOFF] + d_u[indg + KOFF] - 6.0f*d_u[indg]); a = -0.5f*lam; b = 1.0f + lam; c = -0.5f*lam; } d_du[indg] = du; d_ax[indg] = a; d_bx[indg] = b; d_cx[indg] = c; d_ay[indg] = a; d_by[indg] = b; d_cy[indg] = c; d_az[indg] = a; d_bz[indg] = b; d_cz[indg] = c; indg += KOFF; } } } __global__ void GPU_adi_x_float4(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; for (i = 4; i<NX; i += 4) { indg = indg + 1; a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; } // // reverse pass // d4.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4.x = dd; d_d[indg] = d4; for (i = NX - 5; i >= 0; i -= 4) { indg = indg - 1; dd = d2[i] - c2[i] * dd; d4.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4.x = dd; d_d[indg] = d4; } } } __global__ void GPU_adi_x_float4_2(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4, a4_2, b4_2, c4_2, d4_2; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; a4_2 = d_a[indg + 1]; b4 = d_b[indg]; b4_2 = d_b[indg + 1]; c4 = d_c[indg]; c4_2 = d_c[indg + 1]; d4 = d_d[indg]; d4_2 = d_d[indg + 1]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[4] = cc; d2[4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[5] = cc; d2[5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[6] = cc; d2[6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[7] = cc; d2[7] = dd; for (i = 8; i<NX; i += 8) { indg = indg + 2; a4 = d_a[indg]; a4_2 = d_a[indg + 1]; __threadfence_block(); b4 = d_b[indg]; b4_2 = d_b[indg + 1]; __threadfence_block(); c4 = d_c[indg]; c4_2 = d_c[indg + 1]; __threadfence_block(); d4 = d_d[indg]; d4_2 = d_d[indg + 1]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[i + 4] = cc; d2[i + 4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[i + 5] = cc; d2[i + 5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[i + 6] = cc; d2[i + 6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[i + 7] = cc; d2[i + 7] = dd; } // // reverse pass // d4_2.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4_2.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4_2.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4_2.x = dd; dd = d2[NX - 5] - c2[NX - 5] * dd; d4.w = dd; dd = d2[NX - 6] - c2[NX - 6] * dd; d4.z = dd; dd = d2[NX - 7] - c2[NX - 7] * dd; d4.y = dd; dd = d2[NX - 8] - c2[NX - 8] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; for (i = NX - 9; i >= 0; i -= 8) { indg = indg - 2; dd = d2[i] - c2[i] * dd; d4_2.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4_2.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4_2.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4_2.x = dd; dd = d2[i - 4] - c2[i - 4] * dd; d4.w = dd; dd = d2[i - 5] - c2[i - 5] * dd; d4.z = dd; dd = d2[i - 6] - c2[i - 6] * dd; d4.y = dd; dd = d2[i - 7] - c2[i - 7] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; } } } // // new tri-diagonal solve in x-direction // __global__ void GPU_adi_x_new(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int j, k, indg, t, tm, tp, nt, shift = 0; float bbi; __shared__ float a[256], c[256], d[256]; // // set up indices for main block // t = threadIdx.x; j = blockIdx.x; k = blockIdx.y; indg = t + NX*(j + k*NY); bbi = 1.0f / d_b[indg]; a[t] = -bbi * d_a[indg]; c[t] = -bbi * d_c[indg]; d[t] = bbi * d_d[indg]; // forward pass tm = 2 * t; t = tm + 1; tp = tm + 2; for (nt = blockDim.x / 2; nt>0; nt >>= 1) { shift++; __syncthreads(); if (threadIdx.x < nt) { bbi = 1.0f; if (tm >= 0) { bbi -= a[t] * c[tm]; d[t] += a[t] * d[tm]; a[t] = a[t] * a[tm]; } if (tp<NX) { bbi -= c[t] * a[tp]; d[t] += c[t] * d[tp]; c[t] = c[t] * c[tp]; } bbi = 1.0f / bbi; d[t] *= bbi; a[t] *= bbi; c[t] *= bbi; tm = 2 * tm + 1; t = 2 * t + 1; tp = 2 * tp + 1; } } // reverse pass for (; shift>0; shift--) { nt = blockDim.x >> shift; __syncthreads(); if (threadIdx.x < nt) { tm >>= 1; t >>= 1; tp >>= 1; if (tm >= 0) d[tm] += c[tm] * d[t]; if (tp<NX) d[tp] += a[tp] * d[t]; } } __syncthreads(); d_d[indg] = d[threadIdx.x]; } // // old tri-diagonal solve in x-direction // __global__ void GPU_adi_x(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY); if ((j<NY) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (i = 1; i<NX; i++) { indg = indg + 1; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[i] = cc; d[i] = dd; } // // reverse pass // d_d[indg] = dd; for (i = NX - 2; i >= 0; i--) { indg = indg - 1; dd = d[i] - c[i] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in y-direction // __global__ void GPU_adi_y(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = i + k*NX*NY; if ((i<NX) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (j = 1; j<NY; j++) { indg = indg + NX; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[j] = cc; d[j] = dd; } // // reverse pass // d_d[indg] = dd; for (j = NY - 2; j >= 0; j--) { indg = indg - NX; dd = d[j] - c[j] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in z-direction, and update solution // __global__ void GPU_adi_z(int NX, int NY, int NZ, float* __restrict__ d_u, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, const float* __restrict__ d_d) { int i, j, k, indg, off; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices j = threadIdx.y + blockIdx.y*blockDim.y; indg = i + j*NX; off = NX*NY; if ((i<NX) && (j<NY)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (k = 1; k<NZ; k++) { indg = indg + off; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[k] = cc; d[k] = dd; } // // reverse pass // d_u[indg] += dd; for (k = NZ - 2; k >= 0; k--) { indg = indg - off; dd = d[k] - c[k] * dd; d_u[indg] += dd; } } }
.file "tmpxft_0017205a_00000000-6_adi3d_kernel.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 _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .type _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, @function _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .LFB2051: .cfi_startproc endbr64 subq $312, %rsp .cfi_def_cfa_offset 320 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movss %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 296(%rsp) xorl %eax, %eax 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) movq %rcx, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 208(%rsp) movq %r8, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 216(%rsp) movq %r9, 40(%rsp) leaq 40(%rsp), %rax movq %rax, 224(%rsp) movq 320(%rsp), %rax movq %rax, 48(%rsp) leaq 48(%rsp), %rax movq %rax, 232(%rsp) movq 328(%rsp), %rax movq %rax, 56(%rsp) leaq 56(%rsp), %rax movq %rax, 240(%rsp) movq 336(%rsp), %rax movq %rax, 64(%rsp) leaq 64(%rsp), %rax movq %rax, 248(%rsp) movq 344(%rsp), %rax movq %rax, 72(%rsp) leaq 72(%rsp), %rax movq %rax, 256(%rsp) movq 352(%rsp), %rax movq %rax, 80(%rsp) leaq 80(%rsp), %rax movq %rax, 264(%rsp) movq 360(%rsp), %rax movq %rax, 88(%rsp) leaq 88(%rsp), %rax movq %rax, 272(%rsp) movq 368(%rsp), %rax movq %rax, 96(%rsp) leaq 96(%rsp), %rax movq %rax, 280(%rsp) movq 376(%rsp), %rax movq %rax, 104(%rsp) leaq 104(%rsp), %rax movq %rax, 288(%rsp) movl $1, 128(%rsp) movl $1, 132(%rsp) movl $1, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $1, 148(%rsp) leaq 120(%rsp), %rcx leaq 112(%rsp), %rdx leaq 140(%rsp), %rsi leaq 128(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 296(%rsp), %rax subq %fs:40, %rax jne .L8 addq $312, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 120(%rsp) .cfi_def_cfa_offset 328 pushq 120(%rsp) .cfi_def_cfa_offset 336 leaq 192(%rsp), %r9 movq 156(%rsp), %rcx movl 164(%rsp), %r8d movq 144(%rsp), %rsi movl 152(%rsp), %edx leaq _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 320 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .-_Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .globl _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .type _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, @function _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .LFB2052: .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 _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ addq $72, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .-_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .globl _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .type _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, @function _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_: .LFB2053: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L15 .L11: movq 168(%rsp), %rax subq %fs:40, %rax jne .L16 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, .-_Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .globl _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_ .type _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_, @function _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_: .LFB2054: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_, .-_Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_ .globl _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .type _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, @function _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_: .LFB2055: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .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 _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L19 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2055: .size _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, .-_Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .globl _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_ .type _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_, @function _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_: .LFB2056: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2056: .size _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_, .-_Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_ .globl _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2057: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L27 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .type _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, @function _Z13GPU_adi_x_newiiiPKfS0_S0_Pf: .LFB2058: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, .-_Z13GPU_adi_x_newiiiPKfS0_S0_Pf .globl _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2059: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L39 .L35: movq 168(%rsp), %rax subq %fs:40, %rax jne .L40 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L39: .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 _Z9GPU_adi_xiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L35 .L40: call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z9GPU_adi_xiiiPKfS0_S0_Pf .type _Z9GPU_adi_xiiiPKfS0_S0_Pf, @function _Z9GPU_adi_xiiiPKfS0_S0_Pf: .LFB2060: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _Z9GPU_adi_xiiiPKfS0_S0_Pf, .-_Z9GPU_adi_xiiiPKfS0_S0_Pf .globl _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2061: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L47 .L43: movq 168(%rsp), %rax subq %fs:40, %rax jne .L48 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L47: .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 _Z9GPU_adi_yiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L43 .L48: call __stack_chk_fail@PLT .cfi_endproc .LFE2061: .size _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z9GPU_adi_yiiiPKfS0_S0_Pf .type _Z9GPU_adi_yiiiPKfS0_S0_Pf, @function _Z9GPU_adi_yiiiPKfS0_S0_Pf: .LFB2062: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _Z9GPU_adi_yiiiPKfS0_S0_Pf, .-_Z9GPU_adi_yiiiPKfS0_S0_Pf .globl _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ .type _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_, @function _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_: .LFB2063: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 4(%rsp), %rax movq %rax, 144(%rsp) movq %rcx, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 152(%rsp) movq %r8, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 160(%rsp) movq %r9, 40(%rsp) leaq 40(%rsp), %rax movq %rax, 168(%rsp) movq 224(%rsp), %rax movq %rax, 48(%rsp) leaq 48(%rsp), %rax movq %rax, 176(%rsp) movq 232(%rsp), %rax movq %rax, 56(%rsp) leaq 56(%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 .L55 .L51: movq 200(%rsp), %rax subq %fs:40, %rax jne .L56 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L55: .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 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L51 .L56: call __stack_chk_fail@PLT .cfi_endproc .LFE2063: .size _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_, .-_Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ .globl _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .type _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, @function _Z9GPU_adi_ziiiPfPKfS1_S1_S1_: .LFB2064: .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 _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2064: .size _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, .-_Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z9GPU_adi_ziiiPfPKfS1_S1_S1_" .LC1: .string "_Z9GPU_adi_yiiiPKfS0_S0_Pf" .LC2: .string "_Z9GPU_adi_xiiiPKfS0_S0_Pf" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "_Z13GPU_adi_x_newiiiPKfS0_S0_Pf" .align 8 .LC4: .string "_Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_" .align 8 .LC5: .string "_Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_" .align 8 .LC6: .string "_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2066: .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 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_(%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 _Z9GPU_adi_yiiiPKfS0_S0_Pf(%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 _Z9GPU_adi_xiiiPKfS0_S0_Pf(%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 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf(%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 _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_(%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 _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_(%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 _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_(%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 .LFE2066: .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.
//////////////////////////////////////////////////////////////////////// // define kernel block size for //////////////////////////////////////////////////////////////////////// #define BLOCK_X 32 #define BLOCK_Y 8 // device code __global__ void GPU_adi_rhs(int NX, int NY, int NZ, float lam, const float* __restrict__ d_u, float* __restrict__ d_du, float* __restrict__ d_ax, float* __restrict__ d_bx, float* __restrict__ d_cx, float* __restrict__ d_ay, float* __restrict__ d_by, float* __restrict__ d_cy, float* __restrict__ d_az, float* __restrict__ d_bz, float* __restrict__ d_cz) { int i, j, k, indg, active; float du, a, b, c; int NXM1 = NX - 1; int NYM1 = NY - 1; int NZM1 = NZ - 1; #define IOFF 1 #define JOFF NX #define KOFF NX*NY // // set up indices for main block // i = threadIdx.x + blockIdx.x*BLOCK_X; j = threadIdx.y + blockIdx.y*BLOCK_Y; indg = i + j*NX; active = (i<NX) && (j<NY); // // loop over k-planes // for (k = 0; k<NZ; k++) { // // calculate r.h.s. and set a,b,c, coefficients // if (active) { if (i == 0 || i == NXM1 || j == 0 || j == NYM1 || k == 0 || k == NZM1) { du = 0.0f; // Dirichlet b.c.'s a = 0.0f; b = 1.0f; c = 0.0f; } else { du = lam * (d_u[indg - IOFF] + d_u[indg + IOFF] + d_u[indg - JOFF] + d_u[indg + JOFF] + d_u[indg - KOFF] + d_u[indg + KOFF] - 6.0f*d_u[indg]); a = -0.5f*lam; b = 1.0f + lam; c = -0.5f*lam; } d_du[indg] = du; d_ax[indg] = a; d_bx[indg] = b; d_cx[indg] = c; d_ay[indg] = a; d_by[indg] = b; d_cy[indg] = c; d_az[indg] = a; d_bz[indg] = b; d_cz[indg] = c; indg += KOFF; } } } __global__ void GPU_adi_x_float4(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; for (i = 4; i<NX; i += 4) { indg = indg + 1; a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; } // // reverse pass // d4.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4.x = dd; d_d[indg] = d4; for (i = NX - 5; i >= 0; i -= 4) { indg = indg - 1; dd = d2[i] - c2[i] * dd; d4.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4.x = dd; d_d[indg] = d4; } } } __global__ void GPU_adi_x_float4_2(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4, a4_2, b4_2, c4_2, d4_2; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; a4_2 = d_a[indg + 1]; b4 = d_b[indg]; b4_2 = d_b[indg + 1]; c4 = d_c[indg]; c4_2 = d_c[indg + 1]; d4 = d_d[indg]; d4_2 = d_d[indg + 1]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[4] = cc; d2[4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[5] = cc; d2[5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[6] = cc; d2[6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[7] = cc; d2[7] = dd; for (i = 8; i<NX; i += 8) { indg = indg + 2; a4 = d_a[indg]; a4_2 = d_a[indg + 1]; __threadfence_block(); b4 = d_b[indg]; b4_2 = d_b[indg + 1]; __threadfence_block(); c4 = d_c[indg]; c4_2 = d_c[indg + 1]; __threadfence_block(); d4 = d_d[indg]; d4_2 = d_d[indg + 1]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[i + 4] = cc; d2[i + 4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[i + 5] = cc; d2[i + 5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[i + 6] = cc; d2[i + 6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[i + 7] = cc; d2[i + 7] = dd; } // // reverse pass // d4_2.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4_2.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4_2.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4_2.x = dd; dd = d2[NX - 5] - c2[NX - 5] * dd; d4.w = dd; dd = d2[NX - 6] - c2[NX - 6] * dd; d4.z = dd; dd = d2[NX - 7] - c2[NX - 7] * dd; d4.y = dd; dd = d2[NX - 8] - c2[NX - 8] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; for (i = NX - 9; i >= 0; i -= 8) { indg = indg - 2; dd = d2[i] - c2[i] * dd; d4_2.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4_2.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4_2.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4_2.x = dd; dd = d2[i - 4] - c2[i - 4] * dd; d4.w = dd; dd = d2[i - 5] - c2[i - 5] * dd; d4.z = dd; dd = d2[i - 6] - c2[i - 6] * dd; d4.y = dd; dd = d2[i - 7] - c2[i - 7] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; } } } // // new tri-diagonal solve in x-direction // __global__ void GPU_adi_x_new(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int j, k, indg, t, tm, tp, nt, shift = 0; float bbi; __shared__ float a[256], c[256], d[256]; // // set up indices for main block // t = threadIdx.x; j = blockIdx.x; k = blockIdx.y; indg = t + NX*(j + k*NY); bbi = 1.0f / d_b[indg]; a[t] = -bbi * d_a[indg]; c[t] = -bbi * d_c[indg]; d[t] = bbi * d_d[indg]; // forward pass tm = 2 * t; t = tm + 1; tp = tm + 2; for (nt = blockDim.x / 2; nt>0; nt >>= 1) { shift++; __syncthreads(); if (threadIdx.x < nt) { bbi = 1.0f; if (tm >= 0) { bbi -= a[t] * c[tm]; d[t] += a[t] * d[tm]; a[t] = a[t] * a[tm]; } if (tp<NX) { bbi -= c[t] * a[tp]; d[t] += c[t] * d[tp]; c[t] = c[t] * c[tp]; } bbi = 1.0f / bbi; d[t] *= bbi; a[t] *= bbi; c[t] *= bbi; tm = 2 * tm + 1; t = 2 * t + 1; tp = 2 * tp + 1; } } // reverse pass for (; shift>0; shift--) { nt = blockDim.x >> shift; __syncthreads(); if (threadIdx.x < nt) { tm >>= 1; t >>= 1; tp >>= 1; if (tm >= 0) d[tm] += c[tm] * d[t]; if (tp<NX) d[tp] += a[tp] * d[t]; } } __syncthreads(); d_d[indg] = d[threadIdx.x]; } // // old tri-diagonal solve in x-direction // __global__ void GPU_adi_x(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY); if ((j<NY) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (i = 1; i<NX; i++) { indg = indg + 1; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[i] = cc; d[i] = dd; } // // reverse pass // d_d[indg] = dd; for (i = NX - 2; i >= 0; i--) { indg = indg - 1; dd = d[i] - c[i] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in y-direction // __global__ void GPU_adi_y(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = i + k*NX*NY; if ((i<NX) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (j = 1; j<NY; j++) { indg = indg + NX; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[j] = cc; d[j] = dd; } // // reverse pass // d_d[indg] = dd; for (j = NY - 2; j >= 0; j--) { indg = indg - NX; dd = d[j] - c[j] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in z-direction, and update solution // __global__ void GPU_adi_z(int NX, int NY, int NZ, float* __restrict__ d_u, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, const float* __restrict__ d_d) { int i, j, k, indg, off; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices j = threadIdx.y + blockIdx.y*blockDim.y; indg = i + j*NX; off = NX*NY; if ((i<NX) && (j<NY)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (k = 1; k<NZ; k++) { indg = indg + off; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[k] = cc; d[k] = dd; } // // reverse pass // d_u[indg] += dd; for (k = NZ - 2; k >= 0; k--) { indg = indg - off; dd = d[k] - c[k] * dd; d_u[indg] += dd; } } }
#include <hip/hip_runtime.h> //////////////////////////////////////////////////////////////////////// // define kernel block size for //////////////////////////////////////////////////////////////////////// #define BLOCK_X 32 #define BLOCK_Y 8 // device code __global__ void GPU_adi_rhs(int NX, int NY, int NZ, float lam, const float* __restrict__ d_u, float* __restrict__ d_du, float* __restrict__ d_ax, float* __restrict__ d_bx, float* __restrict__ d_cx, float* __restrict__ d_ay, float* __restrict__ d_by, float* __restrict__ d_cy, float* __restrict__ d_az, float* __restrict__ d_bz, float* __restrict__ d_cz) { int i, j, k, indg, active; float du, a, b, c; int NXM1 = NX - 1; int NYM1 = NY - 1; int NZM1 = NZ - 1; #define IOFF 1 #define JOFF NX #define KOFF NX*NY // // set up indices for main block // i = threadIdx.x + blockIdx.x*BLOCK_X; j = threadIdx.y + blockIdx.y*BLOCK_Y; indg = i + j*NX; active = (i<NX) && (j<NY); // // loop over k-planes // for (k = 0; k<NZ; k++) { // // calculate r.h.s. and set a,b,c, coefficients // if (active) { if (i == 0 || i == NXM1 || j == 0 || j == NYM1 || k == 0 || k == NZM1) { du = 0.0f; // Dirichlet b.c.'s a = 0.0f; b = 1.0f; c = 0.0f; } else { du = lam * (d_u[indg - IOFF] + d_u[indg + IOFF] + d_u[indg - JOFF] + d_u[indg + JOFF] + d_u[indg - KOFF] + d_u[indg + KOFF] - 6.0f*d_u[indg]); a = -0.5f*lam; b = 1.0f + lam; c = -0.5f*lam; } d_du[indg] = du; d_ax[indg] = a; d_bx[indg] = b; d_cx[indg] = c; d_ay[indg] = a; d_by[indg] = b; d_cy[indg] = c; d_az[indg] = a; d_bz[indg] = b; d_cz[indg] = c; indg += KOFF; } } } __global__ void GPU_adi_x_float4(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; for (i = 4; i<NX; i += 4) { indg = indg + 1; a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; } // // reverse pass // d4.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4.x = dd; d_d[indg] = d4; for (i = NX - 5; i >= 0; i -= 4) { indg = indg - 1; dd = d2[i] - c2[i] * dd; d4.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4.x = dd; d_d[indg] = d4; } } } __global__ void GPU_adi_x_float4_2(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4, a4_2, b4_2, c4_2, d4_2; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; a4_2 = d_a[indg + 1]; b4 = d_b[indg]; b4_2 = d_b[indg + 1]; c4 = d_c[indg]; c4_2 = d_c[indg + 1]; d4 = d_d[indg]; d4_2 = d_d[indg + 1]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[4] = cc; d2[4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[5] = cc; d2[5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[6] = cc; d2[6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[7] = cc; d2[7] = dd; for (i = 8; i<NX; i += 8) { indg = indg + 2; a4 = d_a[indg]; a4_2 = d_a[indg + 1]; __threadfence_block(); b4 = d_b[indg]; b4_2 = d_b[indg + 1]; __threadfence_block(); c4 = d_c[indg]; c4_2 = d_c[indg + 1]; __threadfence_block(); d4 = d_d[indg]; d4_2 = d_d[indg + 1]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[i + 4] = cc; d2[i + 4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[i + 5] = cc; d2[i + 5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[i + 6] = cc; d2[i + 6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[i + 7] = cc; d2[i + 7] = dd; } // // reverse pass // d4_2.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4_2.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4_2.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4_2.x = dd; dd = d2[NX - 5] - c2[NX - 5] * dd; d4.w = dd; dd = d2[NX - 6] - c2[NX - 6] * dd; d4.z = dd; dd = d2[NX - 7] - c2[NX - 7] * dd; d4.y = dd; dd = d2[NX - 8] - c2[NX - 8] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; for (i = NX - 9; i >= 0; i -= 8) { indg = indg - 2; dd = d2[i] - c2[i] * dd; d4_2.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4_2.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4_2.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4_2.x = dd; dd = d2[i - 4] - c2[i - 4] * dd; d4.w = dd; dd = d2[i - 5] - c2[i - 5] * dd; d4.z = dd; dd = d2[i - 6] - c2[i - 6] * dd; d4.y = dd; dd = d2[i - 7] - c2[i - 7] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; } } } // // new tri-diagonal solve in x-direction // __global__ void GPU_adi_x_new(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int j, k, indg, t, tm, tp, nt, shift = 0; float bbi; __shared__ float a[256], c[256], d[256]; // // set up indices for main block // t = threadIdx.x; j = blockIdx.x; k = blockIdx.y; indg = t + NX*(j + k*NY); bbi = 1.0f / d_b[indg]; a[t] = -bbi * d_a[indg]; c[t] = -bbi * d_c[indg]; d[t] = bbi * d_d[indg]; // forward pass tm = 2 * t; t = tm + 1; tp = tm + 2; for (nt = blockDim.x / 2; nt>0; nt >>= 1) { shift++; __syncthreads(); if (threadIdx.x < nt) { bbi = 1.0f; if (tm >= 0) { bbi -= a[t] * c[tm]; d[t] += a[t] * d[tm]; a[t] = a[t] * a[tm]; } if (tp<NX) { bbi -= c[t] * a[tp]; d[t] += c[t] * d[tp]; c[t] = c[t] * c[tp]; } bbi = 1.0f / bbi; d[t] *= bbi; a[t] *= bbi; c[t] *= bbi; tm = 2 * tm + 1; t = 2 * t + 1; tp = 2 * tp + 1; } } // reverse pass for (; shift>0; shift--) { nt = blockDim.x >> shift; __syncthreads(); if (threadIdx.x < nt) { tm >>= 1; t >>= 1; tp >>= 1; if (tm >= 0) d[tm] += c[tm] * d[t]; if (tp<NX) d[tp] += a[tp] * d[t]; } } __syncthreads(); d_d[indg] = d[threadIdx.x]; } // // old tri-diagonal solve in x-direction // __global__ void GPU_adi_x(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY); if ((j<NY) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (i = 1; i<NX; i++) { indg = indg + 1; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[i] = cc; d[i] = dd; } // // reverse pass // d_d[indg] = dd; for (i = NX - 2; i >= 0; i--) { indg = indg - 1; dd = d[i] - c[i] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in y-direction // __global__ void GPU_adi_y(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = i + k*NX*NY; if ((i<NX) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (j = 1; j<NY; j++) { indg = indg + NX; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[j] = cc; d[j] = dd; } // // reverse pass // d_d[indg] = dd; for (j = NY - 2; j >= 0; j--) { indg = indg - NX; dd = d[j] - c[j] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in z-direction, and update solution // __global__ void GPU_adi_z(int NX, int NY, int NZ, float* __restrict__ d_u, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, const float* __restrict__ d_d) { int i, j, k, indg, off; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices j = threadIdx.y + blockIdx.y*blockDim.y; indg = i + j*NX; off = NX*NY; if ((i<NX) && (j<NY)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (k = 1; k<NZ; k++) { indg = indg + off; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[k] = cc; d[k] = dd; } // // reverse pass // d_u[indg] += dd; for (k = NZ - 2; k >= 0; k--) { indg = indg - off; dd = d[k] - c[k] * dd; d_u[indg] += dd; } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> //////////////////////////////////////////////////////////////////////// // define kernel block size for //////////////////////////////////////////////////////////////////////// #define BLOCK_X 32 #define BLOCK_Y 8 // device code __global__ void GPU_adi_rhs(int NX, int NY, int NZ, float lam, const float* __restrict__ d_u, float* __restrict__ d_du, float* __restrict__ d_ax, float* __restrict__ d_bx, float* __restrict__ d_cx, float* __restrict__ d_ay, float* __restrict__ d_by, float* __restrict__ d_cy, float* __restrict__ d_az, float* __restrict__ d_bz, float* __restrict__ d_cz) { int i, j, k, indg, active; float du, a, b, c; int NXM1 = NX - 1; int NYM1 = NY - 1; int NZM1 = NZ - 1; #define IOFF 1 #define JOFF NX #define KOFF NX*NY // // set up indices for main block // i = threadIdx.x + blockIdx.x*BLOCK_X; j = threadIdx.y + blockIdx.y*BLOCK_Y; indg = i + j*NX; active = (i<NX) && (j<NY); // // loop over k-planes // for (k = 0; k<NZ; k++) { // // calculate r.h.s. and set a,b,c, coefficients // if (active) { if (i == 0 || i == NXM1 || j == 0 || j == NYM1 || k == 0 || k == NZM1) { du = 0.0f; // Dirichlet b.c.'s a = 0.0f; b = 1.0f; c = 0.0f; } else { du = lam * (d_u[indg - IOFF] + d_u[indg + IOFF] + d_u[indg - JOFF] + d_u[indg + JOFF] + d_u[indg - KOFF] + d_u[indg + KOFF] - 6.0f*d_u[indg]); a = -0.5f*lam; b = 1.0f + lam; c = -0.5f*lam; } d_du[indg] = du; d_ax[indg] = a; d_bx[indg] = b; d_cx[indg] = c; d_ay[indg] = a; d_by[indg] = b; d_cy[indg] = c; d_az[indg] = a; d_bz[indg] = b; d_cz[indg] = c; indg += KOFF; } } } __global__ void GPU_adi_x_float4(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; for (i = 4; i<NX; i += 4) { indg = indg + 1; a4 = d_a[indg]; b4 = d_b[indg]; c4 = d_c[indg]; d4 = d_d[indg]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; } // // reverse pass // d4.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4.x = dd; d_d[indg] = d4; for (i = NX - 5; i >= 0; i -= 4) { indg = indg - 1; dd = d2[i] - c2[i] * dd; d4.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4.x = dd; d_d[indg] = d4; } } } __global__ void GPU_adi_x_float4_2(int NX, int NY, int NZ, const float4* __restrict__ d_a, const float4* __restrict__ d_b, const float4* __restrict__ d_c, float4* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c2[256], d2[256]; float4 a4, b4, c4, d4, a4_2, b4_2, c4_2, d4_2; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY) / 4; if ((j<NY) && (k<NZ)) { // // forward pass // a4 = d_a[indg]; a4_2 = d_a[indg + 1]; b4 = d_b[indg]; b4_2 = d_b[indg + 1]; c4 = d_c[indg]; c4_2 = d_c[indg + 1]; d4 = d_d[indg]; d4_2 = d_d[indg + 1]; bb = 1.0f / b4.x; cc = bb * c4.x; dd = bb * d4.x; c2[0] = cc; d2[0] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[1] = cc; d2[1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[2] = cc; d2[2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[3] = cc; d2[3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[4] = cc; d2[4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[5] = cc; d2[5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[6] = cc; d2[6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[7] = cc; d2[7] = dd; for (i = 8; i<NX; i += 8) { indg = indg + 2; a4 = d_a[indg]; a4_2 = d_a[indg + 1]; __threadfence_block(); b4 = d_b[indg]; b4_2 = d_b[indg + 1]; __threadfence_block(); c4 = d_c[indg]; c4_2 = d_c[indg + 1]; __threadfence_block(); d4 = d_d[indg]; d4_2 = d_d[indg + 1]; aa = a4.x; bb = b4.x - aa*cc; dd = d4.x - aa*dd; bb = 1.0f / bb; cc = bb*c4.x; dd = bb*dd; c2[i] = cc; d2[i] = dd; aa = a4.y; bb = b4.y - aa*cc; dd = d4.y - aa*dd; bb = 1.0f / bb; cc = bb*c4.y; dd = bb*dd; c2[i + 1] = cc; d2[i + 1] = dd; aa = a4.z; bb = b4.z - aa*cc; dd = d4.z - aa*dd; bb = 1.0f / bb; cc = bb*c4.z; dd = bb*dd; c2[i + 2] = cc; d2[i + 2] = dd; aa = a4.w; bb = b4.w - aa*cc; dd = d4.w - aa*dd; bb = 1.0f / bb; cc = bb*c4.w; dd = bb*dd; c2[i + 3] = cc; d2[i + 3] = dd; aa = a4_2.x; bb = b4_2.x - aa*cc; dd = d4_2.x - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.x; dd = bb*dd; c2[i + 4] = cc; d2[i + 4] = dd; aa = a4_2.y; bb = b4_2.y - aa*cc; dd = d4_2.y - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.y; dd = bb*dd; c2[i + 5] = cc; d2[i + 5] = dd; aa = a4_2.z; bb = b4_2.z - aa*cc; dd = d4_2.z - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.z; dd = bb*dd; c2[i + 6] = cc; d2[i + 6] = dd; aa = a4_2.w; bb = b4_2.w - aa*cc; dd = d4_2.w - aa*dd; bb = 1.0f / bb; cc = bb*c4_2.w; dd = bb*dd; c2[i + 7] = cc; d2[i + 7] = dd; } // // reverse pass // d4_2.w = dd; dd = d2[NX - 2] - c2[NX - 2] * dd; d4_2.z = dd; dd = d2[NX - 3] - c2[NX - 3] * dd; d4_2.y = dd; dd = d2[NX - 4] - c2[NX - 4] * dd; d4_2.x = dd; dd = d2[NX - 5] - c2[NX - 5] * dd; d4.w = dd; dd = d2[NX - 6] - c2[NX - 6] * dd; d4.z = dd; dd = d2[NX - 7] - c2[NX - 7] * dd; d4.y = dd; dd = d2[NX - 8] - c2[NX - 8] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; for (i = NX - 9; i >= 0; i -= 8) { indg = indg - 2; dd = d2[i] - c2[i] * dd; d4_2.w = dd; dd = d2[i - 1] - c2[i - 1] * dd; d4_2.z = dd; dd = d2[i - 2] - c2[i - 2] * dd; d4_2.y = dd; dd = d2[i - 3] - c2[i - 3] * dd; d4_2.x = dd; dd = d2[i - 4] - c2[i - 4] * dd; d4.w = dd; dd = d2[i - 5] - c2[i - 5] * dd; d4.z = dd; dd = d2[i - 6] - c2[i - 6] * dd; d4.y = dd; dd = d2[i - 7] - c2[i - 7] * dd; d4.x = dd; d_d[indg + 1] = d4_2; d_d[indg] = d4; } } } // // new tri-diagonal solve in x-direction // __global__ void GPU_adi_x_new(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int j, k, indg, t, tm, tp, nt, shift = 0; float bbi; __shared__ float a[256], c[256], d[256]; // // set up indices for main block // t = threadIdx.x; j = blockIdx.x; k = blockIdx.y; indg = t + NX*(j + k*NY); bbi = 1.0f / d_b[indg]; a[t] = -bbi * d_a[indg]; c[t] = -bbi * d_c[indg]; d[t] = bbi * d_d[indg]; // forward pass tm = 2 * t; t = tm + 1; tp = tm + 2; for (nt = blockDim.x / 2; nt>0; nt >>= 1) { shift++; __syncthreads(); if (threadIdx.x < nt) { bbi = 1.0f; if (tm >= 0) { bbi -= a[t] * c[tm]; d[t] += a[t] * d[tm]; a[t] = a[t] * a[tm]; } if (tp<NX) { bbi -= c[t] * a[tp]; d[t] += c[t] * d[tp]; c[t] = c[t] * c[tp]; } bbi = 1.0f / bbi; d[t] *= bbi; a[t] *= bbi; c[t] *= bbi; tm = 2 * tm + 1; t = 2 * t + 1; tp = 2 * tp + 1; } } // reverse pass for (; shift>0; shift--) { nt = blockDim.x >> shift; __syncthreads(); if (threadIdx.x < nt) { tm >>= 1; t >>= 1; tp >>= 1; if (tm >= 0) d[tm] += c[tm] * d[t]; if (tp<NX) d[tp] += a[tp] * d[t]; } } __syncthreads(); d_d[indg] = d[threadIdx.x]; } // // old tri-diagonal solve in x-direction // __global__ void GPU_adi_x(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // j = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = NX*(j + k*NY); if ((j<NY) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (i = 1; i<NX; i++) { indg = indg + 1; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[i] = cc; d[i] = dd; } // // reverse pass // d_d[indg] = dd; for (i = NX - 2; i >= 0; i--) { indg = indg - 1; dd = d[i] - c[i] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in y-direction // __global__ void GPU_adi_y(int NX, int NY, int NZ, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, float* __restrict__ d_d) { int i, j, k, indg; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices k = threadIdx.y + blockIdx.y*blockDim.y; indg = i + k*NX*NY; if ((i<NX) && (k<NZ)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (j = 1; j<NY; j++) { indg = indg + NX; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[j] = cc; d[j] = dd; } // // reverse pass // d_d[indg] = dd; for (j = NY - 2; j >= 0; j--) { indg = indg - NX; dd = d[j] - c[j] * dd; d_d[indg] = dd; } } } // // tri-diagonal solve in z-direction, and update solution // __global__ void GPU_adi_z(int NX, int NY, int NZ, float* __restrict__ d_u, const float* __restrict__ d_a, const float* __restrict__ d_b, const float* __restrict__ d_c, const float* __restrict__ d_d) { int i, j, k, indg, off; float aa, bb, cc, dd, c[256], d[256]; // // set up indices for main block // i = threadIdx.x + blockIdx.x*blockDim.x; // global indices j = threadIdx.y + blockIdx.y*blockDim.y; indg = i + j*NX; off = NX*NY; if ((i<NX) && (j<NY)) { // // forward pass // bb = 1.0f / d_b[indg]; cc = bb*d_c[indg]; dd = bb*d_d[indg]; c[0] = cc; d[0] = dd; for (k = 1; k<NZ; k++) { indg = indg + off; aa = d_a[indg]; bb = d_b[indg] - aa*cc; dd = d_d[indg] - aa*dd; bb = 1.0f / bb; cc = bb*d_c[indg]; dd = bb*dd; c[k] = cc; d[k] = dd; } // // reverse pass // d_u[indg] += dd; for (k = NZ - 2; k >= 0; k--) { indg = indg - off; dd = d[k] - c[k] * dd; d_u[indg] += dd; } } }
.text .file "adi3d_kernel.hip" .globl _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ # -- Begin function _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .p2align 4, 0x90 .type _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_,@function _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: # @_Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .cfi_startproc # %bb.0: subq $216, %rsp .cfi_def_cfa_offset 224 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movss %xmm0, 8(%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 8(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 80(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rax movq %rax, 144(%rsp) leaq 224(%rsp), %rax movq %rax, 152(%rsp) leaq 232(%rsp), %rax movq %rax, 160(%rsp) leaq 240(%rsp), %rax movq %rax, 168(%rsp) leaq 248(%rsp), %rax movq %rax, 176(%rsp) leaq 256(%rsp), %rax movq %rax, 184(%rsp) leaq 264(%rsp), %rax movq %rax, 192(%rsp) leaq 272(%rsp), %rax movq %rax, 200(%rsp) leaq 280(%rsp), %rax movq %rax, 208(%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 $_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $232, %rsp .cfi_adjust_cfa_offset -232 retq .Lfunc_end0: .size _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .Lfunc_end0-_Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .cfi_endproc # -- End function .globl _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ # -- Begin function _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 4, 0x90 .type _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@function _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: # @_Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .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 160(%rsp), %rax movq %rax, 144(%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 $_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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_end1: .size _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, .Lfunc_end1-_Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .cfi_endproc # -- End function .globl _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ # -- Begin function _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 4, 0x90 .type _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@function _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: # @_Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .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 160(%rsp), %rax movq %rax, 144(%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 $_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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_end2: .size _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, .Lfunc_end2-_Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .cfi_endproc # -- End function .globl _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf # -- Begin function _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf,@function _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf: # @_Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z13GPU_adi_x_newiiiPKfS0_S0_Pf, %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_end3: .size _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf, .Lfunc_end3-_Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf # -- Begin function _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf,@function _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf: # @_Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z9GPU_adi_xiiiPKfS0_S0_Pf, %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_end4: .size _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf, .Lfunc_end4-_Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf # -- Begin function _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf,@function _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf: # @_Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z9GPU_adi_yiiiPKfS0_S0_Pf, %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_end5: .size _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf, .Lfunc_end5-_Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ # -- Begin function _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_,@function _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_: # @_Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 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 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%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 $_Z9GPU_adi_ziiiPfPKfS1_S1_S1_, %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_end6: .size _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_, .Lfunc_end6-_Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .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 .LBB7_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB7_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, %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 $_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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 $_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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 $_Z13GPU_adi_x_newiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_xiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_yiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_ziiiPfPKfS1_S1_S1_, %esi movl $.L__unnamed_7, %edx movl $.L__unnamed_7, %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_end7: .size __hip_module_ctor, .Lfunc_end7-__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 .LBB8_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 .LBB8_2: retq .Lfunc_end8: .size __hip_module_dtor, .Lfunc_end8-__hip_module_dtor .cfi_endproc # -- End function .type _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_,@object # @_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .section .rodata,"a",@progbits .globl _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .p2align 3, 0x0 _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .quad _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .size _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, 8 .type _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@object # @_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .globl _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 3, 0x0 _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: .quad _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .size _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, 8 .type _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@object # @_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .globl _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 3, 0x0 _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: .quad _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .size _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, 8 .type _Z13GPU_adi_x_newiiiPKfS0_S0_Pf,@object # @_Z13GPU_adi_x_newiiiPKfS0_S0_Pf .globl _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf: .quad _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .size _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_xiiiPKfS0_S0_Pf,@object # @_Z9GPU_adi_xiiiPKfS0_S0_Pf .globl _Z9GPU_adi_xiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z9GPU_adi_xiiiPKfS0_S0_Pf: .quad _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .size _Z9GPU_adi_xiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_yiiiPKfS0_S0_Pf,@object # @_Z9GPU_adi_yiiiPKfS0_S0_Pf .globl _Z9GPU_adi_yiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z9GPU_adi_yiiiPKfS0_S0_Pf: .quad _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .size _Z9GPU_adi_yiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_ziiiPfPKfS1_S1_S1_,@object # @_Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .globl _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .p2align 3, 0x0 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_: .quad _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .size _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_" .size .L__unnamed_1, 52 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_" .size .L__unnamed_2, 60 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_" .size .L__unnamed_3, 62 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z13GPU_adi_x_newiiiPKfS0_S0_Pf" .size .L__unnamed_4, 32 .type .L__unnamed_5,@object # @4 .L__unnamed_5: .asciz "_Z9GPU_adi_xiiiPKfS0_S0_Pf" .size .L__unnamed_5, 27 .type .L__unnamed_6,@object # @5 .L__unnamed_6: .asciz "_Z9GPU_adi_yiiiPKfS0_S0_Pf" .size .L__unnamed_6, 27 .type .L__unnamed_7,@object # @6 .L__unnamed_7: .asciz "_Z9GPU_adi_ziiiPfPKfS1_S1_S1_" .size .L__unnamed_7, 30 .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 _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .addrsig_sym _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .addrsig_sym _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_xiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_yiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .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_0017205a_00000000-6_adi3d_kernel.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 _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .type _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, @function _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .LFB2051: .cfi_startproc endbr64 subq $312, %rsp .cfi_def_cfa_offset 320 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movss %xmm0, (%rsp) movq %fs:40, %rax movq %rax, 296(%rsp) xorl %eax, %eax 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) movq %rcx, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 208(%rsp) movq %r8, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 216(%rsp) movq %r9, 40(%rsp) leaq 40(%rsp), %rax movq %rax, 224(%rsp) movq 320(%rsp), %rax movq %rax, 48(%rsp) leaq 48(%rsp), %rax movq %rax, 232(%rsp) movq 328(%rsp), %rax movq %rax, 56(%rsp) leaq 56(%rsp), %rax movq %rax, 240(%rsp) movq 336(%rsp), %rax movq %rax, 64(%rsp) leaq 64(%rsp), %rax movq %rax, 248(%rsp) movq 344(%rsp), %rax movq %rax, 72(%rsp) leaq 72(%rsp), %rax movq %rax, 256(%rsp) movq 352(%rsp), %rax movq %rax, 80(%rsp) leaq 80(%rsp), %rax movq %rax, 264(%rsp) movq 360(%rsp), %rax movq %rax, 88(%rsp) leaq 88(%rsp), %rax movq %rax, 272(%rsp) movq 368(%rsp), %rax movq %rax, 96(%rsp) leaq 96(%rsp), %rax movq %rax, 280(%rsp) movq 376(%rsp), %rax movq %rax, 104(%rsp) leaq 104(%rsp), %rax movq %rax, 288(%rsp) movl $1, 128(%rsp) movl $1, 132(%rsp) movl $1, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $1, 148(%rsp) leaq 120(%rsp), %rcx leaq 112(%rsp), %rdx leaq 140(%rsp), %rsi leaq 128(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 296(%rsp), %rax subq %fs:40, %rax jne .L8 addq $312, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 120(%rsp) .cfi_def_cfa_offset 328 pushq 120(%rsp) .cfi_def_cfa_offset 336 leaq 192(%rsp), %r9 movq 156(%rsp), %rcx movl 164(%rsp), %r8d movq 144(%rsp), %rsi movl 152(%rsp), %edx leaq _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 320 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .-_Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .globl _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .type _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, @function _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .LFB2052: .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 _Z65__device_stub__Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_iiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ addq $72, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .-_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .globl _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .type _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, @function _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_: .LFB2053: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L15 .L11: movq 168(%rsp), %rax subq %fs:40, %rax jne .L16 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, .-_Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .globl _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_ .type _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_, @function _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_: .LFB2054: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z55__device_stub__Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_, .-_Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_ .globl _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .type _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, @function _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_: .LFB2055: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .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 _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L19 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2055: .size _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_, .-_Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ .globl _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_ .type _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_, @function _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_: .LFB2056: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z57__device_stub__Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_iiiPK6float4S1_S1_PS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2056: .size _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_, .-_Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_ .globl _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2057: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L27 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .type _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, @function _Z13GPU_adi_x_newiiiPKfS0_S0_Pf: .LFB2058: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z45__device_stub__Z13GPU_adi_x_newiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, .-_Z13GPU_adi_x_newiiiPKfS0_S0_Pf .globl _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2059: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L39 .L35: movq 168(%rsp), %rax subq %fs:40, %rax jne .L40 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L39: .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 _Z9GPU_adi_xiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L35 .L40: call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z9GPU_adi_xiiiPKfS0_S0_Pf .type _Z9GPU_adi_xiiiPKfS0_S0_Pf, @function _Z9GPU_adi_xiiiPKfS0_S0_Pf: .LFB2060: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z40__device_stub__Z9GPU_adi_xiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _Z9GPU_adi_xiiiPKfS0_S0_Pf, .-_Z9GPU_adi_xiiiPKfS0_S0_Pf .globl _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .type _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, @function _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf: .LFB2061: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rcx, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) movq %r8, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) movq %r9, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) movq 192(%rsp), %rax movq %rax, 40(%rsp) leaq 40(%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 .L47 .L43: movq 168(%rsp), %rax subq %fs:40, %rax jne .L48 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L47: .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 _Z9GPU_adi_yiiiPKfS0_S0_Pf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L43 .L48: call __stack_chk_fail@PLT .cfi_endproc .LFE2061: .size _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf, .-_Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf .globl _Z9GPU_adi_yiiiPKfS0_S0_Pf .type _Z9GPU_adi_yiiiPKfS0_S0_Pf, @function _Z9GPU_adi_yiiiPKfS0_S0_Pf: .LFB2062: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z40__device_stub__Z9GPU_adi_yiiiPKfS0_S0_PfiiiPKfS0_S0_Pf addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _Z9GPU_adi_yiiiPKfS0_S0_Pf, .-_Z9GPU_adi_yiiiPKfS0_S0_Pf .globl _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ .type _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_, @function _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_: .LFB2063: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movl %edx, 4(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 4(%rsp), %rax movq %rax, 144(%rsp) movq %rcx, 24(%rsp) leaq 24(%rsp), %rax movq %rax, 152(%rsp) movq %r8, 32(%rsp) leaq 32(%rsp), %rax movq %rax, 160(%rsp) movq %r9, 40(%rsp) leaq 40(%rsp), %rax movq %rax, 168(%rsp) movq 224(%rsp), %rax movq %rax, 48(%rsp) leaq 48(%rsp), %rax movq %rax, 176(%rsp) movq 232(%rsp), %rax movq %rax, 56(%rsp) leaq 56(%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 .L55 .L51: movq 200(%rsp), %rax subq %fs:40, %rax jne .L56 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L55: .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 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L51 .L56: call __stack_chk_fail@PLT .cfi_endproc .LFE2063: .size _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_, .-_Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ .globl _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .type _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, @function _Z9GPU_adi_ziiiPfPKfS1_S1_S1_: .LFB2064: .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 _Z43__device_stub__Z9GPU_adi_ziiiPfPKfS1_S1_S1_iiiPfPKfS1_S1_S1_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2064: .size _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, .-_Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z9GPU_adi_ziiiPfPKfS1_S1_S1_" .LC1: .string "_Z9GPU_adi_yiiiPKfS0_S0_Pf" .LC2: .string "_Z9GPU_adi_xiiiPKfS0_S0_Pf" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "_Z13GPU_adi_x_newiiiPKfS0_S0_Pf" .align 8 .LC4: .string "_Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_" .align 8 .LC5: .string "_Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_" .align 8 .LC6: .string "_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2066: .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 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_(%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 _Z9GPU_adi_yiiiPKfS0_S0_Pf(%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 _Z9GPU_adi_xiiiPKfS0_S0_Pf(%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 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf(%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 _Z18GPU_adi_x_float4_2iiiPK6float4S1_S1_PS_(%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 _Z16GPU_adi_x_float4iiiPK6float4S1_S1_PS_(%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 _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_(%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 .LFE2066: .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 "adi3d_kernel.hip" .globl _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ # -- Begin function _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .p2align 4, 0x90 .type _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_,@function _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: # @_Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .cfi_startproc # %bb.0: subq $216, %rsp .cfi_def_cfa_offset 224 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movl %edx, 12(%rsp) movss %xmm0, 8(%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 8(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 80(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rax movq %rax, 144(%rsp) leaq 224(%rsp), %rax movq %rax, 152(%rsp) leaq 232(%rsp), %rax movq %rax, 160(%rsp) leaq 240(%rsp), %rax movq %rax, 168(%rsp) leaq 248(%rsp), %rax movq %rax, 176(%rsp) leaq 256(%rsp), %rax movq %rax, 184(%rsp) leaq 264(%rsp), %rax movq %rax, 192(%rsp) leaq 272(%rsp), %rax movq %rax, 200(%rsp) leaq 280(%rsp), %rax movq %rax, 208(%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 $_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $232, %rsp .cfi_adjust_cfa_offset -232 retq .Lfunc_end0: .size _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, .Lfunc_end0-_Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .cfi_endproc # -- End function .globl _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ # -- Begin function _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 4, 0x90 .type _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@function _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: # @_Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .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 160(%rsp), %rax movq %rax, 144(%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 $_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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_end1: .size _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, .Lfunc_end1-_Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .cfi_endproc # -- End function .globl _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ # -- Begin function _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 4, 0x90 .type _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@function _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: # @_Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .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 160(%rsp), %rax movq %rax, 144(%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 $_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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_end2: .size _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, .Lfunc_end2-_Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .cfi_endproc # -- End function .globl _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf # -- Begin function _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf,@function _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf: # @_Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z13GPU_adi_x_newiiiPKfS0_S0_Pf, %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_end3: .size _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf, .Lfunc_end3-_Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf # -- Begin function _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf,@function _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf: # @_Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z9GPU_adi_xiiiPKfS0_S0_Pf, %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_end4: .size _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf, .Lfunc_end4-_Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf # -- Begin function _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf,@function _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf: # @_Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .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 160(%rsp), %rax movq %rax, 144(%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 $_Z9GPU_adi_yiiiPKfS0_S0_Pf, %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_end5: .size _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf, .Lfunc_end5-_Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .cfi_endproc # -- End function .globl _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ # -- Begin function _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .p2align 4, 0x90 .type _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_,@function _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_: # @_Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 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 176(%rsp), %rax movq %rax, 144(%rsp) leaq 184(%rsp), %rax movq %rax, 152(%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 $_Z9GPU_adi_ziiiPfPKfS1_S1_S1_, %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_end6: .size _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_, .Lfunc_end6-_Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .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 .LBB7_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB7_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, %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 $_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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 $_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, %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 $_Z13GPU_adi_x_newiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_xiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_yiiiPKfS0_S0_Pf, %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 $_Z9GPU_adi_ziiiPfPKfS1_S1_S1_, %esi movl $.L__unnamed_7, %edx movl $.L__unnamed_7, %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_end7: .size __hip_module_ctor, .Lfunc_end7-__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 .LBB8_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 .LBB8_2: retq .Lfunc_end8: .size __hip_module_dtor, .Lfunc_end8-__hip_module_dtor .cfi_endproc # -- End function .type _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_,@object # @_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .section .rodata,"a",@progbits .globl _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .p2align 3, 0x0 _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_: .quad _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .size _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_, 8 .type _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@object # @_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .globl _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 3, 0x0 _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: .quad _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .size _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, 8 .type _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_,@object # @_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .globl _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .p2align 3, 0x0 _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_: .quad _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .size _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_, 8 .type _Z13GPU_adi_x_newiiiPKfS0_S0_Pf,@object # @_Z13GPU_adi_x_newiiiPKfS0_S0_Pf .globl _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z13GPU_adi_x_newiiiPKfS0_S0_Pf: .quad _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .size _Z13GPU_adi_x_newiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_xiiiPKfS0_S0_Pf,@object # @_Z9GPU_adi_xiiiPKfS0_S0_Pf .globl _Z9GPU_adi_xiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z9GPU_adi_xiiiPKfS0_S0_Pf: .quad _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .size _Z9GPU_adi_xiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_yiiiPKfS0_S0_Pf,@object # @_Z9GPU_adi_yiiiPKfS0_S0_Pf .globl _Z9GPU_adi_yiiiPKfS0_S0_Pf .p2align 3, 0x0 _Z9GPU_adi_yiiiPKfS0_S0_Pf: .quad _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .size _Z9GPU_adi_yiiiPKfS0_S0_Pf, 8 .type _Z9GPU_adi_ziiiPfPKfS1_S1_S1_,@object # @_Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .globl _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .p2align 3, 0x0 _Z9GPU_adi_ziiiPfPKfS1_S1_S1_: .quad _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .size _Z9GPU_adi_ziiiPfPKfS1_S1_S1_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_" .size .L__unnamed_1, 52 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_" .size .L__unnamed_2, 60 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_" .size .L__unnamed_3, 62 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z13GPU_adi_x_newiiiPKfS0_S0_Pf" .size .L__unnamed_4, 32 .type .L__unnamed_5,@object # @4 .L__unnamed_5: .asciz "_Z9GPU_adi_xiiiPKfS0_S0_Pf" .size .L__unnamed_5, 27 .type .L__unnamed_6,@object # @5 .L__unnamed_6: .asciz "_Z9GPU_adi_yiiiPKfS0_S0_Pf" .size .L__unnamed_6, 27 .type .L__unnamed_7,@object # @6 .L__unnamed_7: .asciz "_Z9GPU_adi_ziiiPfPKfS1_S1_S1_" .size .L__unnamed_7, 30 .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 _Z26__device_stub__GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .addrsig_sym _Z31__device_stub__GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z33__device_stub__GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z28__device_stub__GPU_adi_x_newiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_xiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_yiiiPKfS0_S0_Pf .addrsig_sym _Z24__device_stub__GPU_adi_ziiiPfPKfS1_S1_S1_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11GPU_adi_rhsiiifPKfPfS1_S1_S1_S1_S1_S1_S1_S1_S1_ .addrsig_sym _Z16GPU_adi_x_float4iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z18GPU_adi_x_float4_2iiiPK15HIP_vector_typeIfLj4EES2_S2_PS0_ .addrsig_sym _Z13GPU_adi_x_newiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_xiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_yiiiPKfS0_S0_Pf .addrsig_sym _Z9GPU_adi_ziiiPfPKfS1_S1_S1_ .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 "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
//// ////#include "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
.file "tmpxft_00172c4d_00000000-6_nonLinearityFunction.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.
//// ////#include "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
#include <hip/hip_runtime.h> //// ////#include "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> //// ////#include "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
.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 "BC_NN_Functions.h" //// //#include "Tensor.h" // __global__ // void sigmoid(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] = 1 / (1+ pow(2.71828, -x[i])); // } // } // __global__ // void sigmoid_deriv(float* x, unsigned sz) { // for (int i = 0; i < sz; ++i) { // x[i] *= (1 - x[i]); // } // } //// // void sigmoid(Tensor<float, GPU>& x) { sigmoid<<<10, 128>>>(x.data(), x.size());} // void sigmoid_deriv(Tensor<float, GPU>& x) { sigmoid_deriv<<<10, 128>>>(x.data(), x.size());} // ////
.text .file "nonLinearityFunction.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_00172c4d_00000000-6_nonLinearityFunction.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 "nonLinearityFunction.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 device assembly.
#define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
.file "tmpxft_001a8421_00000000-6_Vector2MatrixTensor.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 .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .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 .LFE2053: .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.
#define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
#include <hip/hip_runtime.h> #define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
.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> #define TILE_DIM 32 template<typename T> __device__ void vectorDotMatrix(const T* vector, const T* matrix, T* result, const int rows, const int columns) { __shared__ T vector_tile[TILE_DIM]; __shared__ T matrix_tile[TILE_DIM][TILE_DIM]; int bx = blockIdx.x; int tx = threadIdx.x; int index = bx * blockDim.x + tx; T resultValue = 0; for (int t = 0; t < (rows - 1) / TILE_DIM + 1; t++) { int idx = t * TILE_DIM + tx; if (idx < rows) { vector_tile[tx] = vector[idx]; } else { vector_tile[tx] = 0; } if (index < columns) { int firstTileRow = t * TILE_DIM; for (int i = 0; i < TILE_DIM; i++) { int row = firstTileRow + i; if (row < rows) { matrix_tile[i][tx] = matrix[row * columns + index]; } else { matrix_tile[i][tx] = 0; } } } else { for (int i = 0; i < TILE_DIM; i++) { matrix_tile[i][tx] = 0; } } __syncthreads(); for (int i = 0; i < TILE_DIM; i++) { resultValue += vector_tile[i] * matrix_tile[i][tx]; } __syncthreads(); } if (index < columns) { result[index] = resultValue; } }
.text .file "Vector2MatrixTensor.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_001a8421_00000000-6_Vector2MatrixTensor.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 .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .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 .LFE2053: .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 "Vector2MatrixTensor.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 device assembly.
#include <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; cudaMalloc((void**)&din, n * sizeof(float)); cudaMalloc((void**)&dout, n * sizeof(float)); cudaMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); cudaMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); cudaMallocManaged((void**)&dummy, sizeof(float)); cudaMemcpy(din, in, n * sizeof(float), cudaMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); cudaDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); cudaDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); cudaDeviceSynchronize(); cudaMemcpy(out, dout, n * sizeof(float), cudaMemcpyDeviceToHost); cudaFree(din); cudaFree(dout); cudaFree(block_sums); cudaFree(block_incrs); cudaFree(dummy); }
code for sm_80 Function : _Z13hillis_steelePfS_iS_ .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 R9, SR_CTAID.X ; /* 0x0000000000097919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0040*/ BSSY B0, 0x100 ; /* 0x000000b000007945 */ /* 0x000fe20003800000 */ /*0050*/ S2R R10, SR_TID.X ; /* 0x00000000000a7919 */ /* 0x000e240000002100 */ /*0060*/ IMAD R0, R9, c[0x0][0x0], R10 ; /* 0x0000000009007a24 */ /* 0x001fca00078e020a */ /*0070*/ IMAD.WIDE R2, R0.reuse, R3, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x040fe200078e0203 */ /*0080*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fc80003f06270 */ /*0090*/ ISETP.EQ.OR P1, PT, R10, RZ, P0 ; /* 0x000000ff0a00720c */ /* 0x000fda0000722670 */ /*00a0*/ @P1 BRA 0xe0 ; /* 0x0000003000001947 */ /* 0x000fea0003800000 */ /*00b0*/ LDG.E R4, [R2.64+-0x4] ; /* 0xfffffc0402047981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ STS [R10.X4], R4 ; /* 0x000000040a007388 */ /* 0x0041e20000004800 */ /*00d0*/ BRA 0xf0 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*00e0*/ STS [R10.X4], RZ ; /* 0x000000ff0a007388 */ /* 0x0001e40000004800 */ /*00f0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0100*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff057624 */ /* 0x000fe200078e00ff */ /*0110*/ BSSY B0, 0x1b0 ; /* 0x0000009000007945 */ /* 0x000fe80003800000 */ /*0120*/ IADD3 R4, R5, -0x1, RZ ; /* 0xffffffff05047810 */ /* 0x001fc40007ffe0ff */ /*0130*/ ISETP.GE.AND P2, PT, R5, 0x2, PT ; /* 0x000000020500780c */ /* 0x000fe20003f46270 */ /*0140*/ HFMA2.MMA R5, -RZ, RZ, 0, 0 ; /* 0x00000000ff057435 */ /* 0x000fe200000001ff */ /*0150*/ ISETP.NE.AND P1, PT, R10, R4, PT ; /* 0x000000040a00720c */ /* 0x000fda0003f25270 */ /*0160*/ @P1 BRA 0x1a0 ; /* 0x0000003000001947 */ /* 0x000fea0003800000 */ /*0170*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe200078e00ff */ /*0180*/ @P0 BRA 0x1a0 ; /* 0x0000001000000947 */ /* 0x000fea0003800000 */ /*0190*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000164000c1e1900 */ /*01a0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*01b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*01c0*/ @!P2 BRA 0x2e0 ; /* 0x000001100000a947 */ /* 0x000fea0003800000 */ /*01d0*/ MOV R5, RZ ; /* 0x000000ff00057202 */ /* 0x000fe20000000f00 */ /*01e0*/ IMAD.MOV.U32 R2, RZ, RZ, 0x1 ; /* 0x00000001ff027424 */ /* 0x001fca00078e00ff */ /*01f0*/ ISETP.GE.AND P2, PT, R10, R2, PT ; /* 0x000000020a00720c */ /* 0x000fe20003f46270 */ /*0200*/ IMAD R3, R5.reuse, c[0x0][0x0], R10 ; /* 0x0000000005037a24 */ /* 0x040fe200078e020a */ /*0210*/ IADD3 R5, -R5, 0x1, RZ ; /* 0x0000000105057810 */ /* 0x000fca0007ffe1ff */ /*0220*/ IMAD R6, R5, c[0x0][0x0], R10 ; /* 0x0000000005067a24 */ /* 0x000fc800078e020a */ /*0230*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fe400078e00ff */ /*0240*/ @P2 IADD3 R7, R3, -R2, RZ ; /* 0x8000000203072210 */ /* 0x000fe40007ffe0ff */ /*0250*/ LDS R3, [R3.X4] ; /* 0x0000000003037984 */ /* 0x000fe20000004800 */ /*0260*/ SHF.L.U32 R2, R2, 0x1, RZ ; /* 0x0000000102027819 */ /* 0x000fc600000006ff */ /*0270*/ @P2 LDS R8, [R7.X4] ; /* 0x0000000007082984 */ /* 0x000e240000004800 */ /*0280*/ @P2 FADD R8, R3, R8 ; /* 0x0000000803082221 */ /* 0x001fca0000000000 */ /*0290*/ @P2 STS [R6], R8 ; /* 0x0000000806002388 */ /* 0x0001e80000000800 */ /*02a0*/ @!P2 STS [R6], R3 ; /* 0x000000030600a388 */ /* 0x0001e80000000800 */ /*02b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*02c0*/ ISETP.GE.AND P2, PT, R2, c[0x0][0x0], PT ; /* 0x0000000002007a0c */ /* 0x000fda0003f46270 */ /*02d0*/ @!P2 BRA 0x1f0 ; /* 0xffffff100000a947 */ /* 0x001fea000383ffff */ /*02e0*/ IMAD R6, R5, c[0x0][0x0], R10 ; /* 0x0000000005067a24 */ /* 0x000fe200078e020a */ /*02f0*/ @!P0 MOV R3, 0x4 ; /* 0x0000000400038802 */ /* 0x001fc60000000f00 */ /*0300*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fe400078e00ff */ /*0310*/ @!P0 IMAD.WIDE R2, R0, R3, c[0x0][0x168] ; /* 0x00005a0000028625 */ /* 0x000fc600078e0203 */ /*0320*/ @!P0 LDS R5, [R6] ; /* 0x0000000006058984 */ /* 0x000e280000000800 */ /*0330*/ @!P0 STG.E [R2.64], R5 ; /* 0x0000000502008986 */ /* 0x0011e2000c101904 */ /*0340*/ @P1 EXIT ; /* 0x000000000000194d */ /* 0x000fea0003800000 */ /*0350*/ LDS R3, [R6] ; /* 0x0000000006037984 */ /* 0x001e220000000800 */ /*0360*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fe400078e00ff */ /*0370*/ FADD R5, R3, R4 ; /* 0x0000000403057221 */ /* 0x021fe40000000000 */ /*0380*/ IMAD.WIDE.U32 R2, R9, R2, c[0x0][0x178] ; /* 0x00005e0009027625 */ /* 0x000fca00078e0002 */ /*0390*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*03a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*03b0*/ BRA 0x3b0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*03c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0400*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0410*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0420*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0430*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0440*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0450*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0460*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0470*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z5adderPfS_i .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 R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ BSSY B0, 0xe0 ; /* 0x000000a000007945 */ /* 0x000fe40003800000 */ /*0040*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e620000002500 */ /*0050*/ ISETP.NE.AND P0, PT, R3, RZ, PT ; /* 0x000000ff0300720c */ /* 0x001fe20003f05270 */ /*0060*/ IMAD R0, R2, c[0x0][0x0], R3 ; /* 0x0000000002007a24 */ /* 0x002fca00078e0203 */ /*0070*/ ISETP.GE.AND P1, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fce0003f26270 */ /*0080*/ @P0 BRA 0xd0 ; /* 0x0000004000000947 */ /* 0x000fea0003800000 */ /*0090*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fd400000001ff */ /*00a0*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fcc00078e0003 */ /*00b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ STS [RZ], R2 ; /* 0x00000002ff007388 */ /* 0x0041e40000000800 */ /*00d0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*00e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*00f0*/ @P1 EXIT ; /* 0x000000000000194d */ /* 0x000fea0003800000 */ /*0100*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fca0000000f00 */ /*0110*/ IMAD.WIDE R2, R0, R3, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x001fe400078e0203 */ /*0120*/ LDS R0, [RZ] ; /* 0x00000000ff007984 */ /* 0x000e280000000800 */ /*0130*/ LDG.E R5, [R2.64] ; /* 0x0000000402057981 */ /* 0x000e24000c1e1900 */ /*0140*/ FADD R5, R0, R5 ; /* 0x0000000500057221 */ /* 0x001fca0000000000 */ /*0150*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; cudaMalloc((void**)&din, n * sizeof(float)); cudaMalloc((void**)&dout, n * sizeof(float)); cudaMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); cudaMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); cudaMallocManaged((void**)&dummy, sizeof(float)); cudaMemcpy(din, in, n * sizeof(float), cudaMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); cudaDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); cudaDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); cudaDeviceSynchronize(); cudaMemcpy(out, dout, n * sizeof(float), cudaMemcpyDeviceToHost); cudaFree(din); cudaFree(dout); cudaFree(block_sums); cudaFree(block_incrs); cudaFree(dummy); }
.file "tmpxft_0009604a_00000000-6_scan.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 _Z27__device_stub__Z5adderPfS_iPfS_i .type _Z27__device_stub__Z5adderPfS_iPfS_i, @function _Z27__device_stub__Z5adderPfS_iPfS_i: .LFB2082: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 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 _Z5adderPfS_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z27__device_stub__Z5adderPfS_iPfS_i, .-_Z27__device_stub__Z5adderPfS_iPfS_i .globl _Z5adderPfS_i .type _Z5adderPfS_i, @function _Z5adderPfS_i: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z5adderPfS_iPfS_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5adderPfS_i, .-_Z5adderPfS_i .globl _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ .type _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_, @function _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_: .LFB2084: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%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 12(%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 .L15 .L11: movq 136(%rsp), %rax subq %fs:40, %rax jne .L16 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z13hillis_steelePfS_iS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_, .-_Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ .globl _Z13hillis_steelePfS_iS_ .type _Z13hillis_steelePfS_iS_, @function _Z13hillis_steelePfS_iS_: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z13hillis_steelePfS_iS_, .-_Z13hillis_steelePfS_iS_ .globl _Z4scanPKfPfjj .type _Z4scanPKfPfjj, @function _Z4scanPKfPfjj: .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 $104, %rsp .cfi_def_cfa_offset 160 movq %rdi, (%rsp) movq %rsi, 8(%rsp) movl %edx, %r14d movl %ecx, %ebx movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leal -1(%rcx), %r12d leal (%rdx,%r12), %eax movl $0, %edx divl %ecx movl %eax, %ebp movl %r14d, %r13d salq $2, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT leaq 32(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT movslq %ebp, %r15 salq $2, %r15 leaq 40(%rsp), %rdi movl $1, %edx movq %r15, %rsi call cudaMallocManaged@PLT leaq 48(%rsp), %rdi movl $1, %edx movq %r15, %rsi call cudaMallocManaged@PLT leaq 56(%rsp), %rdi movl $1, %edx movl $4, %esi call cudaMallocManaged@PLT movl $0, %ecx movq %r13, %rdx movq (%rsp), %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT leal (%rbx,%rbx), %r15d salq $2, %r15 movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl %ebp, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movq %r15, %r8 movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L20: call cudaDeviceSynchronize@PLT movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leal 0(%rbp,%r12), %eax movl $0, %edx divl %ebx movl %eax, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movq %r15, %r8 movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L26 .L21: call cudaDeviceSynchronize@PLT movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl %ebp, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movl $4, %r8d movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L27 .L22: call cudaDeviceSynchronize@PLT movl $2, %ecx movq %r13, %rdx movq 32(%rsp), %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@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 .L28 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 .L25: .cfi_restore_state movq 40(%rsp), %rcx movl %r14d, %edx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ jmp .L20 .L26: movq 56(%rsp), %rcx movl %ebp, %edx movq 48(%rsp), %rsi movq 40(%rsp), %rdi call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ jmp .L21 .L27: movl %r14d, %edx movq 48(%rsp), %rsi movq 32(%rsp), %rdi call _Z27__device_stub__Z5adderPfS_iPfS_i jmp .L22 .L28: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z4scanPKfPfjj, .-_Z4scanPKfPfjj .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z13hillis_steelePfS_iS_" .LC1: .string "_Z5adderPfS_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .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 _Z13hillis_steelePfS_iS_(%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 _Z5adderPfS_i(%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 .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 .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 <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; cudaMalloc((void**)&din, n * sizeof(float)); cudaMalloc((void**)&dout, n * sizeof(float)); cudaMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); cudaMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); cudaMallocManaged((void**)&dummy, sizeof(float)); cudaMemcpy(din, in, n * sizeof(float), cudaMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); cudaDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); cudaDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); cudaDeviceSynchronize(); cudaMemcpy(out, dout, n * sizeof(float), cudaMemcpyDeviceToHost); cudaFree(din); cudaFree(dout); cudaFree(block_sums); cudaFree(block_incrs); cudaFree(dummy); }
#include <hip/hip_runtime.h> #include <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; hipMalloc((void**)&din, n * sizeof(float)); hipMalloc((void**)&dout, n * sizeof(float)); hipMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); hipMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); hipMallocManaged((void**)&dummy, sizeof(float)); hipMemcpy(din, in, n * sizeof(float), hipMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); hipDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); hipDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); hipDeviceSynchronize(); hipMemcpy(out, dout, n * sizeof(float), hipMemcpyDeviceToHost); hipFree(din); hipFree(dout); hipFree(block_sums); hipFree(block_incrs); hipFree(dummy); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; hipMalloc((void**)&din, n * sizeof(float)); hipMalloc((void**)&dout, n * sizeof(float)); hipMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); hipMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); hipMallocManaged((void**)&dummy, sizeof(float)); hipMemcpy(din, in, n * sizeof(float), hipMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); hipDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); hipDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); hipDeviceSynchronize(); hipMemcpy(out, dout, n * sizeof(float), hipMemcpyDeviceToHost); hipFree(din); hipFree(dout); hipFree(block_sums); hipFree(block_incrs); hipFree(dummy); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5adderPfS_i .globl _Z5adderPfS_i .p2align 8 .type _Z5adderPfS_i,@function _Z5adderPfS_i: s_load_b32 s4, s[0:1], 0x24 s_mov_b32 s2, s15 s_mov_b32 s3, 0 s_mov_b32 s5, exec_lo v_cmpx_eq_u32_e32 0, v0 s_cbranch_execz .LBB0_2 s_load_b64 s[6:7], s[0:1], 0x8 s_lshl_b64 s[8:9], s[2:3], 2 s_waitcnt lgkmcnt(0) s_add_u32 s6, s6, s8 s_addc_u32 s7, s7, s9 s_load_b32 s3, s[6:7], 0x0 s_waitcnt lgkmcnt(0) v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, s3 ds_store_b32 v1, v2 .LBB0_2: s_or_b32 exec_lo, exec_lo, s5 s_load_b32 s3, s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s4, 0xffff, s4 s_delay_alu instid0(SALU_CYCLE_1) v_mad_u64_u32 v[1:2], null, s2, s4, v[0:1] s_barrier buffer_gl0_inv s_mov_b32 s2, exec_lo v_cmpx_gt_i32_e64 s3, v1 s_cbranch_execz .LBB0_4 s_load_b64 s[0:1], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 v_mov_b32_e32 v3, 0 s_delay_alu instid0(VALU_DEP_2) v_lshlrev_b64 v[0:1], 2, v[1:2] ds_load_b32 v3, v3 s_waitcnt lgkmcnt(0) 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[0:1], off s_waitcnt vmcnt(0) v_add_f32_e32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5adderPfS_i .amdhsa_group_segment_fixed_size 0 .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 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 _Z5adderPfS_i, .Lfunc_end0-_Z5adderPfS_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z13hillis_steelePfS_iS_ .globl _Z13hillis_steelePfS_iS_ .p2align 8 .type _Z13hillis_steelePfS_iS_,@function _Z13hillis_steelePfS_iS_: s_clause 0x2 s_load_b32 s2, s[0:1], 0x2c s_load_b32 s3, s[0:1], 0x10 s_load_b64 s[6:7], s[0:1], 0x0 s_mov_b32 s4, s15 v_mov_b32_e32 v5, 0 v_mov_b32_e32 v3, 0 s_waitcnt lgkmcnt(0) s_and_b32 s5, s2, 0xffff v_cmp_ne_u32_e64 s2, 0, v0 v_mad_u64_u32 v[1:2], null, s4, s5, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_cmp_gt_i32_e32 vcc_lo, s3, v1 v_ashrrev_i32_e32 v2, 31, v1 s_and_b32 s2, s2, vcc_lo s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB1_2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 2, v[1:2] v_add_co_u32 v3, s2, s6, v3 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v4, s2, s7, v4, s2 global_load_b32 v3, v[3:4], off offset:-4 .LBB1_2: s_or_b32 exec_lo, exec_lo, s3 v_lshl_add_u32 v4, v0, 2, 0 s_add_i32 s2, s5, -1 s_mov_b64 s[8:9], src_shared_base v_cmp_eq_u32_e64 s2, s2, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_ne_u32_e64 s3, -1, v4 v_cndmask_b32_e64 v6, 0, v4, s3 v_cndmask_b32_e64 v7, 0, s9, s3 s_delay_alu instid0(VALU_DEP_4) s_and_b32 s3, s2, vcc_lo s_waitcnt vmcnt(0) flat_store_b32 v[6:7], v3 dlc s_waitcnt_vscnt null, 0x0 s_and_saveexec_b32 s8, s3 s_cbranch_execz .LBB1_4 v_lshlrev_b64 v[3:4], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v3, s3, s6, v3 v_add_co_ci_u32_e64 v4, s3, s7, v4, s3 global_load_b32 v5, v[3:4], off .LBB1_4: s_or_b32 exec_lo, exec_lo, s8 s_cmp_lt_u32 s5, 2 s_mov_b32 s8, 0 s_waitcnt vmcnt(0) lgkmcnt(0) s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB1_9 s_mov_b32 s9, 1 s_mov_b64 s[6:7], src_shared_base s_set_inst_prefetch_distance 0x1 s_branch .LBB1_7 .p2align 6 .LBB1_6: s_or_b32 exec_lo, exec_lo, s6 s_sub_i32 s8, 1, s8 s_lshl_b32 s9, s9, 1 v_mad_u64_u32 v[6:7], null, s8, s5, v[0:1] s_cmp_ge_u32 s9, s5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v6, 2, 0 v_cmp_ne_u32_e64 s3, -1, v3 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v3, s3 v_cndmask_b32_e64 v7, 0, s7, s3 s_waitcnt lgkmcnt(0) flat_store_b32 v[6:7], v4 dlc s_waitcnt_vscnt null, 0x0 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB1_9 .LBB1_7: v_mad_u64_u32 v[3:4], null, s8, s5, v[0:1] s_mov_b32 s6, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v4, v3, 2, 0 v_cmp_ne_u32_e64 s3, -1, v4 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v4, s3 v_cndmask_b32_e64 v7, 0, s7, s3 flat_load_b32 v4, v[6:7] glc dlc s_waitcnt vmcnt(0) v_cmpx_le_u32_e64 s9, v0 s_cbranch_execz .LBB1_6 v_subrev_nc_u32_e32 v3, s9, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v3, 2, 0 v_cmp_ne_u32_e64 s3, -1, v3 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v3, s3 v_cndmask_b32_e64 v7, 0, s7, s3 flat_load_b32 v3, v[6:7] glc dlc s_waitcnt vmcnt(0) lgkmcnt(0) v_add_f32_e32 v4, v4, v3 s_branch .LBB1_6 .LBB1_9: s_set_inst_prefetch_distance 0x2 s_and_saveexec_b32 s3, vcc_lo s_cbranch_execz .LBB1_11 v_mad_u64_u32 v[3:4], null, s8, s5, v[0:1] s_mov_b64 s[6:7], src_shared_base v_lshlrev_b64 v[1:2], 2, v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v3, 2, 0 v_cmp_ne_u32_e32 vcc_lo, -1, v3 v_cndmask_b32_e32 v3, 0, v3, vcc_lo v_cndmask_b32_e64 v4, 0, s7, vcc_lo s_load_b64 s[6:7], s[0:1], 0x8 flat_load_b32 v3, v[3:4] glc dlc s_waitcnt vmcnt(0) lgkmcnt(0) v_add_co_u32 v1, vcc_lo, s6, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v2, vcc_lo global_store_b32 v[1:2], v3, off .LBB1_11: s_or_b32 exec_lo, exec_lo, s3 s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB1_13 v_mad_u64_u32 v[1:2], null, s8, s5, v[0:1] s_mov_b64 s[2:3], src_shared_base s_load_b64 s[0:1], s[0:1], 0x18 s_mov_b32 s5, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v0, v1, 2, 0 v_cmp_ne_u32_e32 vcc_lo, -1, v0 v_cndmask_b32_e32 v0, 0, v0, vcc_lo v_cndmask_b32_e64 v1, 0, s3, vcc_lo s_lshl_b64 s[2:3], s[4:5], 2 s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 flat_load_b32 v0, v[0:1] glc dlc s_waitcnt vmcnt(0) v_mov_b32_e32 v1, 0 s_addc_u32 s1, s1, s3 s_waitcnt lgkmcnt(0) v_add_f32_e32 v0, v5, v0 global_store_b32 v1, v0, s[0:1] .LBB1_13: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13hillis_steelePfS_iS_ .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 8 .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_end1: .size _Z13hillis_steelePfS_iS_, .Lfunc_end1-_Z13hillis_steelePfS_iS_ .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 temp .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: by_value - .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 - .offset: 144 .size: 4 .value_kind: hidden_dynamic_lds_size .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5adderPfS_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5adderPfS_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .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: 4 .value_kind: by_value - .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 - .offset: 152 .size: 4 .value_kind: hidden_dynamic_lds_size .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: _Z13hillis_steelePfS_iS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13hillis_steelePfS_iS_.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.
#include <hip/hip_runtime.h> #include <cstdio> using namespace std; __global__ void adder(float* arr, float* block_incrs, int n) { int tid = threadIdx.x; extern __shared__ float sum[]; int gtid = blockIdx.x * blockDim.x + threadIdx.x; if (tid == 0) sum[0] = block_incrs[blockIdx.x]; __syncthreads(); if (gtid < n) arr[gtid] += sum[0]; } __global__ void hillis_steele(float* g_idata, float* g_odata, int n, float* block_sums) { // all memory writes to be serviced immediately extern volatile __shared__ float temp[]; int tid = threadIdx.x; int gtid = blockIdx.x * blockDim.x + threadIdx.x; int block_size = blockDim.x; int pout = 0, pin = 1; float last_element; // load input into shared memory if (tid == 0 || gtid >= n) temp[tid] = 0; else temp[tid] = g_idata[gtid - 1]; if (tid == block_size - 1) if (gtid < n) last_element = g_idata[gtid]; else last_element = 0.0; // entire temp should've gotten populated __syncthreads(); // if (gtid == 5) { // std::printf("global : %f===============\n", g_idata[gtid]); // std::printf("temp : %f===============\n", temp[tid]); // } for (int offset = 1; offset < block_size; offset *= 2) { pout = 1 - pout; // swap double buffer indices pin = 1 - pout; if (tid >= offset) temp[pout * block_size + tid] = temp[pin * block_size + tid] + temp[pin * block_size + tid - offset]; else temp[pout * block_size + tid] = temp[pin * block_size + tid]; __syncthreads(); // I need this here before I start next iteration } if (gtid < n) g_odata[gtid] = temp[pout * block_size + tid]; if (tid == block_size - 1) block_sums[blockIdx.x] = last_element + temp[pout * block_size + tid]; } __host__ void scan(const float* in, float* out, unsigned int n, unsigned int threads_per_block) { int num_blocks = (n + threads_per_block - 1) / threads_per_block; // printf("num blocks : %d\n", num_blocks); float *din, *dout, *block_sums, *block_incrs, *dummy; hipMalloc((void**)&din, n * sizeof(float)); hipMalloc((void**)&dout, n * sizeof(float)); hipMallocManaged((void**)&block_sums, num_blocks * sizeof(float)); hipMallocManaged((void**)&block_incrs, num_blocks * sizeof(float)); hipMallocManaged((void**)&dummy, sizeof(float)); hipMemcpy(din, in, n * sizeof(float), hipMemcpyHostToHost); // // Only applicable if threads_per_block is a power of 2 // reduce_kernel<<<num_blocks, threads_per_block, sizeof(float) * // threads_per_block>>>(din, block_sums, n); hillis_steele<<<num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>(din, dout, n, block_sums); hipDeviceSynchronize(); // get the block increments (scan it once because of assumption) int new_num_blocks = (num_blocks + threads_per_block - 1) / threads_per_block; // will always be 1 hillis_steele<<<new_num_blocks, threads_per_block, 2 * threads_per_block * sizeof(float)>>>( block_sums, block_incrs, num_blocks, dummy); hipDeviceSynchronize(); // for (int i = 0; i < num_blocks; i++) // printf("%f ", block_incrs[i]); // printf("\n"); // add each block increment to each block adder<<<num_blocks, threads_per_block, sizeof(float)>>>(dout, block_incrs, n); hipDeviceSynchronize(); hipMemcpy(out, dout, n * sizeof(float), hipMemcpyDeviceToHost); hipFree(din); hipFree(dout); hipFree(block_sums); hipFree(block_incrs); hipFree(dummy); }
.text .file "scan.hip" .globl _Z20__device_stub__adderPfS_i # -- Begin function _Z20__device_stub__adderPfS_i .p2align 4, 0x90 .type _Z20__device_stub__adderPfS_i,@function _Z20__device_stub__adderPfS_i: # @_Z20__device_stub__adderPfS_i .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%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 $_Z5adderPfS_i, %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__adderPfS_i, .Lfunc_end0-_Z20__device_stub__adderPfS_i .cfi_endproc # -- End function .globl _Z28__device_stub__hillis_steelePfS_iS_ # -- Begin function _Z28__device_stub__hillis_steelePfS_iS_ .p2align 4, 0x90 .type _Z28__device_stub__hillis_steelePfS_iS_,@function _Z28__device_stub__hillis_steelePfS_iS_: # @_Z28__device_stub__hillis_steelePfS_iS_ .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 4(%rsp) movq %rcx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 4(%rsp), %rax movq %rax, 96(%rsp) leaq 56(%rsp), %rax movq %rax, 104(%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 $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z28__device_stub__hillis_steelePfS_iS_, .Lfunc_end1-_Z28__device_stub__hillis_steelePfS_iS_ .cfi_endproc # -- End function .globl _Z4scanPKfPfjj # -- Begin function _Z4scanPKfPfjj .p2align 4, 0x90 .type _Z4scanPKfPfjj,@function _Z4scanPKfPfjj: # @_Z4scanPKfPfjj .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 $184, %rsp .cfi_def_cfa_offset 240 .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 %ecx, %r15d movl %edx, %ecx movq %rsi, 176(%rsp) # 8-byte Spill movq %rdi, %r13 movabsq $4294967296, %rbx # imm = 0x100000000 leal (%r15,%rcx), %eax decl %eax xorl %edx, %edx divl %r15d movl %eax, %r12d movq %rcx, 152(%rsp) # 8-byte Spill movl %ecx, %r14d shlq $2, %r14 leaq 104(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 80(%rsp), %rdi movq %r14, %rsi callq hipMalloc movslq %r12d, %rbp shlq $2, %rbp leaq 96(%rsp), %rdi movq %rbp, %rsi movl $1, %edx callq hipMallocManaged leaq 88(%rsp), %rdi movq %rbp, %rsi movl $1, %edx callq hipMallocManaged leaq 160(%rsp), %rdi movl $4, %esi movl $1, %edx callq hipMallocManaged movq 104(%rsp), %rdi movq %r13, %rsi movq %r14, %rdx xorl %ecx, %ecx callq hipMemcpy leal (%r15,%r15), %ebp shlq $2, %rbp leaq (%r12,%rbx), %rdi movl %r15d, %r13d orq %rbx, %r13 movq %rdi, 168(%rsp) # 8-byte Spill movl $1, %esi movq %r13, %rdx movl $1, %ecx movq %rbp, %r8 xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movq 104(%rsp), %rax movq 80(%rsp), %rcx movq 96(%rsp), %rdx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq 152(%rsp), %rax # 8-byte Reload movl %eax, 4(%rsp) movq %rdx, 24(%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) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_2: callq hipDeviceSynchronize leal (%r15,%r12), %eax decl %eax xorl %edx, %edx divl %r15d # kill: def $eax killed $eax def $rax orq %rax, %rbx movq %rbx, %rdi movl $1, %esi movq %r13, %rdx movl $1, %ecx movq %rbp, %r8 xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 96(%rsp), %rax movq 88(%rsp), %rcx movq 160(%rsp), %rdx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movl %r12d, 4(%rsp) movq %rdx, 24(%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) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_4: callq hipDeviceSynchronize movl $4, %r8d movq 168(%rsp), %rdi # 8-byte Reload movl $1, %esi movq %r13, %rdx movl $1, %ecx xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_6 # %bb.5: movq 80(%rsp), %rax movq 88(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq 152(%rsp), %rax # 8-byte Reload movl %eax, 8(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 8(%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 112(%rsp), %r9 movl $_Z5adderPfS_i, %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 .LBB2_6: callq hipDeviceSynchronize movq 80(%rsp), %rsi movq 176(%rsp), %rdi # 8-byte Reload movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 104(%rsp), %rdi callq hipFree movq 80(%rsp), %rdi callq hipFree movq 96(%rsp), %rdi callq hipFree movq 88(%rsp), %rdi callq hipFree movq 160(%rsp), %rdi callq hipFree addq $184, %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_end2: .size _Z4scanPKfPfjj, .Lfunc_end2-_Z4scanPKfPfjj .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 .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5adderPfS_i, %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 $_Z13hillis_steelePfS_iS_, %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_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 _Z5adderPfS_i,@object # @_Z5adderPfS_i .section .rodata,"a",@progbits .globl _Z5adderPfS_i .p2align 3, 0x0 _Z5adderPfS_i: .quad _Z20__device_stub__adderPfS_i .size _Z5adderPfS_i, 8 .type _Z13hillis_steelePfS_iS_,@object # @_Z13hillis_steelePfS_iS_ .globl _Z13hillis_steelePfS_iS_ .p2align 3, 0x0 _Z13hillis_steelePfS_iS_: .quad _Z28__device_stub__hillis_steelePfS_iS_ .size _Z13hillis_steelePfS_iS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5adderPfS_i" .size .L__unnamed_1, 14 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z13hillis_steelePfS_iS_" .size .L__unnamed_2, 25 .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__adderPfS_i .addrsig_sym _Z28__device_stub__hillis_steelePfS_iS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5adderPfS_i .addrsig_sym _Z13hillis_steelePfS_iS_ .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 : _Z13hillis_steelePfS_iS_ .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 R9, SR_CTAID.X ; /* 0x0000000000097919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0040*/ BSSY B0, 0x100 ; /* 0x000000b000007945 */ /* 0x000fe20003800000 */ /*0050*/ S2R R10, SR_TID.X ; /* 0x00000000000a7919 */ /* 0x000e240000002100 */ /*0060*/ IMAD R0, R9, c[0x0][0x0], R10 ; /* 0x0000000009007a24 */ /* 0x001fca00078e020a */ /*0070*/ IMAD.WIDE R2, R0.reuse, R3, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x040fe200078e0203 */ /*0080*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fc80003f06270 */ /*0090*/ ISETP.EQ.OR P1, PT, R10, RZ, P0 ; /* 0x000000ff0a00720c */ /* 0x000fda0000722670 */ /*00a0*/ @P1 BRA 0xe0 ; /* 0x0000003000001947 */ /* 0x000fea0003800000 */ /*00b0*/ LDG.E R4, [R2.64+-0x4] ; /* 0xfffffc0402047981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ STS [R10.X4], R4 ; /* 0x000000040a007388 */ /* 0x0041e20000004800 */ /*00d0*/ BRA 0xf0 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*00e0*/ STS [R10.X4], RZ ; /* 0x000000ff0a007388 */ /* 0x0001e40000004800 */ /*00f0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0100*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff057624 */ /* 0x000fe200078e00ff */ /*0110*/ BSSY B0, 0x1b0 ; /* 0x0000009000007945 */ /* 0x000fe80003800000 */ /*0120*/ IADD3 R4, R5, -0x1, RZ ; /* 0xffffffff05047810 */ /* 0x001fc40007ffe0ff */ /*0130*/ ISETP.GE.AND P2, PT, R5, 0x2, PT ; /* 0x000000020500780c */ /* 0x000fe20003f46270 */ /*0140*/ HFMA2.MMA R5, -RZ, RZ, 0, 0 ; /* 0x00000000ff057435 */ /* 0x000fe200000001ff */ /*0150*/ ISETP.NE.AND P1, PT, R10, R4, PT ; /* 0x000000040a00720c */ /* 0x000fda0003f25270 */ /*0160*/ @P1 BRA 0x1a0 ; /* 0x0000003000001947 */ /* 0x000fea0003800000 */ /*0170*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe200078e00ff */ /*0180*/ @P0 BRA 0x1a0 ; /* 0x0000001000000947 */ /* 0x000fea0003800000 */ /*0190*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000164000c1e1900 */ /*01a0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*01b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*01c0*/ @!P2 BRA 0x2e0 ; /* 0x000001100000a947 */ /* 0x000fea0003800000 */ /*01d0*/ MOV R5, RZ ; /* 0x000000ff00057202 */ /* 0x000fe20000000f00 */ /*01e0*/ IMAD.MOV.U32 R2, RZ, RZ, 0x1 ; /* 0x00000001ff027424 */ /* 0x001fca00078e00ff */ /*01f0*/ ISETP.GE.AND P2, PT, R10, R2, PT ; /* 0x000000020a00720c */ /* 0x000fe20003f46270 */ /*0200*/ IMAD R3, R5.reuse, c[0x0][0x0], R10 ; /* 0x0000000005037a24 */ /* 0x040fe200078e020a */ /*0210*/ IADD3 R5, -R5, 0x1, RZ ; /* 0x0000000105057810 */ /* 0x000fca0007ffe1ff */ /*0220*/ IMAD R6, R5, c[0x0][0x0], R10 ; /* 0x0000000005067a24 */ /* 0x000fc800078e020a */ /*0230*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fe400078e00ff */ /*0240*/ @P2 IADD3 R7, R3, -R2, RZ ; /* 0x8000000203072210 */ /* 0x000fe40007ffe0ff */ /*0250*/ LDS R3, [R3.X4] ; /* 0x0000000003037984 */ /* 0x000fe20000004800 */ /*0260*/ SHF.L.U32 R2, R2, 0x1, RZ ; /* 0x0000000102027819 */ /* 0x000fc600000006ff */ /*0270*/ @P2 LDS R8, [R7.X4] ; /* 0x0000000007082984 */ /* 0x000e240000004800 */ /*0280*/ @P2 FADD R8, R3, R8 ; /* 0x0000000803082221 */ /* 0x001fca0000000000 */ /*0290*/ @P2 STS [R6], R8 ; /* 0x0000000806002388 */ /* 0x0001e80000000800 */ /*02a0*/ @!P2 STS [R6], R3 ; /* 0x000000030600a388 */ /* 0x0001e80000000800 */ /*02b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*02c0*/ ISETP.GE.AND P2, PT, R2, c[0x0][0x0], PT ; /* 0x0000000002007a0c */ /* 0x000fda0003f46270 */ /*02d0*/ @!P2 BRA 0x1f0 ; /* 0xffffff100000a947 */ /* 0x001fea000383ffff */ /*02e0*/ IMAD R6, R5, c[0x0][0x0], R10 ; /* 0x0000000005067a24 */ /* 0x000fe200078e020a */ /*02f0*/ @!P0 MOV R3, 0x4 ; /* 0x0000000400038802 */ /* 0x001fc60000000f00 */ /*0300*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fe400078e00ff */ /*0310*/ @!P0 IMAD.WIDE R2, R0, R3, c[0x0][0x168] ; /* 0x00005a0000028625 */ /* 0x000fc600078e0203 */ /*0320*/ @!P0 LDS R5, [R6] ; /* 0x0000000006058984 */ /* 0x000e280000000800 */ /*0330*/ @!P0 STG.E [R2.64], R5 ; /* 0x0000000502008986 */ /* 0x0011e2000c101904 */ /*0340*/ @P1 EXIT ; /* 0x000000000000194d */ /* 0x000fea0003800000 */ /*0350*/ LDS R3, [R6] ; /* 0x0000000006037984 */ /* 0x001e220000000800 */ /*0360*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fe400078e00ff */ /*0370*/ FADD R5, R3, R4 ; /* 0x0000000403057221 */ /* 0x021fe40000000000 */ /*0380*/ IMAD.WIDE.U32 R2, R9, R2, c[0x0][0x178] ; /* 0x00005e0009027625 */ /* 0x000fca00078e0002 */ /*0390*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*03a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*03b0*/ BRA 0x3b0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*03c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0400*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0410*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0420*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0430*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0440*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0450*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0460*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0470*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z5adderPfS_i .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 R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ BSSY B0, 0xe0 ; /* 0x000000a000007945 */ /* 0x000fe40003800000 */ /*0040*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e620000002500 */ /*0050*/ ISETP.NE.AND P0, PT, R3, RZ, PT ; /* 0x000000ff0300720c */ /* 0x001fe20003f05270 */ /*0060*/ IMAD R0, R2, c[0x0][0x0], R3 ; /* 0x0000000002007a24 */ /* 0x002fca00078e0203 */ /*0070*/ ISETP.GE.AND P1, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fce0003f26270 */ /*0080*/ @P0 BRA 0xd0 ; /* 0x0000004000000947 */ /* 0x000fea0003800000 */ /*0090*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fd400000001ff */ /*00a0*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fcc00078e0003 */ /*00b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ STS [RZ], R2 ; /* 0x00000002ff007388 */ /* 0x0041e40000000800 */ /*00d0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*00e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*00f0*/ @P1 EXIT ; /* 0x000000000000194d */ /* 0x000fea0003800000 */ /*0100*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fca0000000f00 */ /*0110*/ IMAD.WIDE R2, R0, R3, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x001fe400078e0203 */ /*0120*/ LDS R0, [RZ] ; /* 0x00000000ff007984 */ /* 0x000e280000000800 */ /*0130*/ LDG.E R5, [R2.64] ; /* 0x0000000402057981 */ /* 0x000e24000c1e1900 */ /*0140*/ FADD R5, R0, R5 ; /* 0x0000000500057221 */ /* 0x001fca0000000000 */ /*0150*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 _Z5adderPfS_i .globl _Z5adderPfS_i .p2align 8 .type _Z5adderPfS_i,@function _Z5adderPfS_i: s_load_b32 s4, s[0:1], 0x24 s_mov_b32 s2, s15 s_mov_b32 s3, 0 s_mov_b32 s5, exec_lo v_cmpx_eq_u32_e32 0, v0 s_cbranch_execz .LBB0_2 s_load_b64 s[6:7], s[0:1], 0x8 s_lshl_b64 s[8:9], s[2:3], 2 s_waitcnt lgkmcnt(0) s_add_u32 s6, s6, s8 s_addc_u32 s7, s7, s9 s_load_b32 s3, s[6:7], 0x0 s_waitcnt lgkmcnt(0) v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, s3 ds_store_b32 v1, v2 .LBB0_2: s_or_b32 exec_lo, exec_lo, s5 s_load_b32 s3, s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s4, 0xffff, s4 s_delay_alu instid0(SALU_CYCLE_1) v_mad_u64_u32 v[1:2], null, s2, s4, v[0:1] s_barrier buffer_gl0_inv s_mov_b32 s2, exec_lo v_cmpx_gt_i32_e64 s3, v1 s_cbranch_execz .LBB0_4 s_load_b64 s[0:1], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 v_mov_b32_e32 v3, 0 s_delay_alu instid0(VALU_DEP_2) v_lshlrev_b64 v[0:1], 2, v[1:2] ds_load_b32 v3, v3 s_waitcnt lgkmcnt(0) 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[0:1], off s_waitcnt vmcnt(0) v_add_f32_e32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5adderPfS_i .amdhsa_group_segment_fixed_size 0 .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 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 _Z5adderPfS_i, .Lfunc_end0-_Z5adderPfS_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z13hillis_steelePfS_iS_ .globl _Z13hillis_steelePfS_iS_ .p2align 8 .type _Z13hillis_steelePfS_iS_,@function _Z13hillis_steelePfS_iS_: s_clause 0x2 s_load_b32 s2, s[0:1], 0x2c s_load_b32 s3, s[0:1], 0x10 s_load_b64 s[6:7], s[0:1], 0x0 s_mov_b32 s4, s15 v_mov_b32_e32 v5, 0 v_mov_b32_e32 v3, 0 s_waitcnt lgkmcnt(0) s_and_b32 s5, s2, 0xffff v_cmp_ne_u32_e64 s2, 0, v0 v_mad_u64_u32 v[1:2], null, s4, s5, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_cmp_gt_i32_e32 vcc_lo, s3, v1 v_ashrrev_i32_e32 v2, 31, v1 s_and_b32 s2, s2, vcc_lo s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB1_2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 2, v[1:2] v_add_co_u32 v3, s2, s6, v3 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v4, s2, s7, v4, s2 global_load_b32 v3, v[3:4], off offset:-4 .LBB1_2: s_or_b32 exec_lo, exec_lo, s3 v_lshl_add_u32 v4, v0, 2, 0 s_add_i32 s2, s5, -1 s_mov_b64 s[8:9], src_shared_base v_cmp_eq_u32_e64 s2, s2, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_ne_u32_e64 s3, -1, v4 v_cndmask_b32_e64 v6, 0, v4, s3 v_cndmask_b32_e64 v7, 0, s9, s3 s_delay_alu instid0(VALU_DEP_4) s_and_b32 s3, s2, vcc_lo s_waitcnt vmcnt(0) flat_store_b32 v[6:7], v3 dlc s_waitcnt_vscnt null, 0x0 s_and_saveexec_b32 s8, s3 s_cbranch_execz .LBB1_4 v_lshlrev_b64 v[3:4], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v3, s3, s6, v3 v_add_co_ci_u32_e64 v4, s3, s7, v4, s3 global_load_b32 v5, v[3:4], off .LBB1_4: s_or_b32 exec_lo, exec_lo, s8 s_cmp_lt_u32 s5, 2 s_mov_b32 s8, 0 s_waitcnt vmcnt(0) lgkmcnt(0) s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB1_9 s_mov_b32 s9, 1 s_mov_b64 s[6:7], src_shared_base s_set_inst_prefetch_distance 0x1 s_branch .LBB1_7 .p2align 6 .LBB1_6: s_or_b32 exec_lo, exec_lo, s6 s_sub_i32 s8, 1, s8 s_lshl_b32 s9, s9, 1 v_mad_u64_u32 v[6:7], null, s8, s5, v[0:1] s_cmp_ge_u32 s9, s5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v6, 2, 0 v_cmp_ne_u32_e64 s3, -1, v3 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v3, s3 v_cndmask_b32_e64 v7, 0, s7, s3 s_waitcnt lgkmcnt(0) flat_store_b32 v[6:7], v4 dlc s_waitcnt_vscnt null, 0x0 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB1_9 .LBB1_7: v_mad_u64_u32 v[3:4], null, s8, s5, v[0:1] s_mov_b32 s6, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v4, v3, 2, 0 v_cmp_ne_u32_e64 s3, -1, v4 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v4, s3 v_cndmask_b32_e64 v7, 0, s7, s3 flat_load_b32 v4, v[6:7] glc dlc s_waitcnt vmcnt(0) v_cmpx_le_u32_e64 s9, v0 s_cbranch_execz .LBB1_6 v_subrev_nc_u32_e32 v3, s9, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v3, 2, 0 v_cmp_ne_u32_e64 s3, -1, v3 s_delay_alu instid0(VALU_DEP_1) v_cndmask_b32_e64 v6, 0, v3, s3 v_cndmask_b32_e64 v7, 0, s7, s3 flat_load_b32 v3, v[6:7] glc dlc s_waitcnt vmcnt(0) lgkmcnt(0) v_add_f32_e32 v4, v4, v3 s_branch .LBB1_6 .LBB1_9: s_set_inst_prefetch_distance 0x2 s_and_saveexec_b32 s3, vcc_lo s_cbranch_execz .LBB1_11 v_mad_u64_u32 v[3:4], null, s8, s5, v[0:1] s_mov_b64 s[6:7], src_shared_base v_lshlrev_b64 v[1:2], 2, v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v3, v3, 2, 0 v_cmp_ne_u32_e32 vcc_lo, -1, v3 v_cndmask_b32_e32 v3, 0, v3, vcc_lo v_cndmask_b32_e64 v4, 0, s7, vcc_lo s_load_b64 s[6:7], s[0:1], 0x8 flat_load_b32 v3, v[3:4] glc dlc s_waitcnt vmcnt(0) lgkmcnt(0) v_add_co_u32 v1, vcc_lo, s6, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v2, vcc_lo global_store_b32 v[1:2], v3, off .LBB1_11: s_or_b32 exec_lo, exec_lo, s3 s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB1_13 v_mad_u64_u32 v[1:2], null, s8, s5, v[0:1] s_mov_b64 s[2:3], src_shared_base s_load_b64 s[0:1], s[0:1], 0x18 s_mov_b32 s5, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshl_add_u32 v0, v1, 2, 0 v_cmp_ne_u32_e32 vcc_lo, -1, v0 v_cndmask_b32_e32 v0, 0, v0, vcc_lo v_cndmask_b32_e64 v1, 0, s3, vcc_lo s_lshl_b64 s[2:3], s[4:5], 2 s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 flat_load_b32 v0, v[0:1] glc dlc s_waitcnt vmcnt(0) v_mov_b32_e32 v1, 0 s_addc_u32 s1, s1, s3 s_waitcnt lgkmcnt(0) v_add_f32_e32 v0, v5, v0 global_store_b32 v1, v0, s[0:1] .LBB1_13: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13hillis_steelePfS_iS_ .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 8 .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_end1: .size _Z13hillis_steelePfS_iS_, .Lfunc_end1-_Z13hillis_steelePfS_iS_ .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 temp .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: by_value - .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 - .offset: 144 .size: 4 .value_kind: hidden_dynamic_lds_size .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5adderPfS_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5adderPfS_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .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: 4 .value_kind: by_value - .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 - .offset: 152 .size: 4 .value_kind: hidden_dynamic_lds_size .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: _Z13hillis_steelePfS_iS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13hillis_steelePfS_iS_.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_0009604a_00000000-6_scan.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 _Z27__device_stub__Z5adderPfS_iPfS_i .type _Z27__device_stub__Z5adderPfS_iPfS_i, @function _Z27__device_stub__Z5adderPfS_iPfS_i: .LFB2082: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 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 _Z5adderPfS_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z27__device_stub__Z5adderPfS_iPfS_i, .-_Z27__device_stub__Z5adderPfS_iPfS_i .globl _Z5adderPfS_i .type _Z5adderPfS_i, @function _Z5adderPfS_i: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z5adderPfS_iPfS_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5adderPfS_i, .-_Z5adderPfS_i .globl _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ .type _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_, @function _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_: .LFB2084: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%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 12(%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 .L15 .L11: movq 136(%rsp), %rax subq %fs:40, %rax jne .L16 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z13hillis_steelePfS_iS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_, .-_Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ .globl _Z13hillis_steelePfS_iS_ .type _Z13hillis_steelePfS_iS_, @function _Z13hillis_steelePfS_iS_: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z13hillis_steelePfS_iS_, .-_Z13hillis_steelePfS_iS_ .globl _Z4scanPKfPfjj .type _Z4scanPKfPfjj, @function _Z4scanPKfPfjj: .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 $104, %rsp .cfi_def_cfa_offset 160 movq %rdi, (%rsp) movq %rsi, 8(%rsp) movl %edx, %r14d movl %ecx, %ebx movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leal -1(%rcx), %r12d leal (%rdx,%r12), %eax movl $0, %edx divl %ecx movl %eax, %ebp movl %r14d, %r13d salq $2, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT leaq 32(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT movslq %ebp, %r15 salq $2, %r15 leaq 40(%rsp), %rdi movl $1, %edx movq %r15, %rsi call cudaMallocManaged@PLT leaq 48(%rsp), %rdi movl $1, %edx movq %r15, %rsi call cudaMallocManaged@PLT leaq 56(%rsp), %rdi movl $1, %edx movl $4, %esi call cudaMallocManaged@PLT movl $0, %ecx movq %r13, %rdx movq (%rsp), %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT leal (%rbx,%rbx), %r15d salq $2, %r15 movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl %ebp, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movq %r15, %r8 movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L20: call cudaDeviceSynchronize@PLT movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leal 0(%rbp,%r12), %eax movl $0, %edx divl %ebx movl %eax, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movq %r15, %r8 movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L26 .L21: call cudaDeviceSynchronize@PLT movl %ebx, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl %ebp, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $0, %r9d movl $4, %r8d movq 76(%rsp), %rdx movl $1, %ecx movq 64(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L27 .L22: call cudaDeviceSynchronize@PLT movl $2, %ecx movq %r13, %rdx movq 32(%rsp), %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@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 .L28 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 .L25: .cfi_restore_state movq 40(%rsp), %rcx movl %r14d, %edx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ jmp .L20 .L26: movq 56(%rsp), %rcx movl %ebp, %edx movq 48(%rsp), %rsi movq 40(%rsp), %rdi call _Z38__device_stub__Z13hillis_steelePfS_iS_PfS_iS_ jmp .L21 .L27: movl %r14d, %edx movq 48(%rsp), %rsi movq 32(%rsp), %rdi call _Z27__device_stub__Z5adderPfS_iPfS_i jmp .L22 .L28: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z4scanPKfPfjj, .-_Z4scanPKfPfjj .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z13hillis_steelePfS_iS_" .LC1: .string "_Z5adderPfS_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .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 _Z13hillis_steelePfS_iS_(%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 _Z5adderPfS_i(%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 .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 .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 "scan.hip" .globl _Z20__device_stub__adderPfS_i # -- Begin function _Z20__device_stub__adderPfS_i .p2align 4, 0x90 .type _Z20__device_stub__adderPfS_i,@function _Z20__device_stub__adderPfS_i: # @_Z20__device_stub__adderPfS_i .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%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 $_Z5adderPfS_i, %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__adderPfS_i, .Lfunc_end0-_Z20__device_stub__adderPfS_i .cfi_endproc # -- End function .globl _Z28__device_stub__hillis_steelePfS_iS_ # -- Begin function _Z28__device_stub__hillis_steelePfS_iS_ .p2align 4, 0x90 .type _Z28__device_stub__hillis_steelePfS_iS_,@function _Z28__device_stub__hillis_steelePfS_iS_: # @_Z28__device_stub__hillis_steelePfS_iS_ .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 4(%rsp) movq %rcx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 4(%rsp), %rax movq %rax, 96(%rsp) leaq 56(%rsp), %rax movq %rax, 104(%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 $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z28__device_stub__hillis_steelePfS_iS_, .Lfunc_end1-_Z28__device_stub__hillis_steelePfS_iS_ .cfi_endproc # -- End function .globl _Z4scanPKfPfjj # -- Begin function _Z4scanPKfPfjj .p2align 4, 0x90 .type _Z4scanPKfPfjj,@function _Z4scanPKfPfjj: # @_Z4scanPKfPfjj .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 $184, %rsp .cfi_def_cfa_offset 240 .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 %ecx, %r15d movl %edx, %ecx movq %rsi, 176(%rsp) # 8-byte Spill movq %rdi, %r13 movabsq $4294967296, %rbx # imm = 0x100000000 leal (%r15,%rcx), %eax decl %eax xorl %edx, %edx divl %r15d movl %eax, %r12d movq %rcx, 152(%rsp) # 8-byte Spill movl %ecx, %r14d shlq $2, %r14 leaq 104(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 80(%rsp), %rdi movq %r14, %rsi callq hipMalloc movslq %r12d, %rbp shlq $2, %rbp leaq 96(%rsp), %rdi movq %rbp, %rsi movl $1, %edx callq hipMallocManaged leaq 88(%rsp), %rdi movq %rbp, %rsi movl $1, %edx callq hipMallocManaged leaq 160(%rsp), %rdi movl $4, %esi movl $1, %edx callq hipMallocManaged movq 104(%rsp), %rdi movq %r13, %rsi movq %r14, %rdx xorl %ecx, %ecx callq hipMemcpy leal (%r15,%r15), %ebp shlq $2, %rbp leaq (%r12,%rbx), %rdi movl %r15d, %r13d orq %rbx, %r13 movq %rdi, 168(%rsp) # 8-byte Spill movl $1, %esi movq %r13, %rdx movl $1, %ecx movq %rbp, %r8 xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movq 104(%rsp), %rax movq 80(%rsp), %rcx movq 96(%rsp), %rdx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq 152(%rsp), %rax # 8-byte Reload movl %eax, 4(%rsp) movq %rdx, 24(%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) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_2: callq hipDeviceSynchronize leal (%r15,%r12), %eax decl %eax xorl %edx, %edx divl %r15d # kill: def $eax killed $eax def $rax orq %rax, %rbx movq %rbx, %rdi movl $1, %esi movq %r13, %rdx movl $1, %ecx movq %rbp, %r8 xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 96(%rsp), %rax movq 88(%rsp), %rcx movq 160(%rsp), %rdx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movl %r12d, 4(%rsp) movq %rdx, 24(%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) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z13hillis_steelePfS_iS_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_4: callq hipDeviceSynchronize movl $4, %r8d movq 168(%rsp), %rdi # 8-byte Reload movl $1, %esi movq %r13, %rdx movl $1, %ecx xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_6 # %bb.5: movq 80(%rsp), %rax movq 88(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) movq 152(%rsp), %rax # 8-byte Reload movl %eax, 8(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 8(%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 112(%rsp), %r9 movl $_Z5adderPfS_i, %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 .LBB2_6: callq hipDeviceSynchronize movq 80(%rsp), %rsi movq 176(%rsp), %rdi # 8-byte Reload movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 104(%rsp), %rdi callq hipFree movq 80(%rsp), %rdi callq hipFree movq 96(%rsp), %rdi callq hipFree movq 88(%rsp), %rdi callq hipFree movq 160(%rsp), %rdi callq hipFree addq $184, %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_end2: .size _Z4scanPKfPfjj, .Lfunc_end2-_Z4scanPKfPfjj .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 .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5adderPfS_i, %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 $_Z13hillis_steelePfS_iS_, %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_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 _Z5adderPfS_i,@object # @_Z5adderPfS_i .section .rodata,"a",@progbits .globl _Z5adderPfS_i .p2align 3, 0x0 _Z5adderPfS_i: .quad _Z20__device_stub__adderPfS_i .size _Z5adderPfS_i, 8 .type _Z13hillis_steelePfS_iS_,@object # @_Z13hillis_steelePfS_iS_ .globl _Z13hillis_steelePfS_iS_ .p2align 3, 0x0 _Z13hillis_steelePfS_iS_: .quad _Z28__device_stub__hillis_steelePfS_iS_ .size _Z13hillis_steelePfS_iS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5adderPfS_i" .size .L__unnamed_1, 14 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z13hillis_steelePfS_iS_" .size .L__unnamed_2, 25 .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__adderPfS_i .addrsig_sym _Z28__device_stub__hillis_steelePfS_iS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5adderPfS_i .addrsig_sym _Z13hillis_steelePfS_iS_ .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 <cuda.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
code for sm_80 Function : _Z10KernelTestPiS_S_i .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_CTAID.X ; /* 0x0000000000067919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R6, R6, c[0x0][0x0], R3 ; /* 0x0000000006067a24 */ /* 0x001fca00078e0203 */ /*0060*/ IMAD.WIDE.U32 R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x000fc800078e0007 */ /*0070*/ IMAD.WIDE.U32 R4, R6.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x0c0fe400078e0007 */ /*0080*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0090*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*00a0*/ IMAD.WIDE.U32 R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fe200078e0007 */ /*00b0*/ IADD3 R9, R2, R5, RZ ; /* 0x0000000502097210 */ /* 0x004fca0007ffe0ff */ /*00c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*00d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00e0*/ BRA 0xe0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 <cuda.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
.file "tmpxft_000d0301_00000000-6_test_kernel.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 .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .weak _Z10KernelTestPiS_S_i .type _Z10KernelTestPiS_S_i, @function _Z10KernelTestPiS_S_i: .LFB2053: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax movq %rdi, 8(%rsp) movq %rsi, 16(%rsp) movq %rdx, 24(%rsp) movl %ecx, 4(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 24(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq _Z10KernelTestPiS_S_i(%rip), %rax movq %rax, _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f(%rip) 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 _Z10KernelTestPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z10KernelTestPiS_S_i, .-_Z10KernelTestPiS_S_i .text .globl runKernel .type runKernel, @function runKernel: .LFB2027: .cfi_startproc endbr64 pushq %r13 .cfi_def_cfa_offset 16 .cfi_offset 13, -16 pushq %r12 .cfi_def_cfa_offset 24 .cfi_offset 12, -24 pushq %rbp .cfi_def_cfa_offset 32 .cfi_offset 6, -32 pushq %rbx .cfi_def_cfa_offset 40 .cfi_offset 3, -40 subq $184, %rsp .cfi_def_cfa_offset 224 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movl %ecx, %ebx movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax movl $1, 60(%rsp) movl %ecx, %eax shrl $6, %eax movl %eax, 68(%rsp) movl $1, 72(%rsp) movl $64, 56(%rsp) movl $0, %r9d movl $0, %r8d movq 56(%rsp), %rdx movl $1, %ecx movq 68(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L9: movq 168(%rsp), %rax subq %fs:40, %rax jne .L15 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %rbp .cfi_def_cfa_offset 24 popq %r12 .cfi_def_cfa_offset 16 popq %r13 .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state movq %r13, 16(%rsp) movq %r12, 24(%rsp) movq %rbp, 32(%rsp) movl %ebx, 12(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 32(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%rsp), %rax movq %rax, 152(%rsp) leaq _Z10KernelTestPiS_S_i(%rip), %rax movq %rax, _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f(%rip) 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 48(%rsp), %rcx leaq 40(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax jne .L9 pushq 48(%rsp) .cfi_def_cfa_offset 232 pushq 48(%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 _Z10KernelTestPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L9 .L15: call __stack_chk_fail@PLT .cfi_endproc .LFE2027: .size runKernel, .-runKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10KernelTestPiS_S_i" .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 _Z10KernelTestPiS_S_i(%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 .weak _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f .section .bss._ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f,"awG",@nobits,_ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f,comdat .align 8 .type _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f, @gnu_unique_object .size _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f, 8 _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f: .zero 8 .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 <cuda.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
#include <hip/hip_runtime.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .protected _Z10KernelTestPiS_S_i .globl _Z10KernelTestPiS_S_i .p2align 8 .type _Z10KernelTestPiS_S_i,@function _Z10KernelTestPiS_S_i: s_clause 0x2 s_load_b32 s2, s[0:1], 0x2c s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x10 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] v_mov_b32_e32 v2, 0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s4, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(0) v_add_nc_u32_e32 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 _Z10KernelTestPiS_S_i .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 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 .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .Lfunc_end0: .size _Z10KernelTestPiS_S_i, .Lfunc_end0-_Z10KernelTestPiS_S_i .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: 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: _Z10KernelTestPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10KernelTestPiS_S_i.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 <hip/hip_runtime.h> inline __global__ void KernelTest(int * a, int * b, int * res, int size) { // Calcul de l'indice du tableau unsigned int x = blockIdx.x*blockDim.x + threadIdx.x; res[x] = a[x] + b[x]; } extern "C" void runKernel(int * a, int * b, int * res, int size) { dim3 dimBlock(64, 1, 1); dim3 dimGrid(size / dimBlock.x, 1, 1); KernelTest<<< dimGrid, dimBlock, 0 >>>( a, b, res, size); }
.text .file "test_kernel.hip" .globl runKernel # -- Begin function runKernel .p2align 4, 0x90 .type runKernel,@function runKernel: # @runKernel .cfi_startproc # %bb.0: 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 subq $120, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl %ecx, %ebx movq %rdx, %r14 movq %rsi, %r15 movq %rdi, %r12 movl %ecx, %edi shrl $6, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $64, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_2 # %bb.1: movq %r12, 72(%rsp) movq %r15, 64(%rsp) movq %r14, 56(%rsp) movl %ebx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z10KernelTestPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_2: addq $120, %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 retq .Lfunc_end0: .size runKernel, .Lfunc_end0-runKernel .cfi_endproc # -- End function .section .text._Z25__device_stub__KernelTestPiS_S_i,"axG",@progbits,_Z25__device_stub__KernelTestPiS_S_i,comdat .weak _Z25__device_stub__KernelTestPiS_S_i # -- Begin function _Z25__device_stub__KernelTestPiS_S_i .p2align 4, 0x90 .type _Z25__device_stub__KernelTestPiS_S_i,@function _Z25__device_stub__KernelTestPiS_S_i: # @_Z25__device_stub__KernelTestPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z10KernelTestPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z25__device_stub__KernelTestPiS_S_i, .Lfunc_end1-_Z25__device_stub__KernelTestPiS_S_i .cfi_endproc # -- End function .text .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 $_Z10KernelTestPiS_S_i, %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 _Z10KernelTestPiS_S_i,@object # @_Z10KernelTestPiS_S_i .section .rodata._Z10KernelTestPiS_S_i,"aG",@progbits,_Z10KernelTestPiS_S_i,comdat .weak _Z10KernelTestPiS_S_i .p2align 3, 0x0 _Z10KernelTestPiS_S_i: .quad _Z25__device_stub__KernelTestPiS_S_i .size _Z10KernelTestPiS_S_i, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z10KernelTestPiS_S_i" .size .L__unnamed_1, 22 .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 _Z25__device_stub__KernelTestPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10KernelTestPiS_S_i .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 : _Z10KernelTestPiS_S_i .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_CTAID.X ; /* 0x0000000000067919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R6, R6, c[0x0][0x0], R3 ; /* 0x0000000006067a24 */ /* 0x001fca00078e0203 */ /*0060*/ IMAD.WIDE.U32 R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x000fc800078e0007 */ /*0070*/ IMAD.WIDE.U32 R4, R6.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x0c0fe400078e0007 */ /*0080*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0090*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*00a0*/ IMAD.WIDE.U32 R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fe200078e0007 */ /*00b0*/ IADD3 R9, R2, R5, RZ ; /* 0x0000000502097210 */ /* 0x004fca0007ffe0ff */ /*00c0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*00d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00e0*/ BRA 0xe0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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" .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .protected _Z10KernelTestPiS_S_i .globl _Z10KernelTestPiS_S_i .p2align 8 .type _Z10KernelTestPiS_S_i,@function _Z10KernelTestPiS_S_i: s_clause 0x2 s_load_b32 s2, s[0:1], 0x2c s_load_b128 s[4:7], s[0:1], 0x0 s_load_b64 s[0:1], s[0:1], 0x10 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] v_mov_b32_e32 v2, 0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s4, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(0) v_add_nc_u32_e32 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 _Z10KernelTestPiS_S_i .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 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 .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .Lfunc_end0: .size _Z10KernelTestPiS_S_i, .Lfunc_end0-_Z10KernelTestPiS_S_i .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: 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: _Z10KernelTestPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10KernelTestPiS_S_i.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_000d0301_00000000-6_test_kernel.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 .section .text._Z10KernelTestPiS_S_i,"axG",@progbits,_Z10KernelTestPiS_S_i,comdat .weak _Z10KernelTestPiS_S_i .type _Z10KernelTestPiS_S_i, @function _Z10KernelTestPiS_S_i: .LFB2053: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax movq %rdi, 8(%rsp) movq %rsi, 16(%rsp) movq %rdx, 24(%rsp) movl %ecx, 4(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 24(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq _Z10KernelTestPiS_S_i(%rip), %rax movq %rax, _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f(%rip) 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 _Z10KernelTestPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z10KernelTestPiS_S_i, .-_Z10KernelTestPiS_S_i .text .globl runKernel .type runKernel, @function runKernel: .LFB2027: .cfi_startproc endbr64 pushq %r13 .cfi_def_cfa_offset 16 .cfi_offset 13, -16 pushq %r12 .cfi_def_cfa_offset 24 .cfi_offset 12, -24 pushq %rbp .cfi_def_cfa_offset 32 .cfi_offset 6, -32 pushq %rbx .cfi_def_cfa_offset 40 .cfi_offset 3, -40 subq $184, %rsp .cfi_def_cfa_offset 224 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movl %ecx, %ebx movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax movl $1, 60(%rsp) movl %ecx, %eax shrl $6, %eax movl %eax, 68(%rsp) movl $1, 72(%rsp) movl $64, 56(%rsp) movl $0, %r9d movl $0, %r8d movq 56(%rsp), %rdx movl $1, %ecx movq 68(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L9: movq 168(%rsp), %rax subq %fs:40, %rax jne .L15 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %rbp .cfi_def_cfa_offset 24 popq %r12 .cfi_def_cfa_offset 16 popq %r13 .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state movq %r13, 16(%rsp) movq %r12, 24(%rsp) movq %rbp, 32(%rsp) movl %ebx, 12(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 32(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%rsp), %rax movq %rax, 152(%rsp) leaq _Z10KernelTestPiS_S_i(%rip), %rax movq %rax, _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f(%rip) 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 48(%rsp), %rcx leaq 40(%rsp), %rdx leaq 92(%rsp), %rsi leaq 80(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax jne .L9 pushq 48(%rsp) .cfi_def_cfa_offset 232 pushq 48(%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 _Z10KernelTestPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L9 .L15: call __stack_chk_fail@PLT .cfi_endproc .LFE2027: .size runKernel, .-runKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10KernelTestPiS_S_i" .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 _Z10KernelTestPiS_S_i(%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 .weak _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f .section .bss._ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f,"awG",@nobits,_ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f,comdat .align 8 .type _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f, @gnu_unique_object .size _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f, 8 _ZZ35__device_stub__Z10KernelTestPiS_S_iPiS_S_iE3__f: .zero 8 .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_kernel.hip" .globl runKernel # -- Begin function runKernel .p2align 4, 0x90 .type runKernel,@function runKernel: # @runKernel .cfi_startproc # %bb.0: 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 subq $120, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl %ecx, %ebx movq %rdx, %r14 movq %rsi, %r15 movq %rdi, %r12 movl %ecx, %edi shrl $6, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $64, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_2 # %bb.1: movq %r12, 72(%rsp) movq %r15, 64(%rsp) movq %r14, 56(%rsp) movl %ebx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z10KernelTestPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_2: addq $120, %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 retq .Lfunc_end0: .size runKernel, .Lfunc_end0-runKernel .cfi_endproc # -- End function .section .text._Z25__device_stub__KernelTestPiS_S_i,"axG",@progbits,_Z25__device_stub__KernelTestPiS_S_i,comdat .weak _Z25__device_stub__KernelTestPiS_S_i # -- Begin function _Z25__device_stub__KernelTestPiS_S_i .p2align 4, 0x90 .type _Z25__device_stub__KernelTestPiS_S_i,@function _Z25__device_stub__KernelTestPiS_S_i: # @_Z25__device_stub__KernelTestPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z10KernelTestPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z25__device_stub__KernelTestPiS_S_i, .Lfunc_end1-_Z25__device_stub__KernelTestPiS_S_i .cfi_endproc # -- End function .text .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 $_Z10KernelTestPiS_S_i, %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 _Z10KernelTestPiS_S_i,@object # @_Z10KernelTestPiS_S_i .section .rodata._Z10KernelTestPiS_S_i,"aG",@progbits,_Z10KernelTestPiS_S_i,comdat .weak _Z10KernelTestPiS_S_i .p2align 3, 0x0 _Z10KernelTestPiS_S_i: .quad _Z25__device_stub__KernelTestPiS_S_i .size _Z10KernelTestPiS_S_i, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z10KernelTestPiS_S_i" .size .L__unnamed_1, 22 .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 _Z25__device_stub__KernelTestPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10KernelTestPiS_S_i .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> #include <stdlib.h> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); cudaMalloc((void**) &A_d, size); cudaMalloc((void**) &B_d, size); cudaMalloc((void**) &C_d, size); cudaMemcpy(A_d, A_h, size, cudaMemcpyHostToDevice); cudaMemcpy(B_d, B_h, size, cudaMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); cudaFree(A_d); cudaFree(B_d); cudaFree(C_d); return 0; }
code for sm_80 Function : _Z13sumarMatricesPiS_S_i .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 R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0030*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000e680000002600 */ /*0040*/ S2R R5, SR_TID.Y ; /* 0x0000000000057919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R3, R2, c[0x0][0x4], R5 ; /* 0x0000010002037a24 */ /* 0x002fca00078e0205 */ /*0080*/ ISETP.GE.OR P0, PT, R3, c[0x0][0x178], P0 ; /* 0x00005e0003007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*00b0*/ IMAD R0, R3, c[0x0][0x178], R0 ; /* 0x00005e0003007a24 */ /* 0x000fe200078e0200 */ /*00c0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*00d0*/ IMAD.WIDE R4, R0, R7, c[0x0][0x170] ; /* 0x00005c0000047625 */ /* 0x000fc800078e0207 */ /*00e0*/ IMAD.WIDE R2, R0.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0000027625 */ /* 0x0c0fe400078e0207 */ /*00f0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*0100*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0110*/ IMAD.WIDE R6, R0, R7, c[0x0][0x160] ; /* 0x0000580000067625 */ /* 0x000fe200078e0207 */ /*0120*/ IADD3 R9, R4, R3, RZ ; /* 0x0000000304097210 */ /* 0x004fca0007ffe0ff */ /*0130*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*0140*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0150*/ BRA 0x150; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ .......... Function : _Z19multiplicarMatricesPiS_S_i .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 R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e280000002100 */ /*0030*/ S2R R0, SR_CTAID.Y ; /* 0x0000000000007919 */ /* 0x000e680000002600 */ /*0040*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R5 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0205 */ /*0060*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x178], PT ; /* 0x00005e0002007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R5, R0, c[0x0][0x4], R3 ; /* 0x0000010000057a24 */ /* 0x002fca00078e0203 */ /*0080*/ ISETP.GE.OR P0, PT, R5, c[0x0][0x178], P0 ; /* 0x00005e0005007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ MOV R4, c[0x0][0x178] ; /* 0x00005e0000047a02 */ /* 0x000fe20000000f00 */ /*00b0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00c0*/ HFMA2.MMA R18, -RZ, RZ, 0, 0 ; /* 0x00000000ff127435 */ /* 0x000fe400000001ff */ /*00d0*/ ISETP.GE.AND P0, PT, R4, 0x1, PT ; /* 0x000000010400780c */ /* 0x000fda0003f06270 */ /*00e0*/ @!P0 BRA 0xc10 ; /* 0x00000b2000008947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R6, R4.reuse, -0x1, RZ ; /* 0xffffffff04067810 */ /* 0x040fe40007ffe0ff */ /*0100*/ LOP3.LUT R7, R4, 0x3, RZ, 0xc0, !PT ; /* 0x0000000304077812 */ /* 0x000fe400078ec0ff */ /*0110*/ ISETP.GE.U32.AND P0, PT, R6, 0x3, PT ; /* 0x000000030600780c */ /* 0x000fe40003f06070 */ /*0120*/ MOV R6, RZ ; /* 0x000000ff00067202 */ /* 0x000fe40000000f00 */ /*0130*/ MOV R18, RZ ; /* 0x000000ff00127202 */ /* 0x000fd20000000f00 */ /*0140*/ @!P0 BRA 0xb00 ; /* 0x000009b000008947 */ /* 0x000fea0003800000 */ /*0150*/ IADD3 R8, -R7, c[0x0][0x178], RZ ; /* 0x00005e0007087a10 */ /* 0x000fe20007ffe1ff */ /*0160*/ HFMA2.MMA R27, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff1b7435 */ /* 0x000fe200000001ff */ /*0170*/ IADD3 R9, R3, c[0x0][0x178], RZ ; /* 0x00005e0003097a10 */ /* 0x000fe20007ffe0ff */ /*0180*/ ULDC.64 UR6, c[0x0][0x168] ; /* 0x00005a0000067ab9 */ /* 0x000fe20000000a00 */ /*0190*/ ISETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe40003f04270 */ /*01a0*/ MOV R6, RZ ; /* 0x000000ff00067202 */ /* 0x000fe20000000f00 */ /*01b0*/ IMAD R9, R0, c[0x0][0x4], R9 ; /* 0x0000010000097a24 */ /* 0x000fc800078e0209 */ /*01c0*/ IMAD.WIDE R26, R2, R27, c[0x0][0x170] ; /* 0x00005c00021a7625 */ /* 0x000fcc00078e021b */ /*01d0*/ @!P0 BRA 0x970 ; /* 0x0000079000008947 */ /* 0x000fea0003800000 */ /*01e0*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*01f0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*0200*/ @!P1 BRA 0x6b0 ; /* 0x000004a000009947 */ /* 0x000fea0003800000 */ /*0210*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0220*/ MOV R12, UR6 ; /* 0x00000006000c7c02 */ /* 0x000fe20008000f00 */ /*0230*/ LDG.E R19, [R26.64] ; /* 0x000000041a137981 */ /* 0x0000a2000c1e1900 */ /*0240*/ MOV R13, UR7 ; /* 0x00000007000d7c02 */ /* 0x000fca0008000f00 */ /*0250*/ IMAD.WIDE R12, R9, 0x4, R12 ; /* 0x00000004090c7825 */ /* 0x000fca00078e020c */ /*0260*/ LDG.E R28, [R12.64] ; /* 0x000000040c1c7981 */ /* 0x000ea2000c1e1900 */ /*0270*/ IMAD.WIDE R16, R4, 0x4, R26 ; /* 0x0000000404107825 */ /* 0x000fc600078e021a */ /*0280*/ LDG.E R23, [R12.64+0x4] ; /* 0x000004040c177981 */ /* 0x000ee6000c1e1900 */ /*0290*/ IMAD.WIDE R10, R4.reuse, 0x4, R16 ; /* 0x00000004040a7825 */ /* 0x040fe200078e0210 */ /*02a0*/ LDG.E R22, [R16.64] ; /* 0x0000000410167981 */ /* 0x0002e8000c1e1900 */ /*02b0*/ LDG.E R25, [R10.64] ; /* 0x000000040a197981 */ /* 0x000968000c1e1900 */ /*02c0*/ LDG.E R24, [R12.64+0x8] ; /* 0x000008040c187981 */ /* 0x000f68000c1e1900 */ /*02d0*/ LDG.E R14, [R12.64+0xc] ; /* 0x00000c040c0e7981 */ /* 0x000f62000c1e1900 */ /*02e0*/ IMAD.WIDE R10, R4, 0x4, R10 ; /* 0x00000004040a7825 */ /* 0x010fc600078e020a */ /*02f0*/ LDG.E R26, [R12.64+0x10] ; /* 0x000010040c1a7981 */ /* 0x001f28000c1e1900 */ /*0300*/ LDG.E R15, [R10.64] ; /* 0x000000040a0f7981 */ /* 0x000122000c1e1900 */ /*0310*/ IMAD.WIDE R20, R4, 0x4, R10 ; /* 0x0000000404147825 */ /* 0x000fca00078e020a */ /*0320*/ LDG.E R27, [R20.64] ; /* 0x00000004141b7981 */ /* 0x000122000c1e1900 */ /*0330*/ IMAD.WIDE R16, R4, 0x4, R20 ; /* 0x0000000404107825 */ /* 0x002fc600078e0214 */ /*0340*/ LDG.E R10, [R12.64+0x18] ; /* 0x000018040c0a7981 */ /* 0x001f28000c1e1900 */ /*0350*/ LDG.E R21, [R12.64+0x1c] ; /* 0x00001c040c157981 */ /* 0x000f22000c1e1900 */ /*0360*/ IMAD R28, R19, R28, R18 ; /* 0x0000001c131c7224 */ /* 0x004fc600078e0212 */ /*0370*/ LDG.E R18, [R12.64+0x14] ; /* 0x000014040c127981 */ /* 0x000ea8000c1e1900 */ /*0380*/ LDG.E R19, [R16.64] ; /* 0x0000000410137981 */ /* 0x0000a2000c1e1900 */ /*0390*/ IMAD R28, R22, R23, R28 ; /* 0x00000017161c7224 */ /* 0x008fe400078e021c */ /*03a0*/ IMAD.WIDE R16, R4, 0x4, R16 ; /* 0x0000000404107825 */ /* 0x001fca00078e0210 */ /*03b0*/ LDG.E R11, [R16.64] ; /* 0x00000004100b7981 */ /* 0x000ae2000c1e1900 */ /*03c0*/ IMAD.WIDE R22, R4, 0x4, R16 ; /* 0x0000000404167825 */ /* 0x000fca00078e0210 */ /*03d0*/ LDG.E R20, [R22.64] ; /* 0x0000000416147981 */ /* 0x0000e2000c1e1900 */ /*03e0*/ IMAD R16, R25, R24, R28 ; /* 0x0000001819107224 */ /* 0x020fe400078e021c */ /*03f0*/ IMAD.WIDE R24, R4.reuse, 0x4, R22 ; /* 0x0000000404187825 */ /* 0x040fe200078e0216 */ /*0400*/ LDG.E R28, [R12.64+0x20] ; /* 0x000020040c1c7981 */ /* 0x000f68000c1e1900 */ /*0410*/ LDG.E R29, [R24.64] ; /* 0x00000004181d7981 */ /* 0x000362000c1e1900 */ /*0420*/ IMAD R16, R15, R14, R16 ; /* 0x0000000e0f107224 */ /* 0x010fe400078e0210 */ /*0430*/ IMAD.WIDE R14, R4, 0x4, R24 ; /* 0x00000004040e7825 */ /* 0x000fe200078e0218 */ /*0440*/ LDG.E R23, [R12.64+0x24] ; /* 0x000024040c177981 */ /* 0x001f26000c1e1900 */ /*0450*/ IMAD R26, R27, R26, R16 ; /* 0x0000001a1b1a7224 */ /* 0x000fc400078e0210 */ /*0460*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fe200078e020e */ /*0470*/ LDG.E R27, [R12.64+0x28] ; /* 0x000028040c1b7981 */ /* 0x000f28000c1e1900 */ /*0480*/ LDG.E R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000128000c1e1900 */ /*0490*/ LDG.E R22, [R16.64] ; /* 0x0000000410167981 */ /* 0x000328000c1e1900 */ /*04a0*/ LDG.E R15, [R12.64+0x30] ; /* 0x000030040c0f7981 */ /* 0x001f22000c1e1900 */ /*04b0*/ IMAD R26, R19, R18, R26 ; /* 0x00000012131a7224 */ /* 0x004fc400078e021a */ /*04c0*/ IMAD.WIDE R18, R4, 0x4, R16 ; /* 0x0000000404127825 */ /* 0x000fc800078e0210 */ /*04d0*/ IMAD R26, R11, R10, R26 ; /* 0x0000000a0b1a7224 */ /* 0x008fe400078e021a */ /*04e0*/ IMAD.WIDE R10, R4, 0x4, R18 ; /* 0x00000004040a7825 */ /* 0x000fe400078e0212 */ /*04f0*/ LDG.E R18, [R18.64] ; /* 0x0000000412127981 */ /* 0x0000a4000c1e1900 */ /*0500*/ IMAD R24, R20, R21, R26 ; /* 0x0000001514187224 */ /* 0x002fe400078e021a */ /*0510*/ IMAD.WIDE R20, R4, 0x4, R10 ; /* 0x0000000404147825 */ /* 0x000fe200078e020a */ /*0520*/ LDG.E R26, [R12.64+0x2c] ; /* 0x00002c040c1a7981 */ /* 0x000ea8000c1e1900 */ /*0530*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x0002e2000c1e1900 */ /*0540*/ IMAD R28, R29, R28, R24 ; /* 0x0000001c1d1c7224 */ /* 0x020fc600078e0218 */ /*0550*/ LDG.E R19, [R12.64+0x38] ; /* 0x000038040c137981 */ /* 0x001f62000c1e1900 */ /*0560*/ IMAD.WIDE R24, R4, 0x4, R20 ; /* 0x0000000404187825 */ /* 0x000fc600078e0214 */ /*0570*/ LDG.E R20, [R20.64] ; /* 0x0000000414147981 */ /* 0x000168000c1e1900 */ /*0580*/ LDG.E R11, [R12.64+0x34] ; /* 0x000034040c0b7981 */ /* 0x002f62000c1e1900 */ /*0590*/ IMAD.WIDE R16, R4, 0x4, R24 ; /* 0x0000000404107825 */ /* 0x000fc600078e0218 */ /*05a0*/ LDG.E R29, [R24.64] ; /* 0x00000004181d7981 */ /* 0x000368000c1e1900 */ /*05b0*/ LDG.E R21, [R16.64] ; /* 0x0000000410157981 */ /* 0x001f68000c1e1900 */ /*05c0*/ LDG.E R24, [R12.64+0x3c] ; /* 0x00003c040c187981 */ /* 0x002f62000c1e1900 */ /*05d0*/ IMAD R14, R14, R23, R28 ; /* 0x000000170e0e7224 */ /* 0x010fe200078e021c */ /*05e0*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fc60007ffe0ff */ /*05f0*/ IMAD R27, R22, R27, R14 ; /* 0x0000001b161b7224 */ /* 0x000fe200078e020e */ /*0600*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe20003f24270 */ /*0610*/ UIADD3 UR6, UP0, UR6, 0x40, URZ ; /* 0x0000004006067890 */ /* 0x000fe2000ff1e03f */ /*0620*/ IADD3 R6, R6, 0x10, RZ ; /* 0x0000001006067810 */ /* 0x000fc60007ffe0ff */ /*0630*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0640*/ IMAD R18, R18, R26, R27 ; /* 0x0000001a12127224 */ /* 0x004fc800078e021b */ /*0650*/ IMAD R10, R10, R15, R18 ; /* 0x0000000f0a0a7224 */ /* 0x008fe400078e0212 */ /*0660*/ IMAD.WIDE R26, R4, 0x4, R16 ; /* 0x00000004041a7825 */ /* 0x000fc800078e0210 */ /*0670*/ IMAD R10, R20, R11, R10 ; /* 0x0000000b140a7224 */ /* 0x020fc800078e020a */ /*0680*/ IMAD R10, R29, R19, R10 ; /* 0x000000131d0a7224 */ /* 0x000fc800078e020a */ /*0690*/ IMAD R18, R21, R24, R10 ; /* 0x0000001815127224 */ /* 0x000fe200078e020a */ /*06a0*/ @P1 BRA 0x220 ; /* 0xfffffb7000001947 */ /* 0x000fea000383ffff */ /*06b0*/ ISETP.GT.AND P1, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f24270 */ /*06c0*/ @!P1 BRA 0x950 ; /* 0x0000028000009947 */ /* 0x000fea0003800000 */ /*06d0*/ MOV R10, UR6 ; /* 0x00000006000a7c02 */ /* 0x000fe20008000f00 */ /*06e0*/ LDG.E R19, [R26.64] ; /* 0x000000041a137981 */ /* 0x000ea2000c1e1900 */ /*06f0*/ MOV R11, UR7 ; /* 0x00000007000b7c02 */ /* 0x000fca0008000f00 */ /*0700*/ IMAD.WIDE R10, R9, 0x4, R10 ; /* 0x00000004090a7825 */ /* 0x000fca00078e020a */ /*0710*/ LDG.E R24, [R10.64] ; /* 0x000000040a187981 */ /* 0x000ea2000c1e1900 */ /*0720*/ IMAD.WIDE R22, R4, 0x4, R26 ; /* 0x0000000404167825 */ /* 0x000fc600078e021a */ /*0730*/ LDG.E R25, [R10.64+0x4] ; /* 0x000004040a197981 */ /* 0x000ee6000c1e1900 */ /*0740*/ IMAD.WIDE R14, R4.reuse, 0x4, R22 ; /* 0x00000004040e7825 */ /* 0x040fe200078e0216 */ /*0750*/ LDG.E R29, [R10.64+0x8] ; /* 0x000008040a1d7981 */ /* 0x000f28000c1e1900 */ /*0760*/ LDG.E R22, [R22.64] ; /* 0x0000000416167981 */ /* 0x0000e2000c1e1900 */ /*0770*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fc600078e020e */ /*0780*/ LDG.E R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000326000c1e1900 */ /*0790*/ IMAD.WIDE R12, R4.reuse, 0x4, R16 ; /* 0x00000004040c7825 */ /* 0x040fe200078e0210 */ /*07a0*/ LDG.E R28, [R10.64+0xc] ; /* 0x00000c040a1c7981 */ /* 0x000f68000c1e1900 */ /*07b0*/ LDG.E R17, [R16.64] ; /* 0x0000000410117981 */ /* 0x000362000c1e1900 */ /*07c0*/ IMAD.WIDE R20, R4, 0x4, R12 ; /* 0x0000000404147825 */ /* 0x000fc600078e020c */ /*07d0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000328000c1e1900 */ /*07e0*/ LDG.E R23, [R10.64+0x14] ; /* 0x000014040a177981 */ /* 0x001f68000c1e1900 */ /*07f0*/ LDG.E R16, [R10.64+0x18] ; /* 0x000018040a107981 */ /* 0x002f68000c1e1900 */ /*0800*/ LDG.E R13, [R10.64+0x10] ; /* 0x000010040a0d7981 */ /* 0x000f62000c1e1900 */ /*0810*/ IMAD R24, R19, R24, R18 ; /* 0x0000001813187224 */ /* 0x004fc400078e0212 */ /*0820*/ IMAD.WIDE R18, R4.reuse, 0x4, R20 ; /* 0x0000000404127825 */ /* 0x040fe400078e0214 */ /*0830*/ LDG.E R20, [R20.64] ; /* 0x0000000414147981 */ /* 0x000ea8000c1e1900 */ /*0840*/ IMAD.WIDE R26, R4, 0x4, R18 ; /* 0x00000004041a7825 */ /* 0x000fe200078e0212 */ /*0850*/ LDG.E R15, [R18.64] ; /* 0x00000004120f7981 */ /* 0x0000a8000c1e1900 */ /*0860*/ LDG.E R19, [R10.64+0x1c] ; /* 0x00001c040a137981 */ /* 0x001ea8000c1e1900 */ /*0870*/ LDG.E R18, [R26.64] ; /* 0x000000041a127981 */ /* 0x0000a2000c1e1900 */ /*0880*/ IMAD R22, R22, R25, R24 ; /* 0x0000001916167224 */ /* 0x008fc800078e0218 */ /*0890*/ IMAD R14, R14, R29, R22 ; /* 0x0000001d0e0e7224 */ /* 0x010fc800078e0216 */ /*08a0*/ IMAD R14, R17, R28, R14 ; /* 0x0000001c110e7224 */ /* 0x020fc800078e020e */ /*08b0*/ IMAD R12, R12, R13, R14 ; /* 0x0000000d0c0c7224 */ /* 0x000fe200078e020e */ /*08c0*/ UIADD3 UR6, UP0, UR6, 0x20, URZ ; /* 0x0000002006067890 */ /* 0x000fe2000ff1e03f */ /*08d0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe20003f0e170 */ /*08e0*/ IMAD.WIDE R26, R4, 0x4, R26 ; /* 0x00000004041a7825 */ /* 0x001fe200078e021a */ /*08f0*/ IADD3 R6, R6, 0x8, RZ ; /* 0x0000000806067810 */ /* 0x000fe40007ffe0ff */ /*0900*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe20007ffe0ff */ /*0910*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0920*/ IMAD R12, R20, R23, R12 ; /* 0x00000017140c7224 */ /* 0x004fc800078e020c */ /*0930*/ IMAD R12, R15, R16, R12 ; /* 0x000000100f0c7224 */ /* 0x000fc800078e020c */ /*0940*/ IMAD R18, R18, R19, R12 ; /* 0x0000001312127224 */ /* 0x000fe400078e020c */ /*0950*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0960*/ @!P0 BRA 0xb00 ; /* 0x0000019000008947 */ /* 0x000fea0003800000 */ /*0970*/ MOV R10, UR6 ; /* 0x00000006000a7c02 */ /* 0x000fe20008000f00 */ /*0980*/ IMAD.WIDE R12, R4, 0x4, R26 ; /* 0x00000004040c7825 */ /* 0x000fe200078e021a */ /*0990*/ MOV R11, UR7 ; /* 0x00000007000b7c02 */ /* 0x000fe20008000f00 */ /*09a0*/ LDG.E R27, [R26.64] ; /* 0x000000041a1b7981 */ /* 0x000ea8000c1e1900 */ /*09b0*/ IMAD.WIDE R10, R9, 0x4, R10 ; /* 0x00000004090a7825 */ /* 0x000fc800078e020a */ /*09c0*/ IMAD.WIDE R14, R4.reuse, 0x4, R12 ; /* 0x00000004040e7825 */ /* 0x040fe200078e020c */ /*09d0*/ LDG.E R19, [R10.64] ; /* 0x000000040a137981 */ /* 0x000ea8000c1e1900 */ /*09e0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000ee2000c1e1900 */ /*09f0*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fc600078e020e */ /*0a00*/ LDG.E R20, [R10.64+0x4] ; /* 0x000004040a147981 */ /* 0x000ee8000c1e1900 */ /*0a10*/ LDG.E R22, [R14.64] ; /* 0x000000040e167981 */ /* 0x000f28000c1e1900 */ /*0a20*/ LDG.E R21, [R10.64+0x8] ; /* 0x000008040a157981 */ /* 0x000f28000c1e1900 */ /*0a30*/ LDG.E R23, [R10.64+0xc] ; /* 0x00000c040a177981 */ /* 0x000f68000c1e1900 */ /*0a40*/ LDG.E R24, [R16.64] ; /* 0x0000000410187981 */ /* 0x000f62000c1e1900 */ /*0a50*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fc80007ffe0ff */ /*0a60*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f05270 */ /*0a70*/ UIADD3 UR6, UP0, UR6, 0x10, URZ ; /* 0x0000001006067890 */ /* 0x000fe2000ff1e03f */ /*0a80*/ IADD3 R6, R6, 0x4, RZ ; /* 0x0000000406067810 */ /* 0x000fc60007ffe0ff */ /*0a90*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0aa0*/ IMAD R19, R27, R19, R18 ; /* 0x000000131b137224 */ /* 0x004fc800078e0212 */ /*0ab0*/ IMAD R19, R12, R20, R19 ; /* 0x000000140c137224 */ /* 0x008fe400078e0213 */ /*0ac0*/ IMAD.WIDE R26, R4, 0x4, R16 ; /* 0x00000004041a7825 */ /* 0x000fc800078e0210 */ /*0ad0*/ IMAD R19, R22, R21, R19 ; /* 0x0000001516137224 */ /* 0x010fc800078e0213 */ /*0ae0*/ IMAD R18, R24, R23, R19 ; /* 0x0000001718127224 */ /* 0x020fe200078e0213 */ /*0af0*/ @P0 BRA 0x970 ; /* 0xfffffe7000000947 */ /* 0x000fea000383ffff */ /*0b00*/ ISETP.NE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fda0003f05270 */ /*0b10*/ @!P0 BRA 0xc10 ; /* 0x000000f000008947 */ /* 0x000fea0003800000 */ /*0b20*/ HFMA2.MMA R10, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0a7435 */ /* 0x000fe200000001ff */ /*0b30*/ IADD3 R3, R3, c[0x0][0x178], R6 ; /* 0x00005e0003037a10 */ /* 0x000fe20007ffe006 */ /*0b40*/ IMAD R6, R6, c[0x0][0x178], R2 ; /* 0x00005e0006067a24 */ /* 0x000fc800078e0202 */ /*0b50*/ IMAD R3, R0, c[0x0][0x4], R3 ; /* 0x0000010000037a24 */ /* 0x000fc800078e0203 */ /*0b60*/ IMAD.WIDE R8, R3, R10, c[0x0][0x168] ; /* 0x00005a0003087625 */ /* 0x000fc800078e020a */ /*0b70*/ IMAD.WIDE R10, R6, R10, c[0x0][0x170] ; /* 0x00005c00060a7625 */ /* 0x000fca00078e020a */ /*0b80*/ LDG.E R3, [R10.64] ; /* 0x000000040a037981 */ /* 0x0000a8000c1e1900 */ /*0b90*/ LDG.E R0, [R8.64] ; /* 0x0000000408007981 */ /* 0x0002a2000c1e1900 */ /*0ba0*/ IADD3 R7, R7, -0x1, RZ ; /* 0xffffffff07077810 */ /* 0x000fc80007ffe0ff */ /*0bb0*/ ISETP.NE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fe20003f05270 */ /*0bc0*/ IMAD.WIDE R10, R4, 0x4, R10 ; /* 0x00000004040a7825 */ /* 0x001fe200078e020a */ /*0bd0*/ IADD3 R8, P1, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x002fc80007f3e0ff */ /*0be0*/ IADD3.X R9, RZ, R9, RZ, P1, !PT ; /* 0x00000009ff097210 */ /* 0x000fe20000ffe4ff */ /*0bf0*/ IMAD R18, R3, R0, R18 ; /* 0x0000000003127224 */ /* 0x004fcc00078e0212 */ /*0c00*/ @P0 BRA 0xb80 ; /* 0xffffff7000000947 */ /* 0x000fea000383ffff */ /*0c10*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fe20000000f00 */ /*0c20*/ IMAD R2, R5, c[0x0][0x178], R2 ; /* 0x00005e0005027a24 */ /* 0x000fc800078e0202 */ /*0c30*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*0c40*/ STG.E [R2.64], R18 ; /* 0x0000001202007986 */ /* 0x000fe2000c101904 */ /*0c50*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0c60*/ BRA 0xc60; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0c70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ca0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ce0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cf0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #include <stdlib.h> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); cudaMalloc((void**) &A_d, size); cudaMalloc((void**) &B_d, size); cudaMalloc((void**) &C_d, size); cudaMemcpy(A_d, A_h, size, cudaMemcpyHostToDevice); cudaMemcpy(B_d, B_h, size, cudaMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); cudaFree(A_d); cudaFree(B_d); cudaFree(C_d); return 0; }
.file "tmpxft_0010c29c_00000000-6_CUDA_matrix.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2063: .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 .LFE2063: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z5techoii .type _Z5techoii, @function _Z5techoii: .LFB2057: .cfi_startproc endbr64 movl %edi, %eax cltd idivl %esi cmpl $1, %edx sbbl $-1, %eax ret .cfi_endproc .LFE2057: .size _Z5techoii, .-_Z5techoii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d " .LC1: .string "\n" .text .globl _Z14imprimirMatrizPii .type _Z14imprimirMatrizPii, @function _Z14imprimirMatrizPii: .LFB2058: .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 $24, %rsp .cfi_def_cfa_offset 80 movl %esi, 12(%rsp) testl %esi, %esi jle .L4 movslq %esi, %r14 leaq 0(,%r14,4), %r15 leaq (%rdi,%r15), %rbp negq %r14 salq $2, %r14 movl $0, %r13d leaq .LC0(%rip), %r12 .L6: leaq 0(%rbp,%r14), %rbx .L7: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %rbp, %rbx jne .L7 leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %r13d addq %r15, %rbp cmpl %r13d, 12(%rsp) jne .L6 .L4: addq $24, %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 .cfi_endproc .LFE2058: .size _Z14imprimirMatrizPii, .-_Z14imprimirMatrizPii .globl _Z20newMatrizVectorizadaii .type _Z20newMatrizVectorizadaii, @function _Z20newMatrizVectorizadaii: .LFB2059: .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 movl %edi, %ebp movl %esi, %ebx movslq %edi, %r12 movq %r12, %rdi imulq %r12, %rdi salq $2, %rdi call malloc@PLT testl %ebp, %ebp jle .L10 leaq 0(,%r12,4), %r8 leaq (%rax,%r8), %rcx negq %r12 leaq 0(,%r12,4), %rdi movl $0, %esi .L12: leaq (%rcx,%rdi), %rdx .L13: movl %ebx, (%rdx) addq $4, %rdx cmpq %rcx, %rdx jne .L13 addl $1, %esi addq %r8, %rcx cmpl %esi, %ebp jne .L12 .L10: popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _Z20newMatrizVectorizadaii, .-_Z20newMatrizVectorizadaii .globl _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i .type _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i, @function _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i: .LFB2085: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L20 .L16: movq 136(%rsp), %rax subq %fs:40, %rax jne .L21 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L20: .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 _Z19multiplicarMatricesPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L16 .L21: call __stack_chk_fail@PLT .cfi_endproc .LFE2085: .size _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i, .-_Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i .globl _Z19multiplicarMatricesPiS_S_i .type _Z19multiplicarMatricesPiS_S_i, @function _Z19multiplicarMatricesPiS_S_i: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _Z19multiplicarMatricesPiS_S_i, .-_Z19multiplicarMatricesPiS_S_i .globl _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i .type _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i, @function _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L28 .L24: movq 136(%rsp), %rax subq %fs:40, %rax jne .L29 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L28: .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 _Z13sumarMatricesPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L24 .L29: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i, .-_Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i .globl _Z13sumarMatricesPiS_S_i .type _Z13sumarMatricesPiS_S_i, @function _Z13sumarMatricesPiS_S_i: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z13sumarMatricesPiS_S_i, .-_Z13sumarMatricesPiS_S_i .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "==============================================RESULTADO============================================\n" .align 8 .LC3: .string "===================================================================================================\n" .text .globl main .type main, @function main: .LFB2060: .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 $64, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $12, %ebx cmpl $1, %edi jg .L39 .L33: movl %ebx, %ebp imull %ebx, %ebp movslq %ebp, %rbp salq $2, %rbp movl $2, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r14 movl $3, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r13 movl $0, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r12 leaq 8(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT leaq 16(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT leaq 24(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbp, %rdx movq %r14, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbp, %rdx movq %r13, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 40(%rsp) testb $3, %bl setne %dl movzbl %dl, %edx leal 3(%rbx), %eax testl %ebx, %ebx cmovns %ebx, %eax sarl $2, %eax addl %edx, %eax movl %eax, 44(%rsp) movl %eax, 48(%rsp) movl $1, 52(%rsp) movl $4, 32(%rsp) movl $4, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movl $1, %ecx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L40 .L34: movl $2, %ecx movq %rbp, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movl %ebx, %esi movq %r12, %rdi call _Z14imprimirMatrizPii movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl 52(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L41 .L35: movl $2, %ecx movq %rbp, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %esi movq %r12, %rdi call _Z14imprimirMatrizPii leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r14, %rdi call free@PLT movq %r13, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L42 movl $0, %eax addq $64, %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 .L39: .cfi_restore_state movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movl %eax, %ebx jmp .L33 .L40: movl %ebx, %ecx movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq 24(%rsp), %rdi call _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i jmp .L34 .L41: movl %ebx, %ecx movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i jmp .L35 .L42: call __stack_chk_fail@PLT .cfi_endproc .LFE2060: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z13sumarMatricesPiS_S_i" .section .rodata.str1.8 .align 8 .LC5: .string "_Z19multiplicarMatricesPiS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2090: .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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z13sumarMatricesPiS_S_i(%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 .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z19multiplicarMatricesPiS_S_i(%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 .LFE2090: .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> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); cudaMalloc((void**) &A_d, size); cudaMalloc((void**) &B_d, size); cudaMalloc((void**) &C_d, size); cudaMemcpy(A_d, A_h, size, cudaMemcpyHostToDevice); cudaMemcpy(B_d, B_h, size, cudaMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); cudaMemcpy(C_h, C_d, size, cudaMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); cudaFree(A_d); cudaFree(B_d); cudaFree(C_d); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); hipMalloc((void**) &A_d, size); hipMalloc((void**) &B_d, size); hipMalloc((void**) &C_d, size); hipMemcpy(A_d, A_h, size, hipMemcpyHostToDevice); hipMemcpy(B_d, B_h, size, hipMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); hipFree(A_d); hipFree(B_d); hipFree(C_d); 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> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); hipMalloc((void**) &A_d, size); hipMalloc((void**) &B_d, size); hipMalloc((void**) &C_d, size); hipMemcpy(A_d, A_h, size, hipMemcpyHostToDevice); hipMemcpy(B_d, B_h, size, hipMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); hipFree(A_d); hipFree(B_d); hipFree(C_d); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z19multiplicarMatricesPiS_S_i .globl _Z19multiplicarMatricesPiS_S_i .p2align 8 .type _Z19multiplicarMatricesPiS_S_i,@function _Z19multiplicarMatricesPiS_S_i: s_clause 0x1 s_load_b32 s3, s[0:1], 0x2c s_load_b32 s2, s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s3, 0xffff s_lshr_b32 s3, s3, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mad_u64_u32 v[0:1], null, s14, s4, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s3, v[3:4] s_mov_b32 s3, exec_lo v_max_i32_e32 v2, v0, v1 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s2, v2 s_cbranch_execz .LBB0_6 s_cmp_lt_i32 s2, 1 s_cbranch_scc1 .LBB0_4 s_load_b128 s[4:7], s[0:1], 0x8 v_dual_mov_b32 v2, 0 :: v_dual_add_nc_u32 v5, s2, v1 v_mov_b32_e32 v3, v0 s_mov_b32 s3, 0 .p2align 6 .LBB0_3: s_delay_alu instid0(VALU_DEP_2) | instid1(SALU_CYCLE_1) v_add_nc_u32_e32 v6, s3, v5 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_ashrrev_i32_e32 v4, 31, v3 s_add_i32 s3, s3, 1 s_cmp_eq_u32 s2, s3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v7, 31, v6 v_lshlrev_b64 v[8:9], 2, v[3:4] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_lshlrev_b64 v[6:7], 2, v[6:7] s_waitcnt lgkmcnt(0) v_add_co_u32 v8, vcc_lo, s6, v8 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_ci_u32_e32 v9, vcc_lo, s7, v9, vcc_lo v_add_co_u32 v6, vcc_lo, s4, v6 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v7, vcc_lo, s5, v7, vcc_lo global_load_b32 v4, v[8:9], off global_load_b32 v8, v[6:7], off s_waitcnt vmcnt(0) v_mad_u64_u32 v[6:7], null, v4, v8, v[2:3] v_dual_mov_b32 v2, v6 :: v_dual_add_nc_u32 v3, s2, v3 s_cbranch_scc0 .LBB0_3 s_branch .LBB0_5 .LBB0_4: v_mov_b32_e32 v2, 0 .LBB0_5: s_load_b64 s[0:1], s[0:1], 0x0 v_mad_u64_u32 v[3:4], null, v1, s2, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v4, 31, v3 v_lshlrev_b64 v[0:1], 2, v[3:4] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v2, off .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19multiplicarMatricesPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .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 10 .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 _Z19multiplicarMatricesPiS_S_i, .Lfunc_end0-_Z19multiplicarMatricesPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z13sumarMatricesPiS_S_i .globl _Z13sumarMatricesPiS_S_i .p2align 8 .type _Z13sumarMatricesPiS_S_i,@function _Z13sumarMatricesPiS_S_i: s_clause 0x1 s_load_b32 s3, s[0:1], 0x2c s_load_b32 s2, s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s3, 0xffff s_lshr_b32 s3, s3, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mad_u64_u32 v[0:1], null, s14, s4, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s3, v[3:4] s_mov_b32 s3, exec_lo v_max_i32_e32 v2, v0, v1 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s2, v2 s_cbranch_execz .LBB1_2 s_load_b128 s[4:7], s[0:1], 0x0 v_mad_u64_u32 v[2:3], null, v1, s2, v[0:1] s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[0:1], 2, v[2:3] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s1, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s4, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s5, v1, vcc_lo s_waitcnt vmcnt(0) v_add_nc_u32_e32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB1_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13sumarMatricesPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .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 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_end1: .size _Z13sumarMatricesPiS_S_i, .Lfunc_end1-_Z13sumarMatricesPiS_S_i .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: 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: _Z19multiplicarMatricesPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19multiplicarMatricesPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 10 .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 - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .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: _Z13sumarMatricesPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13sumarMatricesPiS_S_i.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 <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> typedef int* Vector; int techo(int tam, int block_size) { return (tam / block_size) + (tam % block_size == 0? 0: 1); } __global__ void multiplicarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) { int suma = 0; for(int k = 0; k < tam; k++) suma += A[idy + tam + k] * B [k * tam + idx]; C[index] = suma; } } __global__ void sumarMatrices(int *C, int *A, int *B, int tam) { int idx = blockIdx.x * blockDim.x + threadIdx.x; int idy = blockIdx.y * blockDim.y + threadIdx.y; int index = idy * tam + idx; if(idy < tam && idx < tam) C[index] = A[index] + B[index]; } void imprimirMatriz(int *A, int tam) { for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { printf("%d ", A[tam * i + j]); } printf("\n"); } } Vector newMatrizVectorizada(int tam, int val) { Vector v = (Vector) malloc(sizeof(int) * tam * tam); for(int i = 0; i < tam; i++) { for(int j = 0; j < tam; j++) { v[i * tam + j] = val; } } return v; } int main(int argc, char **argv) { Vector A_h, B_h, C_h; Vector A_d, B_d, C_d; int tam = 12; if(argc > 1) tam = atoi(argv[1]); int N = tam * tam; size_t size = N * sizeof(int); A_h = newMatrizVectorizada(tam, 2); B_h = newMatrizVectorizada(tam, 3); C_h = newMatrizVectorizada(tam, 0); hipMalloc((void**) &A_d, size); hipMalloc((void**) &B_d, size); hipMalloc((void**) &C_d, size); hipMemcpy(A_d, A_h, size, hipMemcpyHostToDevice); hipMemcpy(B_d, B_h, size, hipMemcpyHostToDevice); int BLOCK_SIZE = 4; dim3 block_size(BLOCK_SIZE, BLOCK_SIZE); dim3 n_blocks(techo(tam, block_size.x), techo(tam, block_size.y)); multiplicarMatrices<<< n_blocks, block_size >>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); imprimirMatriz(C_h, tam); sumarMatrices <<< n_blocks, block_size>>> (C_d, A_d, B_d, tam); hipMemcpy(C_h, C_d, size, hipMemcpyDeviceToHost); printf("==============================================RESULTADO============================================\n"); imprimirMatriz(C_h, tam); printf("===================================================================================================\n"); free(A_h); free(B_h); free(C_h); hipFree(A_d); hipFree(B_d); hipFree(C_d); return 0; }
.text .file "CUDA_matrix.hip" .globl _Z5techoii # -- Begin function _Z5techoii .p2align 4, 0x90 .type _Z5techoii,@function _Z5techoii: # @_Z5techoii .cfi_startproc # %bb.0: movl %edi, %eax cltd idivl %esi cmpl $1, %edx sbbl $-1, %eax retq .Lfunc_end0: .size _Z5techoii, .Lfunc_end0-_Z5techoii .cfi_endproc # -- End function .globl _Z34__device_stub__multiplicarMatricesPiS_S_i # -- Begin function _Z34__device_stub__multiplicarMatricesPiS_S_i .p2align 4, 0x90 .type _Z34__device_stub__multiplicarMatricesPiS_S_i,@function _Z34__device_stub__multiplicarMatricesPiS_S_i: # @_Z34__device_stub__multiplicarMatricesPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z19multiplicarMatricesPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z34__device_stub__multiplicarMatricesPiS_S_i, .Lfunc_end1-_Z34__device_stub__multiplicarMatricesPiS_S_i .cfi_endproc # -- End function .globl _Z28__device_stub__sumarMatricesPiS_S_i # -- Begin function _Z28__device_stub__sumarMatricesPiS_S_i .p2align 4, 0x90 .type _Z28__device_stub__sumarMatricesPiS_S_i,@function _Z28__device_stub__sumarMatricesPiS_S_i: # @_Z28__device_stub__sumarMatricesPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z13sumarMatricesPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z28__device_stub__sumarMatricesPiS_S_i, .Lfunc_end2-_Z28__device_stub__sumarMatricesPiS_S_i .cfi_endproc # -- End function .globl _Z14imprimirMatrizPii # -- Begin function _Z14imprimirMatrizPii .p2align 4, 0x90 .type _Z14imprimirMatrizPii,@function _Z14imprimirMatrizPii: # @_Z14imprimirMatrizPii .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 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 movq %rdi, (%rsp) # 8-byte Spill testl %esi, %esi jle .LBB3_5 # %bb.1: # %.preheader.lr.ph movl %esi, %ebx movl %esi, %r15d xorl %ebp, %ebp xorl %r12d, %r12d .p2align 4, 0x90 .LBB3_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB3_3 Depth 2 movl %ebp, %eax movq (%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r13 xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_3: # Parent Loop BB3_2 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r13,%r14,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r14 cmpq %r14, %r15 jne .LBB3_3 # %bb.4: # %._crit_edge # in Loop: Header=BB3_2 Depth=1 movl $10, %edi callq putchar@PLT incq %r12 addl %ebx, %ebp cmpq %r15, %r12 jne .LBB3_2 .LBB3_5: # %._crit_edge13 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 retq .Lfunc_end3: .size _Z14imprimirMatrizPii, .Lfunc_end3-_Z14imprimirMatrizPii .cfi_endproc # -- End function .globl _Z20newMatrizVectorizadaii # -- Begin function _Z20newMatrizVectorizadaii .p2align 4, 0x90 .type _Z20newMatrizVectorizadaii,@function _Z20newMatrizVectorizadaii: # @_Z20newMatrizVectorizadaii .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 movl %esi, %ebx movl %edi, %ebp movslq %edi, %r14 movq %r14, %rdi imulq %r14, %rdi shlq $2, %rdi callq malloc testl %r14d, %r14d jle .LBB4_5 # %bb.1: # %.preheader.lr.ph movl %ebp, %ecx xorl %edx, %edx xorl %esi, %esi .p2align 4, 0x90 .LBB4_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB4_3 Depth 2 movl %edx, %edi leaq (%rax,%rdi,4), %rdi xorl %r8d, %r8d .p2align 4, 0x90 .LBB4_3: # Parent Loop BB4_2 Depth=1 # => This Inner Loop Header: Depth=2 movl %ebx, (%rdi,%r8,4) incq %r8 cmpq %r8, %rcx jne .LBB4_3 # %bb.4: # %._crit_edge # in Loop: Header=BB4_2 Depth=1 incq %rsi addl %ebp, %edx cmpq %rcx, %rsi jne .LBB4_2 .LBB4_5: # %._crit_edge17 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size _Z20newMatrizVectorizadaii, .Lfunc_end4-_Z20newMatrizVectorizadaii .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 $184, %rsp .cfi_def_cfa_offset 240 .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 $12, %ebx cmpl $2, %edi jl .LBB5_2 # %bb.1: movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %rbx .LBB5_2: movl %ebx, %r15d imull %r15d, %r15d movslq %ebx, %r14 movq %r14, %r13 imulq %r14, %r13 shlq $2, %r13 movq %r13, %rdi callq malloc movl %ebx, %r12d testl %r14d, %r14d jle .LBB5_7 # %bb.3: # %.preheader.lr.ph.i xorl %edi, %edi xorl %ecx, %ecx .p2align 4, 0x90 .LBB5_4: # %.preheader.i # =>This Loop Header: Depth=1 # Child Loop BB5_5 Depth 2 movl %edi, %edx leaq (%rax,%rdx,4), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB5_5: # Parent Loop BB5_4 Depth=1 # => This Inner Loop Header: Depth=2 movl $2, (%rdx,%rsi,4) incq %rsi cmpq %rsi, %r12 jne .LBB5_5 # %bb.6: # %._crit_edge.i # in Loop: Header=BB5_4 Depth=1 incq %rcx addl %ebx, %edi cmpq %r12, %rcx jne .LBB5_4 .LBB5_7: # %_Z20newMatrizVectorizadaii.exit movq %rax, 128(%rsp) # 8-byte Spill movq %r13, %rdi callq malloc testl %ebx, %ebx jle .LBB5_12 # %bb.8: # %.preheader.lr.ph.i47 xorl %edi, %edi xorl %ecx, %ecx .p2align 4, 0x90 .LBB5_9: # %.preheader.i49 # =>This Loop Header: Depth=1 # Child Loop BB5_10 Depth 2 movl %edi, %edx leaq (%rax,%rdx,4), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB5_10: # Parent Loop BB5_9 Depth=1 # => This Inner Loop Header: Depth=2 movl $3, (%rdx,%rsi,4) incq %rsi cmpq %rsi, %r12 jne .LBB5_10 # %bb.11: # %._crit_edge.i55 # in Loop: Header=BB5_9 Depth=1 incq %rcx addl %ebx, %edi cmpq %r12, %rcx jne .LBB5_9 .LBB5_12: # %_Z20newMatrizVectorizadaii.exit58 movq %rax, 120(%rsp) # 8-byte Spill shlq $2, %r15 movq %r15, 112(%rsp) # 8-byte Spill movq %r13, %rdi callq malloc movq %rax, %r13 testl %ebx, %ebx jle .LBB5_15 # %bb.13: # %.preheader.lr.ph.i59 leaq (,%r12,4), %rbp xorl %r14d, %r14d movq %r12, %r15 .p2align 4, 0x90 .LBB5_14: # %.preheader.i61 # =>This Inner Loop Header: Depth=1 movl %r14d, %eax leaq (,%rax,4), %rdi addq %r13, %rdi xorl %esi, %esi movq %rbp, %rdx callq memset@PLT addl %ebx, %r14d decq %r15 jne .LBB5_14 .LBB5_15: # %_Z20newMatrizVectorizadaii.exit70 movabsq $17179869188, %r15 # imm = 0x400000004 leaq 32(%rsp), %rdi movq 112(%rsp), %r14 # 8-byte Reload movq %r14, %rsi callq hipMalloc leaq 24(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 32(%rsp), %rdi movq 128(%rsp), %rsi # 8-byte Reload movq %r14, %rdx movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movq 120(%rsp), %rsi # 8-byte Reload movq %r14, %rdx movl $1, %ecx callq hipMemcpy leal 3(%rbx), %eax testl %ebx, %ebx cmovnsl %ebx, %eax sarl $2, %eax movl %ebx, %ecx andl $3, %ecx cmpl $1, %ecx sbbl $-1, %eax movq %rax, %rdi shlq $32, %rdi orq %rax, %rdi movq %rdi, 176(%rsp) # 8-byte Spill movl $1, %esi movq %r15, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_17 # %bb.16: movq 16(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %ebx, 12(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 96(%rsp), %rax movq %rax, 152(%rsp) leaq 88(%rsp), %rax movq %rax, 160(%rsp) leaq 12(%rsp), %rax movq %rax, 168(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z19multiplicarMatricesPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_17: movq 16(%rsp), %rsi movq %r13, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy testl %ebx, %ebx movq %r13, 136(%rsp) # 8-byte Spill jle .LBB5_22 # %bb.18: # %.preheader.lr.ph.i71 xorl %ebp, %ebp xorl %r14d, %r14d .p2align 4, 0x90 .LBB5_19: # %.preheader.i72 # =>This Loop Header: Depth=1 # Child Loop BB5_20 Depth 2 movl %ebp, %eax leaq (,%rax,4), %r15 addq %r13, %r15 xorl %r13d, %r13d .p2align 4, 0x90 .LBB5_20: # Parent Loop BB5_19 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r13,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB5_20 # %bb.21: # %._crit_edge.i77 # in Loop: Header=BB5_19 Depth=1 movl $10, %edi callq putchar@PLT incq %r14 addl %ebx, %ebp cmpq %r12, %r14 movq 136(%rsp), %r13 # 8-byte Reload jne .LBB5_19 .LBB5_22: # %_Z14imprimirMatrizPii.exit movq 176(%rsp), %rdi # 8-byte Reload movl $1, %esi movabsq $17179869188, %rdx # imm = 0x400000004 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_24 # %bb.23: movq 16(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %ebx, 12(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 96(%rsp), %rax movq %rax, 152(%rsp) leaq 88(%rsp), %rax movq %rax, 160(%rsp) leaq 12(%rsp), %rax movq %rax, 168(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z13sumarMatricesPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_24: movq 16(%rsp), %rsi movq %r13, %rdi movq 112(%rsp), %rdx # 8-byte Reload movl $2, %ecx callq hipMemcpy movl $.Lstr, %edi callq puts@PLT testl %ebx, %ebx jle .LBB5_29 # %bb.25: # %.preheader.lr.ph.i84 xorl %ebp, %ebp xorl %r14d, %r14d .p2align 4, 0x90 .LBB5_26: # %.preheader.i86 # =>This Loop Header: Depth=1 # Child Loop BB5_27 Depth 2 movl %ebp, %eax leaq (,%rax,4), %r15 addq %r13, %r15 xorl %r13d, %r13d .p2align 4, 0x90 .LBB5_27: # Parent Loop BB5_26 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r13,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB5_27 # %bb.28: # %._crit_edge.i92 # in Loop: Header=BB5_26 Depth=1 movl $10, %edi callq putchar@PLT incq %r14 addl %ebx, %ebp cmpq %r12, %r14 movq 136(%rsp), %r13 # 8-byte Reload jne .LBB5_26 .LBB5_29: # %_Z14imprimirMatrizPii.exit96 movl $.Lstr.1, %edi callq puts@PLT movq 128(%rsp), %rdi # 8-byte Reload callq free movq 120(%rsp), %rdi # 8-byte Reload callq free movq %r13, %rdi callq free movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree xorl %eax, %eax addq $184, %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_end5: .size main, .Lfunc_end5-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 .LBB6_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB6_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19multiplicarMatricesPiS_S_i, %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 $_Z13sumarMatricesPiS_S_i, %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_end6: .size __hip_module_ctor, .Lfunc_end6-__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 .LBB7_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 .LBB7_2: retq .Lfunc_end7: .size __hip_module_dtor, .Lfunc_end7-__hip_module_dtor .cfi_endproc # -- End function .type _Z19multiplicarMatricesPiS_S_i,@object # @_Z19multiplicarMatricesPiS_S_i .section .rodata,"a",@progbits .globl _Z19multiplicarMatricesPiS_S_i .p2align 3, 0x0 _Z19multiplicarMatricesPiS_S_i: .quad _Z34__device_stub__multiplicarMatricesPiS_S_i .size _Z19multiplicarMatricesPiS_S_i, 8 .type _Z13sumarMatricesPiS_S_i,@object # @_Z13sumarMatricesPiS_S_i .globl _Z13sumarMatricesPiS_S_i .p2align 3, 0x0 _Z13sumarMatricesPiS_S_i: .quad _Z28__device_stub__sumarMatricesPiS_S_i .size _Z13sumarMatricesPiS_S_i, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d " .size .L.str, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19multiplicarMatricesPiS_S_i" .size .L__unnamed_1, 31 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z13sumarMatricesPiS_S_i" .size .L__unnamed_2, 25 .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 "==============================================RESULTADO============================================" .size .Lstr, 100 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "===================================================================================================" .size .Lstr.1, 100 .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__multiplicarMatricesPiS_S_i .addrsig_sym _Z28__device_stub__sumarMatricesPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19multiplicarMatricesPiS_S_i .addrsig_sym _Z13sumarMatricesPiS_S_i .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 : _Z13sumarMatricesPiS_S_i .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 R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0030*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000e680000002600 */ /*0040*/ S2R R5, SR_TID.Y ; /* 0x0000000000057919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R3, R2, c[0x0][0x4], R5 ; /* 0x0000010002037a24 */ /* 0x002fca00078e0205 */ /*0080*/ ISETP.GE.OR P0, PT, R3, c[0x0][0x178], P0 ; /* 0x00005e0003007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*00b0*/ IMAD R0, R3, c[0x0][0x178], R0 ; /* 0x00005e0003007a24 */ /* 0x000fe200078e0200 */ /*00c0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*00d0*/ IMAD.WIDE R4, R0, R7, c[0x0][0x170] ; /* 0x00005c0000047625 */ /* 0x000fc800078e0207 */ /*00e0*/ IMAD.WIDE R2, R0.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0000027625 */ /* 0x0c0fe400078e0207 */ /*00f0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*0100*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0110*/ IMAD.WIDE R6, R0, R7, c[0x0][0x160] ; /* 0x0000580000067625 */ /* 0x000fe200078e0207 */ /*0120*/ IADD3 R9, R4, R3, RZ ; /* 0x0000000304097210 */ /* 0x004fca0007ffe0ff */ /*0130*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*0140*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0150*/ BRA 0x150; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ .......... Function : _Z19multiplicarMatricesPiS_S_i .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 R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e280000002100 */ /*0030*/ S2R R0, SR_CTAID.Y ; /* 0x0000000000007919 */ /* 0x000e680000002600 */ /*0040*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000e620000002200 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R5 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0205 */ /*0060*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x178], PT ; /* 0x00005e0002007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R5, R0, c[0x0][0x4], R3 ; /* 0x0000010000057a24 */ /* 0x002fca00078e0203 */ /*0080*/ ISETP.GE.OR P0, PT, R5, c[0x0][0x178], P0 ; /* 0x00005e0005007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ MOV R4, c[0x0][0x178] ; /* 0x00005e0000047a02 */ /* 0x000fe20000000f00 */ /*00b0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00c0*/ HFMA2.MMA R18, -RZ, RZ, 0, 0 ; /* 0x00000000ff127435 */ /* 0x000fe400000001ff */ /*00d0*/ ISETP.GE.AND P0, PT, R4, 0x1, PT ; /* 0x000000010400780c */ /* 0x000fda0003f06270 */ /*00e0*/ @!P0 BRA 0xc10 ; /* 0x00000b2000008947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R6, R4.reuse, -0x1, RZ ; /* 0xffffffff04067810 */ /* 0x040fe40007ffe0ff */ /*0100*/ LOP3.LUT R7, R4, 0x3, RZ, 0xc0, !PT ; /* 0x0000000304077812 */ /* 0x000fe400078ec0ff */ /*0110*/ ISETP.GE.U32.AND P0, PT, R6, 0x3, PT ; /* 0x000000030600780c */ /* 0x000fe40003f06070 */ /*0120*/ MOV R6, RZ ; /* 0x000000ff00067202 */ /* 0x000fe40000000f00 */ /*0130*/ MOV R18, RZ ; /* 0x000000ff00127202 */ /* 0x000fd20000000f00 */ /*0140*/ @!P0 BRA 0xb00 ; /* 0x000009b000008947 */ /* 0x000fea0003800000 */ /*0150*/ IADD3 R8, -R7, c[0x0][0x178], RZ ; /* 0x00005e0007087a10 */ /* 0x000fe20007ffe1ff */ /*0160*/ HFMA2.MMA R27, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff1b7435 */ /* 0x000fe200000001ff */ /*0170*/ IADD3 R9, R3, c[0x0][0x178], RZ ; /* 0x00005e0003097a10 */ /* 0x000fe20007ffe0ff */ /*0180*/ ULDC.64 UR6, c[0x0][0x168] ; /* 0x00005a0000067ab9 */ /* 0x000fe20000000a00 */ /*0190*/ ISETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe40003f04270 */ /*01a0*/ MOV R6, RZ ; /* 0x000000ff00067202 */ /* 0x000fe20000000f00 */ /*01b0*/ IMAD R9, R0, c[0x0][0x4], R9 ; /* 0x0000010000097a24 */ /* 0x000fc800078e0209 */ /*01c0*/ IMAD.WIDE R26, R2, R27, c[0x0][0x170] ; /* 0x00005c00021a7625 */ /* 0x000fcc00078e021b */ /*01d0*/ @!P0 BRA 0x970 ; /* 0x0000079000008947 */ /* 0x000fea0003800000 */ /*01e0*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*01f0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*0200*/ @!P1 BRA 0x6b0 ; /* 0x000004a000009947 */ /* 0x000fea0003800000 */ /*0210*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0220*/ MOV R12, UR6 ; /* 0x00000006000c7c02 */ /* 0x000fe20008000f00 */ /*0230*/ LDG.E R19, [R26.64] ; /* 0x000000041a137981 */ /* 0x0000a2000c1e1900 */ /*0240*/ MOV R13, UR7 ; /* 0x00000007000d7c02 */ /* 0x000fca0008000f00 */ /*0250*/ IMAD.WIDE R12, R9, 0x4, R12 ; /* 0x00000004090c7825 */ /* 0x000fca00078e020c */ /*0260*/ LDG.E R28, [R12.64] ; /* 0x000000040c1c7981 */ /* 0x000ea2000c1e1900 */ /*0270*/ IMAD.WIDE R16, R4, 0x4, R26 ; /* 0x0000000404107825 */ /* 0x000fc600078e021a */ /*0280*/ LDG.E R23, [R12.64+0x4] ; /* 0x000004040c177981 */ /* 0x000ee6000c1e1900 */ /*0290*/ IMAD.WIDE R10, R4.reuse, 0x4, R16 ; /* 0x00000004040a7825 */ /* 0x040fe200078e0210 */ /*02a0*/ LDG.E R22, [R16.64] ; /* 0x0000000410167981 */ /* 0x0002e8000c1e1900 */ /*02b0*/ LDG.E R25, [R10.64] ; /* 0x000000040a197981 */ /* 0x000968000c1e1900 */ /*02c0*/ LDG.E R24, [R12.64+0x8] ; /* 0x000008040c187981 */ /* 0x000f68000c1e1900 */ /*02d0*/ LDG.E R14, [R12.64+0xc] ; /* 0x00000c040c0e7981 */ /* 0x000f62000c1e1900 */ /*02e0*/ IMAD.WIDE R10, R4, 0x4, R10 ; /* 0x00000004040a7825 */ /* 0x010fc600078e020a */ /*02f0*/ LDG.E R26, [R12.64+0x10] ; /* 0x000010040c1a7981 */ /* 0x001f28000c1e1900 */ /*0300*/ LDG.E R15, [R10.64] ; /* 0x000000040a0f7981 */ /* 0x000122000c1e1900 */ /*0310*/ IMAD.WIDE R20, R4, 0x4, R10 ; /* 0x0000000404147825 */ /* 0x000fca00078e020a */ /*0320*/ LDG.E R27, [R20.64] ; /* 0x00000004141b7981 */ /* 0x000122000c1e1900 */ /*0330*/ IMAD.WIDE R16, R4, 0x4, R20 ; /* 0x0000000404107825 */ /* 0x002fc600078e0214 */ /*0340*/ LDG.E R10, [R12.64+0x18] ; /* 0x000018040c0a7981 */ /* 0x001f28000c1e1900 */ /*0350*/ LDG.E R21, [R12.64+0x1c] ; /* 0x00001c040c157981 */ /* 0x000f22000c1e1900 */ /*0360*/ IMAD R28, R19, R28, R18 ; /* 0x0000001c131c7224 */ /* 0x004fc600078e0212 */ /*0370*/ LDG.E R18, [R12.64+0x14] ; /* 0x000014040c127981 */ /* 0x000ea8000c1e1900 */ /*0380*/ LDG.E R19, [R16.64] ; /* 0x0000000410137981 */ /* 0x0000a2000c1e1900 */ /*0390*/ IMAD R28, R22, R23, R28 ; /* 0x00000017161c7224 */ /* 0x008fe400078e021c */ /*03a0*/ IMAD.WIDE R16, R4, 0x4, R16 ; /* 0x0000000404107825 */ /* 0x001fca00078e0210 */ /*03b0*/ LDG.E R11, [R16.64] ; /* 0x00000004100b7981 */ /* 0x000ae2000c1e1900 */ /*03c0*/ IMAD.WIDE R22, R4, 0x4, R16 ; /* 0x0000000404167825 */ /* 0x000fca00078e0210 */ /*03d0*/ LDG.E R20, [R22.64] ; /* 0x0000000416147981 */ /* 0x0000e2000c1e1900 */ /*03e0*/ IMAD R16, R25, R24, R28 ; /* 0x0000001819107224 */ /* 0x020fe400078e021c */ /*03f0*/ IMAD.WIDE R24, R4.reuse, 0x4, R22 ; /* 0x0000000404187825 */ /* 0x040fe200078e0216 */ /*0400*/ LDG.E R28, [R12.64+0x20] ; /* 0x000020040c1c7981 */ /* 0x000f68000c1e1900 */ /*0410*/ LDG.E R29, [R24.64] ; /* 0x00000004181d7981 */ /* 0x000362000c1e1900 */ /*0420*/ IMAD R16, R15, R14, R16 ; /* 0x0000000e0f107224 */ /* 0x010fe400078e0210 */ /*0430*/ IMAD.WIDE R14, R4, 0x4, R24 ; /* 0x00000004040e7825 */ /* 0x000fe200078e0218 */ /*0440*/ LDG.E R23, [R12.64+0x24] ; /* 0x000024040c177981 */ /* 0x001f26000c1e1900 */ /*0450*/ IMAD R26, R27, R26, R16 ; /* 0x0000001a1b1a7224 */ /* 0x000fc400078e0210 */ /*0460*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fe200078e020e */ /*0470*/ LDG.E R27, [R12.64+0x28] ; /* 0x000028040c1b7981 */ /* 0x000f28000c1e1900 */ /*0480*/ LDG.E R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000128000c1e1900 */ /*0490*/ LDG.E R22, [R16.64] ; /* 0x0000000410167981 */ /* 0x000328000c1e1900 */ /*04a0*/ LDG.E R15, [R12.64+0x30] ; /* 0x000030040c0f7981 */ /* 0x001f22000c1e1900 */ /*04b0*/ IMAD R26, R19, R18, R26 ; /* 0x00000012131a7224 */ /* 0x004fc400078e021a */ /*04c0*/ IMAD.WIDE R18, R4, 0x4, R16 ; /* 0x0000000404127825 */ /* 0x000fc800078e0210 */ /*04d0*/ IMAD R26, R11, R10, R26 ; /* 0x0000000a0b1a7224 */ /* 0x008fe400078e021a */ /*04e0*/ IMAD.WIDE R10, R4, 0x4, R18 ; /* 0x00000004040a7825 */ /* 0x000fe400078e0212 */ /*04f0*/ LDG.E R18, [R18.64] ; /* 0x0000000412127981 */ /* 0x0000a4000c1e1900 */ /*0500*/ IMAD R24, R20, R21, R26 ; /* 0x0000001514187224 */ /* 0x002fe400078e021a */ /*0510*/ IMAD.WIDE R20, R4, 0x4, R10 ; /* 0x0000000404147825 */ /* 0x000fe200078e020a */ /*0520*/ LDG.E R26, [R12.64+0x2c] ; /* 0x00002c040c1a7981 */ /* 0x000ea8000c1e1900 */ /*0530*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x0002e2000c1e1900 */ /*0540*/ IMAD R28, R29, R28, R24 ; /* 0x0000001c1d1c7224 */ /* 0x020fc600078e0218 */ /*0550*/ LDG.E R19, [R12.64+0x38] ; /* 0x000038040c137981 */ /* 0x001f62000c1e1900 */ /*0560*/ IMAD.WIDE R24, R4, 0x4, R20 ; /* 0x0000000404187825 */ /* 0x000fc600078e0214 */ /*0570*/ LDG.E R20, [R20.64] ; /* 0x0000000414147981 */ /* 0x000168000c1e1900 */ /*0580*/ LDG.E R11, [R12.64+0x34] ; /* 0x000034040c0b7981 */ /* 0x002f62000c1e1900 */ /*0590*/ IMAD.WIDE R16, R4, 0x4, R24 ; /* 0x0000000404107825 */ /* 0x000fc600078e0218 */ /*05a0*/ LDG.E R29, [R24.64] ; /* 0x00000004181d7981 */ /* 0x000368000c1e1900 */ /*05b0*/ LDG.E R21, [R16.64] ; /* 0x0000000410157981 */ /* 0x001f68000c1e1900 */ /*05c0*/ LDG.E R24, [R12.64+0x3c] ; /* 0x00003c040c187981 */ /* 0x002f62000c1e1900 */ /*05d0*/ IMAD R14, R14, R23, R28 ; /* 0x000000170e0e7224 */ /* 0x010fe200078e021c */ /*05e0*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fc60007ffe0ff */ /*05f0*/ IMAD R27, R22, R27, R14 ; /* 0x0000001b161b7224 */ /* 0x000fe200078e020e */ /*0600*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe20003f24270 */ /*0610*/ UIADD3 UR6, UP0, UR6, 0x40, URZ ; /* 0x0000004006067890 */ /* 0x000fe2000ff1e03f */ /*0620*/ IADD3 R6, R6, 0x10, RZ ; /* 0x0000001006067810 */ /* 0x000fc60007ffe0ff */ /*0630*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0640*/ IMAD R18, R18, R26, R27 ; /* 0x0000001a12127224 */ /* 0x004fc800078e021b */ /*0650*/ IMAD R10, R10, R15, R18 ; /* 0x0000000f0a0a7224 */ /* 0x008fe400078e0212 */ /*0660*/ IMAD.WIDE R26, R4, 0x4, R16 ; /* 0x00000004041a7825 */ /* 0x000fc800078e0210 */ /*0670*/ IMAD R10, R20, R11, R10 ; /* 0x0000000b140a7224 */ /* 0x020fc800078e020a */ /*0680*/ IMAD R10, R29, R19, R10 ; /* 0x000000131d0a7224 */ /* 0x000fc800078e020a */ /*0690*/ IMAD R18, R21, R24, R10 ; /* 0x0000001815127224 */ /* 0x000fe200078e020a */ /*06a0*/ @P1 BRA 0x220 ; /* 0xfffffb7000001947 */ /* 0x000fea000383ffff */ /*06b0*/ ISETP.GT.AND P1, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f24270 */ /*06c0*/ @!P1 BRA 0x950 ; /* 0x0000028000009947 */ /* 0x000fea0003800000 */ /*06d0*/ MOV R10, UR6 ; /* 0x00000006000a7c02 */ /* 0x000fe20008000f00 */ /*06e0*/ LDG.E R19, [R26.64] ; /* 0x000000041a137981 */ /* 0x000ea2000c1e1900 */ /*06f0*/ MOV R11, UR7 ; /* 0x00000007000b7c02 */ /* 0x000fca0008000f00 */ /*0700*/ IMAD.WIDE R10, R9, 0x4, R10 ; /* 0x00000004090a7825 */ /* 0x000fca00078e020a */ /*0710*/ LDG.E R24, [R10.64] ; /* 0x000000040a187981 */ /* 0x000ea2000c1e1900 */ /*0720*/ IMAD.WIDE R22, R4, 0x4, R26 ; /* 0x0000000404167825 */ /* 0x000fc600078e021a */ /*0730*/ LDG.E R25, [R10.64+0x4] ; /* 0x000004040a197981 */ /* 0x000ee6000c1e1900 */ /*0740*/ IMAD.WIDE R14, R4.reuse, 0x4, R22 ; /* 0x00000004040e7825 */ /* 0x040fe200078e0216 */ /*0750*/ LDG.E R29, [R10.64+0x8] ; /* 0x000008040a1d7981 */ /* 0x000f28000c1e1900 */ /*0760*/ LDG.E R22, [R22.64] ; /* 0x0000000416167981 */ /* 0x0000e2000c1e1900 */ /*0770*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fc600078e020e */ /*0780*/ LDG.E R14, [R14.64] ; /* 0x000000040e0e7981 */ /* 0x000326000c1e1900 */ /*0790*/ IMAD.WIDE R12, R4.reuse, 0x4, R16 ; /* 0x00000004040c7825 */ /* 0x040fe200078e0210 */ /*07a0*/ LDG.E R28, [R10.64+0xc] ; /* 0x00000c040a1c7981 */ /* 0x000f68000c1e1900 */ /*07b0*/ LDG.E R17, [R16.64] ; /* 0x0000000410117981 */ /* 0x000362000c1e1900 */ /*07c0*/ IMAD.WIDE R20, R4, 0x4, R12 ; /* 0x0000000404147825 */ /* 0x000fc600078e020c */ /*07d0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000328000c1e1900 */ /*07e0*/ LDG.E R23, [R10.64+0x14] ; /* 0x000014040a177981 */ /* 0x001f68000c1e1900 */ /*07f0*/ LDG.E R16, [R10.64+0x18] ; /* 0x000018040a107981 */ /* 0x002f68000c1e1900 */ /*0800*/ LDG.E R13, [R10.64+0x10] ; /* 0x000010040a0d7981 */ /* 0x000f62000c1e1900 */ /*0810*/ IMAD R24, R19, R24, R18 ; /* 0x0000001813187224 */ /* 0x004fc400078e0212 */ /*0820*/ IMAD.WIDE R18, R4.reuse, 0x4, R20 ; /* 0x0000000404127825 */ /* 0x040fe400078e0214 */ /*0830*/ LDG.E R20, [R20.64] ; /* 0x0000000414147981 */ /* 0x000ea8000c1e1900 */ /*0840*/ IMAD.WIDE R26, R4, 0x4, R18 ; /* 0x00000004041a7825 */ /* 0x000fe200078e0212 */ /*0850*/ LDG.E R15, [R18.64] ; /* 0x00000004120f7981 */ /* 0x0000a8000c1e1900 */ /*0860*/ LDG.E R19, [R10.64+0x1c] ; /* 0x00001c040a137981 */ /* 0x001ea8000c1e1900 */ /*0870*/ LDG.E R18, [R26.64] ; /* 0x000000041a127981 */ /* 0x0000a2000c1e1900 */ /*0880*/ IMAD R22, R22, R25, R24 ; /* 0x0000001916167224 */ /* 0x008fc800078e0218 */ /*0890*/ IMAD R14, R14, R29, R22 ; /* 0x0000001d0e0e7224 */ /* 0x010fc800078e0216 */ /*08a0*/ IMAD R14, R17, R28, R14 ; /* 0x0000001c110e7224 */ /* 0x020fc800078e020e */ /*08b0*/ IMAD R12, R12, R13, R14 ; /* 0x0000000d0c0c7224 */ /* 0x000fe200078e020e */ /*08c0*/ UIADD3 UR6, UP0, UR6, 0x20, URZ ; /* 0x0000002006067890 */ /* 0x000fe2000ff1e03f */ /*08d0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe20003f0e170 */ /*08e0*/ IMAD.WIDE R26, R4, 0x4, R26 ; /* 0x00000004041a7825 */ /* 0x001fe200078e021a */ /*08f0*/ IADD3 R6, R6, 0x8, RZ ; /* 0x0000000806067810 */ /* 0x000fe40007ffe0ff */ /*0900*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe20007ffe0ff */ /*0910*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0920*/ IMAD R12, R20, R23, R12 ; /* 0x00000017140c7224 */ /* 0x004fc800078e020c */ /*0930*/ IMAD R12, R15, R16, R12 ; /* 0x000000100f0c7224 */ /* 0x000fc800078e020c */ /*0940*/ IMAD R18, R18, R19, R12 ; /* 0x0000001312127224 */ /* 0x000fe400078e020c */ /*0950*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0960*/ @!P0 BRA 0xb00 ; /* 0x0000019000008947 */ /* 0x000fea0003800000 */ /*0970*/ MOV R10, UR6 ; /* 0x00000006000a7c02 */ /* 0x000fe20008000f00 */ /*0980*/ IMAD.WIDE R12, R4, 0x4, R26 ; /* 0x00000004040c7825 */ /* 0x000fe200078e021a */ /*0990*/ MOV R11, UR7 ; /* 0x00000007000b7c02 */ /* 0x000fe20008000f00 */ /*09a0*/ LDG.E R27, [R26.64] ; /* 0x000000041a1b7981 */ /* 0x000ea8000c1e1900 */ /*09b0*/ IMAD.WIDE R10, R9, 0x4, R10 ; /* 0x00000004090a7825 */ /* 0x000fc800078e020a */ /*09c0*/ IMAD.WIDE R14, R4.reuse, 0x4, R12 ; /* 0x00000004040e7825 */ /* 0x040fe200078e020c */ /*09d0*/ LDG.E R19, [R10.64] ; /* 0x000000040a137981 */ /* 0x000ea8000c1e1900 */ /*09e0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000ee2000c1e1900 */ /*09f0*/ IMAD.WIDE R16, R4, 0x4, R14 ; /* 0x0000000404107825 */ /* 0x000fc600078e020e */ /*0a00*/ LDG.E R20, [R10.64+0x4] ; /* 0x000004040a147981 */ /* 0x000ee8000c1e1900 */ /*0a10*/ LDG.E R22, [R14.64] ; /* 0x000000040e167981 */ /* 0x000f28000c1e1900 */ /*0a20*/ LDG.E R21, [R10.64+0x8] ; /* 0x000008040a157981 */ /* 0x000f28000c1e1900 */ /*0a30*/ LDG.E R23, [R10.64+0xc] ; /* 0x00000c040a177981 */ /* 0x000f68000c1e1900 */ /*0a40*/ LDG.E R24, [R16.64] ; /* 0x0000000410187981 */ /* 0x000f62000c1e1900 */ /*0a50*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fc80007ffe0ff */ /*0a60*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f05270 */ /*0a70*/ UIADD3 UR6, UP0, UR6, 0x10, URZ ; /* 0x0000001006067890 */ /* 0x000fe2000ff1e03f */ /*0a80*/ IADD3 R6, R6, 0x4, RZ ; /* 0x0000000406067810 */ /* 0x000fc60007ffe0ff */ /*0a90*/ UIADD3.X UR7, URZ, UR7, URZ, UP0, !UPT ; /* 0x000000073f077290 */ /* 0x000fe200087fe43f */ /*0aa0*/ IMAD R19, R27, R19, R18 ; /* 0x000000131b137224 */ /* 0x004fc800078e0212 */ /*0ab0*/ IMAD R19, R12, R20, R19 ; /* 0x000000140c137224 */ /* 0x008fe400078e0213 */ /*0ac0*/ IMAD.WIDE R26, R4, 0x4, R16 ; /* 0x00000004041a7825 */ /* 0x000fc800078e0210 */ /*0ad0*/ IMAD R19, R22, R21, R19 ; /* 0x0000001516137224 */ /* 0x010fc800078e0213 */ /*0ae0*/ IMAD R18, R24, R23, R19 ; /* 0x0000001718127224 */ /* 0x020fe200078e0213 */ /*0af0*/ @P0 BRA 0x970 ; /* 0xfffffe7000000947 */ /* 0x000fea000383ffff */ /*0b00*/ ISETP.NE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fda0003f05270 */ /*0b10*/ @!P0 BRA 0xc10 ; /* 0x000000f000008947 */ /* 0x000fea0003800000 */ /*0b20*/ HFMA2.MMA R10, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0a7435 */ /* 0x000fe200000001ff */ /*0b30*/ IADD3 R3, R3, c[0x0][0x178], R6 ; /* 0x00005e0003037a10 */ /* 0x000fe20007ffe006 */ /*0b40*/ IMAD R6, R6, c[0x0][0x178], R2 ; /* 0x00005e0006067a24 */ /* 0x000fc800078e0202 */ /*0b50*/ IMAD R3, R0, c[0x0][0x4], R3 ; /* 0x0000010000037a24 */ /* 0x000fc800078e0203 */ /*0b60*/ IMAD.WIDE R8, R3, R10, c[0x0][0x168] ; /* 0x00005a0003087625 */ /* 0x000fc800078e020a */ /*0b70*/ IMAD.WIDE R10, R6, R10, c[0x0][0x170] ; /* 0x00005c00060a7625 */ /* 0x000fca00078e020a */ /*0b80*/ LDG.E R3, [R10.64] ; /* 0x000000040a037981 */ /* 0x0000a8000c1e1900 */ /*0b90*/ LDG.E R0, [R8.64] ; /* 0x0000000408007981 */ /* 0x0002a2000c1e1900 */ /*0ba0*/ IADD3 R7, R7, -0x1, RZ ; /* 0xffffffff07077810 */ /* 0x000fc80007ffe0ff */ /*0bb0*/ ISETP.NE.AND P0, PT, R7, RZ, PT ; /* 0x000000ff0700720c */ /* 0x000fe20003f05270 */ /*0bc0*/ IMAD.WIDE R10, R4, 0x4, R10 ; /* 0x00000004040a7825 */ /* 0x001fe200078e020a */ /*0bd0*/ IADD3 R8, P1, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x002fc80007f3e0ff */ /*0be0*/ IADD3.X R9, RZ, R9, RZ, P1, !PT ; /* 0x00000009ff097210 */ /* 0x000fe20000ffe4ff */ /*0bf0*/ IMAD R18, R3, R0, R18 ; /* 0x0000000003127224 */ /* 0x004fcc00078e0212 */ /*0c00*/ @P0 BRA 0xb80 ; /* 0xffffff7000000947 */ /* 0x000fea000383ffff */ /*0c10*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fe20000000f00 */ /*0c20*/ IMAD R2, R5, c[0x0][0x178], R2 ; /* 0x00005e0005027a24 */ /* 0x000fc800078e0202 */ /*0c30*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*0c40*/ STG.E [R2.64], R18 ; /* 0x0000001202007986 */ /* 0x000fe2000c101904 */ /*0c50*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0c60*/ BRA 0xc60; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0c70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ca0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ce0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cf0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z19multiplicarMatricesPiS_S_i .globl _Z19multiplicarMatricesPiS_S_i .p2align 8 .type _Z19multiplicarMatricesPiS_S_i,@function _Z19multiplicarMatricesPiS_S_i: s_clause 0x1 s_load_b32 s3, s[0:1], 0x2c s_load_b32 s2, s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s3, 0xffff s_lshr_b32 s3, s3, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mad_u64_u32 v[0:1], null, s14, s4, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s3, v[3:4] s_mov_b32 s3, exec_lo v_max_i32_e32 v2, v0, v1 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s2, v2 s_cbranch_execz .LBB0_6 s_cmp_lt_i32 s2, 1 s_cbranch_scc1 .LBB0_4 s_load_b128 s[4:7], s[0:1], 0x8 v_dual_mov_b32 v2, 0 :: v_dual_add_nc_u32 v5, s2, v1 v_mov_b32_e32 v3, v0 s_mov_b32 s3, 0 .p2align 6 .LBB0_3: s_delay_alu instid0(VALU_DEP_2) | instid1(SALU_CYCLE_1) v_add_nc_u32_e32 v6, s3, v5 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_ashrrev_i32_e32 v4, 31, v3 s_add_i32 s3, s3, 1 s_cmp_eq_u32 s2, s3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v7, 31, v6 v_lshlrev_b64 v[8:9], 2, v[3:4] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_lshlrev_b64 v[6:7], 2, v[6:7] s_waitcnt lgkmcnt(0) v_add_co_u32 v8, vcc_lo, s6, v8 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_ci_u32_e32 v9, vcc_lo, s7, v9, vcc_lo v_add_co_u32 v6, vcc_lo, s4, v6 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v7, vcc_lo, s5, v7, vcc_lo global_load_b32 v4, v[8:9], off global_load_b32 v8, v[6:7], off s_waitcnt vmcnt(0) v_mad_u64_u32 v[6:7], null, v4, v8, v[2:3] v_dual_mov_b32 v2, v6 :: v_dual_add_nc_u32 v3, s2, v3 s_cbranch_scc0 .LBB0_3 s_branch .LBB0_5 .LBB0_4: v_mov_b32_e32 v2, 0 .LBB0_5: s_load_b64 s[0:1], s[0:1], 0x0 v_mad_u64_u32 v[3:4], null, v1, s2, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v4, 31, v3 v_lshlrev_b64 v[0:1], 2, v[3:4] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v2, off .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z19multiplicarMatricesPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .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 10 .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 _Z19multiplicarMatricesPiS_S_i, .Lfunc_end0-_Z19multiplicarMatricesPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z13sumarMatricesPiS_S_i .globl _Z13sumarMatricesPiS_S_i .p2align 8 .type _Z13sumarMatricesPiS_S_i,@function _Z13sumarMatricesPiS_S_i: s_clause 0x1 s_load_b32 s3, s[0:1], 0x2c s_load_b32 s2, s[0:1], 0x18 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s4, s3, 0xffff s_lshr_b32 s3, s3, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mad_u64_u32 v[0:1], null, s14, s4, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s3, v[3:4] s_mov_b32 s3, exec_lo v_max_i32_e32 v2, v0, v1 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s2, v2 s_cbranch_execz .LBB1_2 s_load_b128 s[4:7], s[0:1], 0x0 v_mad_u64_u32 v[2:3], null, v1, s2, v[0:1] s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[0:1], 2, v[2:3] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s1, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s4, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s5, v1, vcc_lo s_waitcnt vmcnt(0) v_add_nc_u32_e32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB1_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13sumarMatricesPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .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 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_end1: .size _Z13sumarMatricesPiS_S_i, .Lfunc_end1-_Z13sumarMatricesPiS_S_i .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: 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: _Z19multiplicarMatricesPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z19multiplicarMatricesPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 10 .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 - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .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: _Z13sumarMatricesPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13sumarMatricesPiS_S_i.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_0010c29c_00000000-6_CUDA_matrix.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2063: .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 .LFE2063: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z5techoii .type _Z5techoii, @function _Z5techoii: .LFB2057: .cfi_startproc endbr64 movl %edi, %eax cltd idivl %esi cmpl $1, %edx sbbl $-1, %eax ret .cfi_endproc .LFE2057: .size _Z5techoii, .-_Z5techoii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d " .LC1: .string "\n" .text .globl _Z14imprimirMatrizPii .type _Z14imprimirMatrizPii, @function _Z14imprimirMatrizPii: .LFB2058: .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 $24, %rsp .cfi_def_cfa_offset 80 movl %esi, 12(%rsp) testl %esi, %esi jle .L4 movslq %esi, %r14 leaq 0(,%r14,4), %r15 leaq (%rdi,%r15), %rbp negq %r14 salq $2, %r14 movl $0, %r13d leaq .LC0(%rip), %r12 .L6: leaq 0(%rbp,%r14), %rbx .L7: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %rbp, %rbx jne .L7 leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %r13d addq %r15, %rbp cmpl %r13d, 12(%rsp) jne .L6 .L4: addq $24, %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 .cfi_endproc .LFE2058: .size _Z14imprimirMatrizPii, .-_Z14imprimirMatrizPii .globl _Z20newMatrizVectorizadaii .type _Z20newMatrizVectorizadaii, @function _Z20newMatrizVectorizadaii: .LFB2059: .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 movl %edi, %ebp movl %esi, %ebx movslq %edi, %r12 movq %r12, %rdi imulq %r12, %rdi salq $2, %rdi call malloc@PLT testl %ebp, %ebp jle .L10 leaq 0(,%r12,4), %r8 leaq (%rax,%r8), %rcx negq %r12 leaq 0(,%r12,4), %rdi movl $0, %esi .L12: leaq (%rcx,%rdi), %rdx .L13: movl %ebx, (%rdx) addq $4, %rdx cmpq %rcx, %rdx jne .L13 addl $1, %esi addq %r8, %rcx cmpl %esi, %ebp jne .L12 .L10: popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _Z20newMatrizVectorizadaii, .-_Z20newMatrizVectorizadaii .globl _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i .type _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i, @function _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i: .LFB2085: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L20 .L16: movq 136(%rsp), %rax subq %fs:40, %rax jne .L21 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L20: .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 _Z19multiplicarMatricesPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L16 .L21: call __stack_chk_fail@PLT .cfi_endproc .LFE2085: .size _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i, .-_Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i .globl _Z19multiplicarMatricesPiS_S_i .type _Z19multiplicarMatricesPiS_S_i, @function _Z19multiplicarMatricesPiS_S_i: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _Z19multiplicarMatricesPiS_S_i, .-_Z19multiplicarMatricesPiS_S_i .globl _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i .type _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i, @function _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L28 .L24: movq 136(%rsp), %rax subq %fs:40, %rax jne .L29 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L28: .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 _Z13sumarMatricesPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L24 .L29: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i, .-_Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i .globl _Z13sumarMatricesPiS_S_i .type _Z13sumarMatricesPiS_S_i, @function _Z13sumarMatricesPiS_S_i: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z13sumarMatricesPiS_S_i, .-_Z13sumarMatricesPiS_S_i .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "==============================================RESULTADO============================================\n" .align 8 .LC3: .string "===================================================================================================\n" .text .globl main .type main, @function main: .LFB2060: .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 $64, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $12, %ebx cmpl $1, %edi jg .L39 .L33: movl %ebx, %ebp imull %ebx, %ebp movslq %ebp, %rbp salq $2, %rbp movl $2, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r14 movl $3, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r13 movl $0, %esi movl %ebx, %edi call _Z20newMatrizVectorizadaii movq %rax, %r12 leaq 8(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT leaq 16(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT leaq 24(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbp, %rdx movq %r14, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %rbp, %rdx movq %r13, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 40(%rsp) testb $3, %bl setne %dl movzbl %dl, %edx leal 3(%rbx), %eax testl %ebx, %ebx cmovns %ebx, %eax sarl $2, %eax addl %edx, %eax movl %eax, 44(%rsp) movl %eax, 48(%rsp) movl $1, 52(%rsp) movl $4, 32(%rsp) movl $4, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movl $1, %ecx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L40 .L34: movl $2, %ecx movq %rbp, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movl %ebx, %esi movq %r12, %rdi call _Z14imprimirMatrizPii movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl 52(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L41 .L35: movl $2, %ecx movq %rbp, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %esi movq %r12, %rdi call _Z14imprimirMatrizPii leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r14, %rdi call free@PLT movq %r13, %rdi call free@PLT movq %r12, %rdi call free@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L42 movl $0, %eax addq $64, %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 .L39: .cfi_restore_state movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movl %eax, %ebx jmp .L33 .L40: movl %ebx, %ecx movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq 24(%rsp), %rdi call _Z44__device_stub__Z19multiplicarMatricesPiS_S_iPiS_S_i jmp .L34 .L41: movl %ebx, %ecx movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z13sumarMatricesPiS_S_iPiS_S_i jmp .L35 .L42: call __stack_chk_fail@PLT .cfi_endproc .LFE2060: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z13sumarMatricesPiS_S_i" .section .rodata.str1.8 .align 8 .LC5: .string "_Z19multiplicarMatricesPiS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2090: .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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z13sumarMatricesPiS_S_i(%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 .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z19multiplicarMatricesPiS_S_i(%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 .LFE2090: .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 "CUDA_matrix.hip" .globl _Z5techoii # -- Begin function _Z5techoii .p2align 4, 0x90 .type _Z5techoii,@function _Z5techoii: # @_Z5techoii .cfi_startproc # %bb.0: movl %edi, %eax cltd idivl %esi cmpl $1, %edx sbbl $-1, %eax retq .Lfunc_end0: .size _Z5techoii, .Lfunc_end0-_Z5techoii .cfi_endproc # -- End function .globl _Z34__device_stub__multiplicarMatricesPiS_S_i # -- Begin function _Z34__device_stub__multiplicarMatricesPiS_S_i .p2align 4, 0x90 .type _Z34__device_stub__multiplicarMatricesPiS_S_i,@function _Z34__device_stub__multiplicarMatricesPiS_S_i: # @_Z34__device_stub__multiplicarMatricesPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z19multiplicarMatricesPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end1: .size _Z34__device_stub__multiplicarMatricesPiS_S_i, .Lfunc_end1-_Z34__device_stub__multiplicarMatricesPiS_S_i .cfi_endproc # -- End function .globl _Z28__device_stub__sumarMatricesPiS_S_i # -- Begin function _Z28__device_stub__sumarMatricesPiS_S_i .p2align 4, 0x90 .type _Z28__device_stub__sumarMatricesPiS_S_i,@function _Z28__device_stub__sumarMatricesPiS_S_i: # @_Z28__device_stub__sumarMatricesPiS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z13sumarMatricesPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z28__device_stub__sumarMatricesPiS_S_i, .Lfunc_end2-_Z28__device_stub__sumarMatricesPiS_S_i .cfi_endproc # -- End function .globl _Z14imprimirMatrizPii # -- Begin function _Z14imprimirMatrizPii .p2align 4, 0x90 .type _Z14imprimirMatrizPii,@function _Z14imprimirMatrizPii: # @_Z14imprimirMatrizPii .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 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 movq %rdi, (%rsp) # 8-byte Spill testl %esi, %esi jle .LBB3_5 # %bb.1: # %.preheader.lr.ph movl %esi, %ebx movl %esi, %r15d xorl %ebp, %ebp xorl %r12d, %r12d .p2align 4, 0x90 .LBB3_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB3_3 Depth 2 movl %ebp, %eax movq (%rsp), %rcx # 8-byte Reload leaq (%rcx,%rax,4), %r13 xorl %r14d, %r14d .p2align 4, 0x90 .LBB3_3: # Parent Loop BB3_2 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r13,%r14,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r14 cmpq %r14, %r15 jne .LBB3_3 # %bb.4: # %._crit_edge # in Loop: Header=BB3_2 Depth=1 movl $10, %edi callq putchar@PLT incq %r12 addl %ebx, %ebp cmpq %r15, %r12 jne .LBB3_2 .LBB3_5: # %._crit_edge13 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 retq .Lfunc_end3: .size _Z14imprimirMatrizPii, .Lfunc_end3-_Z14imprimirMatrizPii .cfi_endproc # -- End function .globl _Z20newMatrizVectorizadaii # -- Begin function _Z20newMatrizVectorizadaii .p2align 4, 0x90 .type _Z20newMatrizVectorizadaii,@function _Z20newMatrizVectorizadaii: # @_Z20newMatrizVectorizadaii .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 movl %esi, %ebx movl %edi, %ebp movslq %edi, %r14 movq %r14, %rdi imulq %r14, %rdi shlq $2, %rdi callq malloc testl %r14d, %r14d jle .LBB4_5 # %bb.1: # %.preheader.lr.ph movl %ebp, %ecx xorl %edx, %edx xorl %esi, %esi .p2align 4, 0x90 .LBB4_2: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB4_3 Depth 2 movl %edx, %edi leaq (%rax,%rdi,4), %rdi xorl %r8d, %r8d .p2align 4, 0x90 .LBB4_3: # Parent Loop BB4_2 Depth=1 # => This Inner Loop Header: Depth=2 movl %ebx, (%rdi,%r8,4) incq %r8 cmpq %r8, %rcx jne .LBB4_3 # %bb.4: # %._crit_edge # in Loop: Header=BB4_2 Depth=1 incq %rsi addl %ebp, %edx cmpq %rcx, %rsi jne .LBB4_2 .LBB4_5: # %._crit_edge17 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size _Z20newMatrizVectorizadaii, .Lfunc_end4-_Z20newMatrizVectorizadaii .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 $184, %rsp .cfi_def_cfa_offset 240 .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 $12, %ebx cmpl $2, %edi jl .LBB5_2 # %bb.1: movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %rbx .LBB5_2: movl %ebx, %r15d imull %r15d, %r15d movslq %ebx, %r14 movq %r14, %r13 imulq %r14, %r13 shlq $2, %r13 movq %r13, %rdi callq malloc movl %ebx, %r12d testl %r14d, %r14d jle .LBB5_7 # %bb.3: # %.preheader.lr.ph.i xorl %edi, %edi xorl %ecx, %ecx .p2align 4, 0x90 .LBB5_4: # %.preheader.i # =>This Loop Header: Depth=1 # Child Loop BB5_5 Depth 2 movl %edi, %edx leaq (%rax,%rdx,4), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB5_5: # Parent Loop BB5_4 Depth=1 # => This Inner Loop Header: Depth=2 movl $2, (%rdx,%rsi,4) incq %rsi cmpq %rsi, %r12 jne .LBB5_5 # %bb.6: # %._crit_edge.i # in Loop: Header=BB5_4 Depth=1 incq %rcx addl %ebx, %edi cmpq %r12, %rcx jne .LBB5_4 .LBB5_7: # %_Z20newMatrizVectorizadaii.exit movq %rax, 128(%rsp) # 8-byte Spill movq %r13, %rdi callq malloc testl %ebx, %ebx jle .LBB5_12 # %bb.8: # %.preheader.lr.ph.i47 xorl %edi, %edi xorl %ecx, %ecx .p2align 4, 0x90 .LBB5_9: # %.preheader.i49 # =>This Loop Header: Depth=1 # Child Loop BB5_10 Depth 2 movl %edi, %edx leaq (%rax,%rdx,4), %rdx xorl %esi, %esi .p2align 4, 0x90 .LBB5_10: # Parent Loop BB5_9 Depth=1 # => This Inner Loop Header: Depth=2 movl $3, (%rdx,%rsi,4) incq %rsi cmpq %rsi, %r12 jne .LBB5_10 # %bb.11: # %._crit_edge.i55 # in Loop: Header=BB5_9 Depth=1 incq %rcx addl %ebx, %edi cmpq %r12, %rcx jne .LBB5_9 .LBB5_12: # %_Z20newMatrizVectorizadaii.exit58 movq %rax, 120(%rsp) # 8-byte Spill shlq $2, %r15 movq %r15, 112(%rsp) # 8-byte Spill movq %r13, %rdi callq malloc movq %rax, %r13 testl %ebx, %ebx jle .LBB5_15 # %bb.13: # %.preheader.lr.ph.i59 leaq (,%r12,4), %rbp xorl %r14d, %r14d movq %r12, %r15 .p2align 4, 0x90 .LBB5_14: # %.preheader.i61 # =>This Inner Loop Header: Depth=1 movl %r14d, %eax leaq (,%rax,4), %rdi addq %r13, %rdi xorl %esi, %esi movq %rbp, %rdx callq memset@PLT addl %ebx, %r14d decq %r15 jne .LBB5_14 .LBB5_15: # %_Z20newMatrizVectorizadaii.exit70 movabsq $17179869188, %r15 # imm = 0x400000004 leaq 32(%rsp), %rdi movq 112(%rsp), %r14 # 8-byte Reload movq %r14, %rsi callq hipMalloc leaq 24(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 32(%rsp), %rdi movq 128(%rsp), %rsi # 8-byte Reload movq %r14, %rdx movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movq 120(%rsp), %rsi # 8-byte Reload movq %r14, %rdx movl $1, %ecx callq hipMemcpy leal 3(%rbx), %eax testl %ebx, %ebx cmovnsl %ebx, %eax sarl $2, %eax movl %ebx, %ecx andl $3, %ecx cmpl $1, %ecx sbbl $-1, %eax movq %rax, %rdi shlq $32, %rdi orq %rax, %rdi movq %rdi, 176(%rsp) # 8-byte Spill movl $1, %esi movq %r15, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_17 # %bb.16: movq 16(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %ebx, 12(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 96(%rsp), %rax movq %rax, 152(%rsp) leaq 88(%rsp), %rax movq %rax, 160(%rsp) leaq 12(%rsp), %rax movq %rax, 168(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z19multiplicarMatricesPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_17: movq 16(%rsp), %rsi movq %r13, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy testl %ebx, %ebx movq %r13, 136(%rsp) # 8-byte Spill jle .LBB5_22 # %bb.18: # %.preheader.lr.ph.i71 xorl %ebp, %ebp xorl %r14d, %r14d .p2align 4, 0x90 .LBB5_19: # %.preheader.i72 # =>This Loop Header: Depth=1 # Child Loop BB5_20 Depth 2 movl %ebp, %eax leaq (,%rax,4), %r15 addq %r13, %r15 xorl %r13d, %r13d .p2align 4, 0x90 .LBB5_20: # Parent Loop BB5_19 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r13,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB5_20 # %bb.21: # %._crit_edge.i77 # in Loop: Header=BB5_19 Depth=1 movl $10, %edi callq putchar@PLT incq %r14 addl %ebx, %ebp cmpq %r12, %r14 movq 136(%rsp), %r13 # 8-byte Reload jne .LBB5_19 .LBB5_22: # %_Z14imprimirMatrizPii.exit movq 176(%rsp), %rdi # 8-byte Reload movl $1, %esi movabsq $17179869188, %rdx # imm = 0x400000004 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_24 # %bb.23: movq 16(%rsp), %rax movq 32(%rsp), %rcx movq 24(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %ebx, 12(%rsp) leaq 104(%rsp), %rax movq %rax, 144(%rsp) leaq 96(%rsp), %rax movq %rax, 152(%rsp) leaq 88(%rsp), %rax movq %rax, 160(%rsp) leaq 12(%rsp), %rax movq %rax, 168(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 144(%rsp), %r9 movl $_Z13sumarMatricesPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_24: movq 16(%rsp), %rsi movq %r13, %rdi movq 112(%rsp), %rdx # 8-byte Reload movl $2, %ecx callq hipMemcpy movl $.Lstr, %edi callq puts@PLT testl %ebx, %ebx jle .LBB5_29 # %bb.25: # %.preheader.lr.ph.i84 xorl %ebp, %ebp xorl %r14d, %r14d .p2align 4, 0x90 .LBB5_26: # %.preheader.i86 # =>This Loop Header: Depth=1 # Child Loop BB5_27 Depth 2 movl %ebp, %eax leaq (,%rax,4), %r15 addq %r13, %r15 xorl %r13d, %r13d .p2align 4, 0x90 .LBB5_27: # Parent Loop BB5_26 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r15,%r13,4), %esi movl $.L.str, %edi xorl %eax, %eax callq printf incq %r13 cmpq %r13, %r12 jne .LBB5_27 # %bb.28: # %._crit_edge.i92 # in Loop: Header=BB5_26 Depth=1 movl $10, %edi callq putchar@PLT incq %r14 addl %ebx, %ebp cmpq %r12, %r14 movq 136(%rsp), %r13 # 8-byte Reload jne .LBB5_26 .LBB5_29: # %_Z14imprimirMatrizPii.exit96 movl $.Lstr.1, %edi callq puts@PLT movq 128(%rsp), %rdi # 8-byte Reload callq free movq 120(%rsp), %rdi # 8-byte Reload callq free movq %r13, %rdi callq free movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree xorl %eax, %eax addq $184, %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_end5: .size main, .Lfunc_end5-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 .LBB6_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB6_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19multiplicarMatricesPiS_S_i, %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 $_Z13sumarMatricesPiS_S_i, %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_end6: .size __hip_module_ctor, .Lfunc_end6-__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 .LBB7_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 .LBB7_2: retq .Lfunc_end7: .size __hip_module_dtor, .Lfunc_end7-__hip_module_dtor .cfi_endproc # -- End function .type _Z19multiplicarMatricesPiS_S_i,@object # @_Z19multiplicarMatricesPiS_S_i .section .rodata,"a",@progbits .globl _Z19multiplicarMatricesPiS_S_i .p2align 3, 0x0 _Z19multiplicarMatricesPiS_S_i: .quad _Z34__device_stub__multiplicarMatricesPiS_S_i .size _Z19multiplicarMatricesPiS_S_i, 8 .type _Z13sumarMatricesPiS_S_i,@object # @_Z13sumarMatricesPiS_S_i .globl _Z13sumarMatricesPiS_S_i .p2align 3, 0x0 _Z13sumarMatricesPiS_S_i: .quad _Z28__device_stub__sumarMatricesPiS_S_i .size _Z13sumarMatricesPiS_S_i, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d " .size .L.str, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19multiplicarMatricesPiS_S_i" .size .L__unnamed_1, 31 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z13sumarMatricesPiS_S_i" .size .L__unnamed_2, 25 .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 "==============================================RESULTADO============================================" .size .Lstr, 100 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "===================================================================================================" .size .Lstr.1, 100 .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__multiplicarMatricesPiS_S_i .addrsig_sym _Z28__device_stub__sumarMatricesPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z19multiplicarMatricesPiS_S_i .addrsig_sym _Z13sumarMatricesPiS_S_i .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 <cuda.h> #include <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; cudaMalloc(&d_A, sizeM); cudaMemcpy(d_A, h_A, sizeM, cudaMemcpyHostToDevice); cudaMalloc(&d_B, sizeV); cudaMemcpy(d_B, h_B, sizeV, cudaMemcpyHostToDevice); cudaMalloc(&d_C, sizeV); cudaMemcpy(d_C, h_C, sizeV, cudaMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); cudaMemcpy(h_C, d_C, sizeV, cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_C); return 0; }
code for sm_80 Function : _Z16MatVecMul_KernelPfS_S_i .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 R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e240000002100 */ /*0020*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x001fda0003f06270 */ /*0030*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0040*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff067624 */ /* 0x000fe200078e00ff */ /*0050*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0060*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fc600000001ff */ /*0070*/ ISETP.GE.AND P0, PT, R6, 0x1, PT ; /* 0x000000010600780c */ /* 0x000fda0003f06270 */ /*0080*/ @!P0 BRA 0xa50 ; /* 0x000009c000008947 */ /* 0x000fea0003800000 */ /*0090*/ IADD3 R2, R6.reuse, -0x1, RZ ; /* 0xffffffff06027810 */ /* 0x040fe20007ffe0ff */ /*00a0*/ IMAD.MOV.U32 R15, RZ, RZ, RZ ; /* 0x000000ffff0f7224 */ /* 0x000fe200078e00ff */ /*00b0*/ LOP3.LUT R6, R6, 0x3, RZ, 0xc0, !PT ; /* 0x0000000306067812 */ /* 0x000fe400078ec0ff */ /*00c0*/ ISETP.GE.U32.AND P0, PT, R2, 0x3, PT ; /* 0x000000030200780c */ /* 0x000fe40003f06070 */ /*00d0*/ MOV R7, RZ ; /* 0x000000ff00077202 */ /* 0x000fd60000000f00 */ /*00e0*/ @!P0 BRA 0x910 ; /* 0x0000082000008947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R8, -R6, c[0x0][0x178], RZ ; /* 0x00005e0006087a10 */ /* 0x000fe20007ffe1ff */ /*0100*/ IMAD.MOV.U32 R3, RZ, RZ, 0x4 ; /* 0x00000004ff037424 */ /* 0x000fe200078e00ff */ /*0110*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fe200000001ff */ /*0120*/ IMAD R2, R0, c[0x0][0x178], RZ ; /* 0x00005e0000027a24 */ /* 0x000fe200078e02ff */ /*0130*/ ISETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f04270 */ /*0140*/ IMAD.MOV.U32 R7, RZ, RZ, RZ ; /* 0x000000ffff077224 */ /* 0x000fe200078e00ff */ /*0150*/ MOV R4, c[0x0][0x168] ; /* 0x00005a0000047a02 */ /* 0x000fe20000000f00 */ /*0160*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x16c] ; /* 0x00005b00ff057624 */ /* 0x000fe400078e00ff */ /*0170*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fd000078e0203 */ /*0180*/ @!P0 BRA 0x790 ; /* 0x0000060000008947 */ /* 0x000fea0003800000 */ /*0190*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*01a0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*01b0*/ @!P1 BRA 0x550 ; /* 0x0000039000009947 */ /* 0x000fea0003800000 */ /*01c0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*01d0*/ LDG.E R16, [R4.64] ; /* 0x0000000404107981 */ /* 0x000ea8000c1e1900 */ /*01e0*/ LDG.E R17, [R2.64] ; /* 0x0000000402117981 */ /* 0x000ea8000c1e1900 */ /*01f0*/ LDG.E R18, [R4.64+0x4] ; /* 0x0000040404127981 */ /* 0x000ee8000c1e1900 */ /*0200*/ LDG.E R25, [R2.64+0x4] ; /* 0x0000040402197981 */ /* 0x000ee8000c1e1900 */ /*0210*/ LDG.E R19, [R4.64+0x8] ; /* 0x0000080404137981 */ /* 0x000f28000c1e1900 */ /*0220*/ LDG.E R20, [R2.64+0x8] ; /* 0x0000080402147981 */ /* 0x000f28000c1e1900 */ /*0230*/ LDG.E R22, [R4.64+0xc] ; /* 0x00000c0404167981 */ /* 0x000f68000c1e1900 */ /*0240*/ LDG.E R21, [R2.64+0xc] ; /* 0x00000c0402157981 */ /* 0x000f68000c1e1900 */ /*0250*/ LDG.E R23, [R4.64+0x10] ; /* 0x0000100404177981 */ /* 0x000f68000c1e1900 */ /*0260*/ LDG.E R24, [R2.64+0x10] ; /* 0x0000100402187981 */ /* 0x000f68000c1e1900 */ /*0270*/ LDG.E R9, [R4.64+0x14] ; /* 0x0000140404097981 */ /* 0x000f68000c1e1900 */ /*0280*/ LDG.E R10, [R2.64+0x14] ; /* 0x00001404020a7981 */ /* 0x000f68000c1e1900 */ /*0290*/ LDG.E R11, [R4.64+0x18] ; /* 0x00001804040b7981 */ /* 0x000f68000c1e1900 */ /*02a0*/ LDG.E R12, [R2.64+0x18] ; /* 0x00001804020c7981 */ /* 0x000f68000c1e1900 */ /*02b0*/ LDG.E R13, [R4.64+0x1c] ; /* 0x00001c04040d7981 */ /* 0x000f68000c1e1900 */ /*02c0*/ LDG.E R14, [R2.64+0x1c] ; /* 0x00001c04020e7981 */ /* 0x000f62000c1e1900 */ /*02d0*/ FFMA R17, R16, R17, R15 ; /* 0x0000001110117223 */ /* 0x004fc6000000000f */ /*02e0*/ LDG.E R15, [R4.64+0x20] ; /* 0x00002004040f7981 */ /* 0x0000a8000c1e1900 */ /*02f0*/ LDG.E R16, [R2.64+0x20] ; /* 0x0000200402107981 */ /* 0x0002a2000c1e1900 */ /*0300*/ FFMA R25, R18, R25, R17 ; /* 0x0000001912197223 */ /* 0x008fc60000000011 */ /*0310*/ LDG.E R17, [R4.64+0x24] ; /* 0x0000240404117981 */ /* 0x0000e8000c1e1900 */ /*0320*/ LDG.E R18, [R2.64+0x24] ; /* 0x0000240402127981 */ /* 0x0002e2000c1e1900 */ /*0330*/ FFMA R25, R19, R20, R25 ; /* 0x0000001413197223 */ /* 0x010fc60000000019 */ /*0340*/ LDG.E R19, [R4.64+0x28] ; /* 0x0000280404137981 */ /* 0x000128000c1e1900 */ /*0350*/ LDG.E R20, [R2.64+0x28] ; /* 0x0000280402147981 */ /* 0x000322000c1e1900 */ /*0360*/ FFMA R25, R22, R21, R25 ; /* 0x0000001516197223 */ /* 0x020fc60000000019 */ /*0370*/ LDG.E R21, [R4.64+0x2c] ; /* 0x00002c0404157981 */ /* 0x000168000c1e1900 */ /*0380*/ LDG.E R22, [R2.64+0x2c] ; /* 0x00002c0402167981 */ /* 0x000362000c1e1900 */ /*0390*/ FFMA R25, R23, R24, R25 ; /* 0x0000001817197223 */ /* 0x000fc60000000019 */ /*03a0*/ LDG.E R23, [R4.64+0x30] ; /* 0x0000300404177981 */ /* 0x000168000c1e1900 */ /*03b0*/ LDG.E R24, [R2.64+0x30] ; /* 0x0000300402187981 */ /* 0x000362000c1e1900 */ /*03c0*/ FFMA R25, R9, R10, R25 ; /* 0x0000000a09197223 */ /* 0x000fc60000000019 */ /*03d0*/ LDG.E R9, [R4.64+0x34] ; /* 0x0000340404097981 */ /* 0x000168000c1e1900 */ /*03e0*/ LDG.E R10, [R2.64+0x34] ; /* 0x00003404020a7981 */ /* 0x000362000c1e1900 */ /*03f0*/ FFMA R25, R11, R12, R25 ; /* 0x0000000c0b197223 */ /* 0x000fc60000000019 */ /*0400*/ LDG.E R11, [R4.64+0x38] ; /* 0x00003804040b7981 */ /* 0x000168000c1e1900 */ /*0410*/ LDG.E R12, [R2.64+0x38] ; /* 0x00003804020c7981 */ /* 0x000362000c1e1900 */ /*0420*/ FFMA R25, R13, R14, R25 ; /* 0x0000000e0d197223 */ /* 0x000fc60000000019 */ /*0430*/ LDG.E R13, [R4.64+0x3c] ; /* 0x00003c04040d7981 */ /* 0x000168000c1e1900 */ /*0440*/ LDG.E R14, [R2.64+0x3c] ; /* 0x00003c04020e7981 */ /* 0x000362000c1e1900 */ /*0450*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fe40007ffe0ff */ /*0460*/ IADD3 R7, R7, 0x10, RZ ; /* 0x0000001007077810 */ /* 0x000fe40007ffe0ff */ /*0470*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*0480*/ IADD3 R4, P3, R4, 0x40, RZ ; /* 0x0000004004047810 */ /* 0x001fc40007f7e0ff */ /*0490*/ IADD3 R2, P2, R2, 0x40, RZ ; /* 0x0000004002027810 */ /* 0x002fe40007f5e0ff */ /*04a0*/ IADD3.X R5, RZ, R5, RZ, P3, !PT ; /* 0x00000005ff057210 */ /* 0x000fc60001ffe4ff */ /*04b0*/ IMAD.X R3, RZ, RZ, R3, P2 ; /* 0x000000ffff037224 */ /* 0x000fe400010e0603 */ /*04c0*/ FFMA R15, R15, R16, R25 ; /* 0x000000100f0f7223 */ /* 0x004fc80000000019 */ /*04d0*/ FFMA R15, R17, R18, R15 ; /* 0x00000012110f7223 */ /* 0x008fc8000000000f */ /*04e0*/ FFMA R15, R19, R20, R15 ; /* 0x00000014130f7223 */ /* 0x010fc8000000000f */ /*04f0*/ FFMA R15, R21, R22, R15 ; /* 0x00000016150f7223 */ /* 0x020fc8000000000f */ /*0500*/ FFMA R15, R23, R24, R15 ; /* 0x00000018170f7223 */ /* 0x000fc8000000000f */ /*0510*/ FFMA R9, R9, R10, R15 ; /* 0x0000000a09097223 */ /* 0x000fc8000000000f */ /*0520*/ FFMA R9, R11, R12, R9 ; /* 0x0000000c0b097223 */ /* 0x000fc80000000009 */ /*0530*/ FFMA R15, R13, R14, R9 ; /* 0x0000000e0d0f7223 */ /* 0x000fe20000000009 */ /*0540*/ @P1 BRA 0x1d0 ; /* 0xfffffc8000001947 */ /* 0x000fea000383ffff */ /*0550*/ ISETP.GT.AND P1, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f24270 */ /*0560*/ @!P1 BRA 0x770 ; /* 0x0000020000009947 */ /* 0x000fea0003800000 */ /*0570*/ LDG.E R16, [R4.64] ; /* 0x0000000404107981 */ /* 0x0000a8000c1e1900 */ /*0580*/ LDG.E R17, [R2.64] ; /* 0x0000000402117981 */ /* 0x000ea8000c1e1900 */ /*0590*/ LDG.E R19, [R4.64+0x4] ; /* 0x0000040404137981 */ /* 0x0000e8000c1e1900 */ /*05a0*/ LDG.E R18, [R2.64+0x4] ; /* 0x0000040402127981 */ /* 0x000ee8000c1e1900 */ /*05b0*/ LDG.E R21, [R4.64+0x8] ; /* 0x0000080404157981 */ /* 0x000128000c1e1900 */ /*05c0*/ LDG.E R20, [R2.64+0x8] ; /* 0x0000080402147981 */ /* 0x000f28000c1e1900 */ /*05d0*/ LDG.E R23, [R4.64+0xc] ; /* 0x00000c0404177981 */ /* 0x000168000c1e1900 */ /*05e0*/ LDG.E R22, [R2.64+0xc] ; /* 0x00000c0402167981 */ /* 0x000f68000c1e1900 */ /*05f0*/ LDG.E R25, [R4.64+0x10] ; /* 0x0000100404197981 */ /* 0x000168000c1e1900 */ /*0600*/ LDG.E R24, [R2.64+0x10] ; /* 0x0000100402187981 */ /* 0x000f68000c1e1900 */ /*0610*/ LDG.E R13, [R4.64+0x14] ; /* 0x00001404040d7981 */ /* 0x000168000c1e1900 */ /*0620*/ LDG.E R14, [R2.64+0x14] ; /* 0x00001404020e7981 */ /* 0x000f68000c1e1900 */ /*0630*/ LDG.E R11, [R4.64+0x18] ; /* 0x00001804040b7981 */ /* 0x000168000c1e1900 */ /*0640*/ LDG.E R12, [R2.64+0x18] ; /* 0x00001804020c7981 */ /* 0x000f68000c1e1900 */ /*0650*/ LDG.E R9, [R4.64+0x1c] ; /* 0x00001c0404097981 */ /* 0x000168000c1e1900 */ /*0660*/ LDG.E R10, [R2.64+0x1c] ; /* 0x00001c04020a7981 */ /* 0x000362000c1e1900 */ /*0670*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fc40003f0e170 */ /*0680*/ IADD3 R7, R7, 0x8, RZ ; /* 0x0000000807077810 */ /* 0x000fe40007ffe0ff */ /*0690*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe40007ffe0ff */ /*06a0*/ IADD3 R4, P2, R4, 0x20, RZ ; /* 0x0000002004047810 */ /* 0x001fca0007f5e0ff */ /*06b0*/ IMAD.X R5, RZ, RZ, R5, P2 ; /* 0x000000ffff057224 */ /* 0x000fe400010e0605 */ /*06c0*/ FFMA R16, R16, R17, R15 ; /* 0x0000001110107223 */ /* 0x004fc8000000000f */ /*06d0*/ FFMA R16, R19, R18, R16 ; /* 0x0000001213107223 */ /* 0x008fc80000000010 */ /*06e0*/ FFMA R16, R21, R20, R16 ; /* 0x0000001415107223 */ /* 0x010fc80000000010 */ /*06f0*/ FFMA R16, R23, R22, R16 ; /* 0x0000001617107223 */ /* 0x020fc80000000010 */ /*0700*/ FFMA R16, R25, R24, R16 ; /* 0x0000001819107223 */ /* 0x000fc80000000010 */ /*0710*/ FFMA R13, R13, R14, R16 ; /* 0x0000000e0d0d7223 */ /* 0x000fc80000000010 */ /*0720*/ FFMA R11, R11, R12, R13 ; /* 0x0000000c0b0b7223 */ /* 0x000fe2000000000d */ /*0730*/ IADD3 R12, P1, R2, 0x20, RZ ; /* 0x00000020020c7810 */ /* 0x000fc80007f3e0ff */ /*0740*/ IADD3.X R3, RZ, R3, RZ, P1, !PT ; /* 0x00000003ff037210 */ /* 0x002fe40000ffe4ff */ /*0750*/ MOV R2, R12 ; /* 0x0000000c00027202 */ /* 0x000fe20000000f00 */ /*0760*/ FFMA R15, R9, R10, R11 ; /* 0x0000000a090f7223 */ /* 0x000fe4000000000b */ /*0770*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0780*/ @!P0 BRA 0x910 ; /* 0x0000018000008947 */ /* 0x000fea0003800000 */ /*0790*/ LDG.E R10, [R4.64] ; /* 0x00000004040a7981 */ /* 0x000ea8000c1e1900 */ /*07a0*/ LDG.E R9, [R2.64] ; /* 0x0000000402097981 */ /* 0x000ea8000c1e1900 */ /*07b0*/ LDG.E R12, [R4.64+0x4] ; /* 0x00000404040c7981 */ /* 0x0000e8000c1e1900 */ /*07c0*/ LDG.E R11, [R2.64+0x4] ; /* 0x00000404020b7981 */ /* 0x000ee8000c1e1900 */ /*07d0*/ LDG.E R14, [R4.64+0x8] ; /* 0x00000804040e7981 */ /* 0x000128000c1e1900 */ /*07e0*/ LDG.E R13, [R2.64+0x8] ; /* 0x00000804020d7981 */ /* 0x000328000c1e1900 */ /*07f0*/ LDG.E R16, [R4.64+0xc] ; /* 0x00000c0404107981 */ /* 0x000168000c1e1900 */ /*0800*/ LDG.E R17, [R2.64+0xc] ; /* 0x00000c0402117981 */ /* 0x000362000c1e1900 */ /*0810*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fc40007ffe0ff */ /*0820*/ IADD3 R7, R7, 0x4, RZ ; /* 0x0000000407077810 */ /* 0x000fe40007ffe0ff */ /*0830*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f05270 */ /*0840*/ FFMA R9, R10, R9, R15 ; /* 0x000000090a097223 */ /* 0x004fe2000000000f */ /*0850*/ IADD3 R10, P2, R4, 0x10, RZ ; /* 0x00000010040a7810 */ /* 0x000fca0007f5e0ff */ /*0860*/ IMAD.MOV.U32 R4, RZ, RZ, R10 ; /* 0x000000ffff047224 */ /* 0x001fe400078e000a */ /*0870*/ FFMA R9, R12, R11, R9 ; /* 0x0000000b0c097223 */ /* 0x008fe20000000009 */ /*0880*/ IADD3 R12, P1, R2, 0x10, RZ ; /* 0x00000010020c7810 */ /* 0x000fe20007f3e0ff */ /*0890*/ IMAD.X R11, RZ, RZ, R5, P2 ; /* 0x000000ffff0b7224 */ /* 0x000fc800010e0605 */ /*08a0*/ IMAD.MOV.U32 R2, RZ, RZ, R12 ; /* 0x000000ffff027224 */ /* 0x002fe200078e000c */ /*08b0*/ MOV R5, R11 ; /* 0x0000000b00057202 */ /* 0x000fe20000000f00 */ /*08c0*/ FFMA R9, R14, R13, R9 ; /* 0x0000000d0e097223 */ /* 0x010fe20000000009 */ /*08d0*/ IADD3.X R13, RZ, R3, RZ, P1, !PT ; /* 0x00000003ff0d7210 */ /* 0x000fc80000ffe4ff */ /*08e0*/ MOV R3, R13 ; /* 0x0000000d00037202 */ /* 0x000fe20000000f00 */ /*08f0*/ FFMA R15, R16, R17, R9 ; /* 0x00000011100f7223 */ /* 0x020fe20000000009 */ /*0900*/ @P0 BRA 0x790 ; /* 0xfffffe8000000947 */ /* 0x000fea000383ffff */ /*0910*/ ISETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600720c */ /* 0x000fda0003f05270 */ /*0920*/ @!P0 BRA 0xa50 ; /* 0x0000012000008947 */ /* 0x000fea0003800000 */ /*0930*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe400078e00ff */ /*0940*/ IMAD R4, R0, c[0x0][0x178], R7 ; /* 0x00005e0000047a24 */ /* 0x000fe400078e0207 */ /*0950*/ IMAD.WIDE R2, R7, R5, c[0x0][0x168] ; /* 0x00005a0007027625 */ /* 0x000fc800078e0205 */ /*0960*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0970*/ MOV R8, R2 ; /* 0x0000000200087202 */ /* 0x000fc60000000f00 */ /*0980*/ IMAD.MOV.U32 R7, RZ, RZ, R4 ; /* 0x000000ffff077224 */ /* 0x000fc600078e0004 */ /*0990*/ MOV R2, R8 ; /* 0x0000000800027202 */ /* 0x000fe40000000f00 */ /*09a0*/ MOV R4, R7 ; /* 0x0000000700047202 */ /* 0x000fc80000000f00 */ /*09b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x0000a8000c1e1900 */ /*09c0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x0002a2000c1e1900 */ /*09d0*/ IADD3 R6, R6, -0x1, RZ ; /* 0xffffffff06067810 */ /* 0x000fe40007ffe0ff */ /*09e0*/ IADD3 R8, P1, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x000fe40007f3e0ff */ /*09f0*/ ISETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600720c */ /* 0x000fe40003f05270 */ /*0a00*/ IADD3 R7, P2, R7, 0x4, RZ ; /* 0x0000000407077810 */ /* 0x000fe20007f5e0ff */ /*0a10*/ IMAD.X R3, RZ, RZ, R3, P1 ; /* 0x000000ffff037224 */ /* 0x001fc600008e0603 */ /*0a20*/ IADD3.X R5, RZ, R5, RZ, P2, !PT ; /* 0x00000005ff057210 */ /* 0x002fe200017fe4ff */ /*0a30*/ FFMA R15, R2, R4, R15 ; /* 0x00000004020f7223 */ /* 0x004fcc000000000f */ /*0a40*/ @P0 BRA 0x990 ; /* 0xffffff4000000947 */ /* 0x000fea000383ffff */ /*0a50*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fca0000000f00 */ /*0a60*/ IMAD.WIDE R2, R0, R3, c[0x0][0x170] ; /* 0x00005c0000027625 */ /* 0x000fca00078e0203 */ /*0a70*/ STG.E [R2.64], R15 ; /* 0x0000000f02007986 */ /* 0x000fe2000c101904 */ /*0a80*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0a90*/ BRA 0xa90; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b00*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b10*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b20*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cuda.h> #include <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; cudaMalloc(&d_A, sizeM); cudaMemcpy(d_A, h_A, sizeM, cudaMemcpyHostToDevice); cudaMalloc(&d_B, sizeV); cudaMemcpy(d_B, h_B, sizeV, cudaMemcpyHostToDevice); cudaMalloc(&d_C, sizeV); cudaMemcpy(d_C, h_C, sizeV, cudaMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); cudaMemcpy(h_C, d_C, sizeV, cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_C); return 0; }
.file "tmpxft_0018df27_00000000-6_e2.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3778: .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 .LFE3778: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i .type _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i, @function _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i: .LFB3800: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 _Z16MatVecMul_KernelPfS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3800: .size _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i, .-_Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i .globl _Z16MatVecMul_KernelPfS_S_i .type _Z16MatVecMul_KernelPfS_S_i, @function _Z16MatVecMul_KernelPfS_S_i: .LFB3801: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3801: .size _Z16MatVecMul_KernelPfS_S_i, .-_Z16MatVecMul_KernelPfS_S_i .globl _Z13MatVecMul_GPUPfS_S_i .type _Z13MatVecMul_GPUPfS_S_i, @function _Z13MatVecMul_GPUPfS_S_i: .LFB3768: .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 $64, %rsp .cfi_def_cfa_offset 112 movq %rdi, %r14 movq %rsi, %r13 movq %rdx, %r12 movl %ecx, %ebp movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl %ecx, %ebx imull %ecx, %ebx sall $2, %ebx movslq %ebx, %rbx leaq 8(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r14, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT leal 0(,%rbp,4), %ebx movslq %ebx, %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r13, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r12, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $10, 44(%rsp) movl $1, 48(%rsp) movl $1, 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 .L15 .L12: movl $2, %ecx movq %rbx, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L16 addq $64, %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 .L15: .cfi_restore_state movl %ebp, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE3768: .size _Z13MatVecMul_GPUPfS_S_i, .-_Z13MatVecMul_GPUPfS_S_i .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "MatVecMul_GPU: " .LC2: .string "ns." .LC3: .string " " .text .globl main .type main, @function main: .LFB3769: .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 $24, %rsp .cfi_def_cfa_offset 80 movl $400, %edi call malloc@PLT movq %rax, %rbx movq %rax, 8(%rsp) movl $40, %edi call malloc@PLT movq %rax, %r12 movl $40, %edi call malloc@PLT movq %rax, %rbp movq %rbx, %rax leaq 400(%rbx), %rdx movss .LC0(%rip), %xmm0 .L18: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L18 movl $0, %eax movss .LC0(%rip), %xmm0 .L19: movss %xmm0, (%r12,%rax) movss %xmm0, 0(%rbp,%rax) addq $4, %rax cmpq $40, %rax jne .L19 call _ZNSt6chrono3_V212system_clock3nowEv@PLT movq %rax, %r13 movl $10, %ecx movq %rbp, %rdx movq %r12, %rsi movq 8(%rsp), %rdi call _Z13MatVecMul_GPUPfS_S_i call _ZNSt6chrono3_V212system_clock3nowEv@PLT movq %rax, %rbx leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %rsi subq %r13, %rsi call _ZNSo9_M_insertIlEERSoT_@PLT movq %rax, %rdi leaq .LC2(%rip), %rsi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq %rbp, %rbx leaq 40(%rbp), %r15 leaq _ZSt4cout(%rip), %r14 leaq .LC3(%rip), %r13 .L20: pxor %xmm0, %xmm0 cvtss2sd (%rbx), %xmm0 movq %r14, %rdi call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi movl $1, %edx movq %r13, %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $4, %rbx cmpq %r15, %rbx jne .L20 leaq _ZSt4cout(%rip), %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 8(%rsp), %rdi call free@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movl $0, %eax addq $24, %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 .cfi_endproc .LFE3769: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z16MatVecMul_KernelPfS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3803: .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 _Z16MatVecMul_KernelPfS_S_i(%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 .LFE3803: .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.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1065353216 .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 <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; cudaMalloc(&d_A, sizeM); cudaMemcpy(d_A, h_A, sizeM, cudaMemcpyHostToDevice); cudaMalloc(&d_B, sizeV); cudaMemcpy(d_B, h_B, sizeV, cudaMemcpyHostToDevice); cudaMalloc(&d_C, sizeV); cudaMemcpy(d_C, h_C, sizeV, cudaMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); cudaMemcpy(h_C, d_C, sizeV, cudaMemcpyDeviceToHost); cudaFree(d_A); cudaFree(d_B); cudaFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_C); return 0; }
#include <hip/hip_runtime.h> #include <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; hipMalloc(&d_A, sizeM); hipMemcpy(d_A, h_A, sizeM, hipMemcpyHostToDevice); hipMalloc(&d_B, sizeV); hipMemcpy(d_B, h_B, sizeV, hipMemcpyHostToDevice); hipMalloc(&d_C, sizeV); hipMemcpy(d_C, h_C, sizeV, hipMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); hipMemcpy(h_C, d_C, sizeV, hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_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 <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; hipMalloc(&d_A, sizeM); hipMemcpy(d_A, h_A, sizeM, hipMemcpyHostToDevice); hipMalloc(&d_B, sizeV); hipMemcpy(d_B, h_B, sizeV, hipMemcpyHostToDevice); hipMalloc(&d_C, sizeV); hipMemcpy(d_C, h_C, sizeV, hipMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); hipMemcpy(h_C, d_C, sizeV, hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_C); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z16MatVecMul_KernelPfS_S_i .globl _Z16MatVecMul_KernelPfS_S_i .p2align 8 .type _Z16MatVecMul_KernelPfS_S_i,@function _Z16MatVecMul_KernelPfS_S_i: s_load_b32 s2, s[0:1], 0x18 s_mov_b32 s3, exec_lo s_waitcnt lgkmcnt(0) v_cmpx_gt_i32_e64 s2, v0 s_cbranch_execz .LBB0_6 s_cmp_lt_i32 s2, 1 s_cbranch_scc1 .LBB0_4 s_load_b128 s[4:7], s[0:1], 0x0 v_mul_lo_u32 v1, v0, s2 v_mov_b32_e32 v2, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v3, vcc_lo, s4, v3 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s5, v4, vcc_lo .LBB0_3: global_load_b32 v1, v[3:4], off s_load_b32 s3, s[6:7], 0x0 v_add_co_u32 v3, vcc_lo, v3, 4 s_add_i32 s2, s2, -1 v_add_co_ci_u32_e32 v4, vcc_lo, 0, v4, vcc_lo s_add_u32 s6, s6, 4 s_addc_u32 s7, s7, 0 s_cmp_eq_u32 s2, 0 s_waitcnt vmcnt(0) lgkmcnt(0) v_fmac_f32_e32 v2, s3, v1 s_cbranch_scc0 .LBB0_3 s_branch .LBB0_5 .LBB0_4: v_mov_b32_e32 v2, 0 .LBB0_5: s_load_b64 s[0:1], s[0:1], 0x10 v_lshlrev_b32_e32 v0, 2, v0 s_waitcnt lgkmcnt(0) global_store_b32 v0, v2, s[0:1] .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16MatVecMul_KernelPfS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 28 .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 8 .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 _Z16MatVecMul_KernelPfS_S_i, .Lfunc_end0-_Z16MatVecMul_KernelPfS_S_i .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: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 28 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z16MatVecMul_KernelPfS_S_i .private_segment_fixed_size: 0 .sgpr_count: 10 .sgpr_spill_count: 0 .symbol: _Z16MatVecMul_KernelPfS_S_i.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 <cmath> #include <cstdio> #include <iostream> #include <chrono> using namespace std; /*MatVecMul_Kernel*/ __global__ void MatVecMul_Kernel(float* A, float* B, float* C, int n) { int i = threadIdx.x; int offset; float sum = 0; if (i < n) { for (int j = 0; j < n; j++) { offset = i*n + j; sum += A[offset] * B[j]; } C[i] = sum; } } /*MatVecMul_GPU*/ void MatVecMul_GPU(float* h_A, float* h_B, float* h_C, int n) { int sizeM = n*n * sizeof(float); int sizeV = n * sizeof(float); float *d_A; float *d_B; float *d_C; hipMalloc(&d_A, sizeM); hipMemcpy(d_A, h_A, sizeM, hipMemcpyHostToDevice); hipMalloc(&d_B, sizeV); hipMemcpy(d_B, h_B, sizeV, hipMemcpyHostToDevice); hipMalloc(&d_C, sizeV); hipMemcpy(d_C, h_C, sizeV, hipMemcpyHostToDevice); //dim3 dimGrid(ceil(n / 32.0), 1, 1); //dim3 dimBlock(32.0, 1, 1); MatVecMul_Kernel <<< 1, 10 >>> (d_A, d_B, d_C, n); hipMemcpy(h_C, d_C, sizeV, hipMemcpyDeviceToHost); hipFree(d_A); hipFree(d_B); hipFree(d_C); } int main() { //Host Matrix float *h_A, *h_B, *h_C; int n = 10; h_A = (float*)malloc(n*n * sizeof(float)); h_B = (float*)malloc(n * sizeof(float)); h_C = (float*)malloc(n * sizeof(float)); //Create Matrix for (int i = 0; i < n*n; i++) { h_A[i] = 1.0; } //Create Vector for (int i = 0; i < n; i++) { h_B[i] = 1.0; h_C[i] = 1.0; } //MatVecMul (Main) chrono::time_point<chrono::system_clock> MatVecMul_GPU_Start, MatVecMul_GPU_End; MatVecMul_GPU_Start = chrono::system_clock::now(); MatVecMul_GPU(h_A, h_B, h_C, n); MatVecMul_GPU_End = chrono::system_clock::now(); cout << "MatVecMul_GPU: " << chrono::duration_cast<chrono::nanoseconds>(MatVecMul_GPU_End - MatVecMul_GPU_Start).count() << "ns." << endl; //Print MatVecMul for (int i = 0; i < n; i++) { cout << h_C[i] << " "; } cout << endl; //Free free(h_A); free(h_B); free(h_C); return 0; }
.text .file "e2.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z31__device_stub__MatVecMul_KernelPfS_S_i # -- Begin function _Z31__device_stub__MatVecMul_KernelPfS_S_i .p2align 4, 0x90 .type _Z31__device_stub__MatVecMul_KernelPfS_S_i,@function _Z31__device_stub__MatVecMul_KernelPfS_S_i: # @_Z31__device_stub__MatVecMul_KernelPfS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z16MatVecMul_KernelPfS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z31__device_stub__MatVecMul_KernelPfS_S_i, .Lfunc_end0-_Z31__device_stub__MatVecMul_KernelPfS_S_i .cfi_endproc # -- End function .globl _Z13MatVecMul_GPUPfS_S_i # -- Begin function _Z13MatVecMul_GPUPfS_S_i .p2align 4, 0x90 .type _Z13MatVecMul_GPUPfS_S_i,@function _Z13MatVecMul_GPUPfS_S_i: # @_Z13MatVecMul_GPUPfS_S_i .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 %ecx, %r15d movq %rdx, %rbx movq %rsi, %r12 movq %rdi, %r14 leal (,%r15,4), %ebp movl %ebp, %eax imull %ecx, %eax movslq %eax, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 24(%rsp), %rdi movq %r14, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movslq %ebp, %r14 leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %r12, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 8(%rsp), %rdi movq %rbx, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy 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: movq 24(%rsp), %rax movq 16(%rsp), %rcx movq 8(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %r15d, 36(%rsp) leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 36(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z16MatVecMul_KernelPfS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: movq 8(%rsp), %rsi movq %rbx, %rdi 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 $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 .Lfunc_end1: .size _Z13MatVecMul_GPUPfS_S_i, .Lfunc_end1-_Z13MatVecMul_GPUPfS_S_i .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 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $400, %edi # imm = 0x190 callq malloc movq %rax, %rbx movl $40, %edi callq malloc movq %rax, %r14 movl $40, %edi callq malloc movq %rax, %r15 xorl %eax, %eax .p2align 4, 0x90 .LBB2_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%rbx,%rax,4) # imm = 0x3F800000 incq %rax cmpq $100, %rax jne .LBB2_1 # %bb.2: # %.preheader.preheader xorl %eax, %eax .p2align 4, 0x90 .LBB2_3: # %.preheader # =>This Inner Loop Header: Depth=1 movl $1065353216, (%r14,%rax,4) # imm = 0x3F800000 movl $1065353216, (%r15,%rax,4) # imm = 0x3F800000 incq %rax cmpq $10, %rax jne .LBB2_3 # %bb.4: callq _ZNSt6chrono3_V212system_clock3nowEv movq %rax, %r12 movq %rbx, %rdi movq %r14, %rsi movq %r15, %rdx movl $10, %ecx callq _Z13MatVecMul_GPUPfS_S_i callq _ZNSt6chrono3_V212system_clock3nowEv movq %rax, %r13 movl $_ZSt4cout, %edi movl $.L.str, %esi movl $15, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l subq %r12, %r13 movl $_ZSt4cout, %edi movq %r13, %rsi callq _ZNSo9_M_insertIlEERSoT_ movq %rax, %r12 movl $.L.str.1, %esi movl $3, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movq (%r12), %rax movq -24(%rax), %rax movq 240(%r12,%rax), %r13 testq %r13, %r13 je .LBB2_15 # %bb.5: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r13) je .LBB2_7 # %bb.6: movzbl 67(%r13), %eax jmp .LBB2_8 .LBB2_7: movq %r13, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r13), %rax movq %r13, %rdi movl $10, %esi callq *48(%rax) .LBB2_8: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movq %r12, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %r12d, %r12d .p2align 4, 0x90 .LBB2_9: # =>This Inner Loop Header: Depth=1 movss (%r15,%r12,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movl $.L.str.2, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r12 cmpq $10, %r12 jne .LBB2_9 # %bb.10: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r12 testq %r12, %r12 je .LBB2_15 # %bb.11: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i32 cmpb $0, 56(%r12) je .LBB2_13 # %bb.12: movzbl 67(%r12), %eax jmp .LBB2_14 .LBB2_13: movq %r12, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r12), %rax movq %r12, %rdi movl $10, %esi callq *48(%rax) .LBB2_14: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit35 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax 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 .LBB2_15: .cfi_def_cfa_offset 48 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 $_Z16MatVecMul_KernelPfS_S_i, %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 _Z16MatVecMul_KernelPfS_S_i,@object # @_Z16MatVecMul_KernelPfS_S_i .section .rodata,"a",@progbits .globl _Z16MatVecMul_KernelPfS_S_i .p2align 3, 0x0 _Z16MatVecMul_KernelPfS_S_i: .quad _Z31__device_stub__MatVecMul_KernelPfS_S_i .size _Z16MatVecMul_KernelPfS_S_i, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "MatVecMul_GPU: " .size .L.str, 16 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "ns." .size .L.str.1, 4 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz " " .size .L.str.2, 2 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z16MatVecMul_KernelPfS_S_i" .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 _Z31__device_stub__MatVecMul_KernelPfS_S_i .addrsig_sym __gxx_personality_v0 .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16MatVecMul_KernelPfS_S_i .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 : _Z16MatVecMul_KernelPfS_S_i .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 R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e240000002100 */ /*0020*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x001fda0003f06270 */ /*0030*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0040*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x178] ; /* 0x00005e00ff067624 */ /* 0x000fe200078e00ff */ /*0050*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0060*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fc600000001ff */ /*0070*/ ISETP.GE.AND P0, PT, R6, 0x1, PT ; /* 0x000000010600780c */ /* 0x000fda0003f06270 */ /*0080*/ @!P0 BRA 0xa50 ; /* 0x000009c000008947 */ /* 0x000fea0003800000 */ /*0090*/ IADD3 R2, R6.reuse, -0x1, RZ ; /* 0xffffffff06027810 */ /* 0x040fe20007ffe0ff */ /*00a0*/ IMAD.MOV.U32 R15, RZ, RZ, RZ ; /* 0x000000ffff0f7224 */ /* 0x000fe200078e00ff */ /*00b0*/ LOP3.LUT R6, R6, 0x3, RZ, 0xc0, !PT ; /* 0x0000000306067812 */ /* 0x000fe400078ec0ff */ /*00c0*/ ISETP.GE.U32.AND P0, PT, R2, 0x3, PT ; /* 0x000000030200780c */ /* 0x000fe40003f06070 */ /*00d0*/ MOV R7, RZ ; /* 0x000000ff00077202 */ /* 0x000fd60000000f00 */ /*00e0*/ @!P0 BRA 0x910 ; /* 0x0000082000008947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R8, -R6, c[0x0][0x178], RZ ; /* 0x00005e0006087a10 */ /* 0x000fe20007ffe1ff */ /*0100*/ IMAD.MOV.U32 R3, RZ, RZ, 0x4 ; /* 0x00000004ff037424 */ /* 0x000fe200078e00ff */ /*0110*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fe200000001ff */ /*0120*/ IMAD R2, R0, c[0x0][0x178], RZ ; /* 0x00005e0000027a24 */ /* 0x000fe200078e02ff */ /*0130*/ ISETP.GT.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f04270 */ /*0140*/ IMAD.MOV.U32 R7, RZ, RZ, RZ ; /* 0x000000ffff077224 */ /* 0x000fe200078e00ff */ /*0150*/ MOV R4, c[0x0][0x168] ; /* 0x00005a0000047a02 */ /* 0x000fe20000000f00 */ /*0160*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x0][0x16c] ; /* 0x00005b00ff057624 */ /* 0x000fe400078e00ff */ /*0170*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fd000078e0203 */ /*0180*/ @!P0 BRA 0x790 ; /* 0x0000060000008947 */ /* 0x000fea0003800000 */ /*0190*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*01a0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fd60003f0f070 */ /*01b0*/ @!P1 BRA 0x550 ; /* 0x0000039000009947 */ /* 0x000fea0003800000 */ /*01c0*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*01d0*/ LDG.E R16, [R4.64] ; /* 0x0000000404107981 */ /* 0x000ea8000c1e1900 */ /*01e0*/ LDG.E R17, [R2.64] ; /* 0x0000000402117981 */ /* 0x000ea8000c1e1900 */ /*01f0*/ LDG.E R18, [R4.64+0x4] ; /* 0x0000040404127981 */ /* 0x000ee8000c1e1900 */ /*0200*/ LDG.E R25, [R2.64+0x4] ; /* 0x0000040402197981 */ /* 0x000ee8000c1e1900 */ /*0210*/ LDG.E R19, [R4.64+0x8] ; /* 0x0000080404137981 */ /* 0x000f28000c1e1900 */ /*0220*/ LDG.E R20, [R2.64+0x8] ; /* 0x0000080402147981 */ /* 0x000f28000c1e1900 */ /*0230*/ LDG.E R22, [R4.64+0xc] ; /* 0x00000c0404167981 */ /* 0x000f68000c1e1900 */ /*0240*/ LDG.E R21, [R2.64+0xc] ; /* 0x00000c0402157981 */ /* 0x000f68000c1e1900 */ /*0250*/ LDG.E R23, [R4.64+0x10] ; /* 0x0000100404177981 */ /* 0x000f68000c1e1900 */ /*0260*/ LDG.E R24, [R2.64+0x10] ; /* 0x0000100402187981 */ /* 0x000f68000c1e1900 */ /*0270*/ LDG.E R9, [R4.64+0x14] ; /* 0x0000140404097981 */ /* 0x000f68000c1e1900 */ /*0280*/ LDG.E R10, [R2.64+0x14] ; /* 0x00001404020a7981 */ /* 0x000f68000c1e1900 */ /*0290*/ LDG.E R11, [R4.64+0x18] ; /* 0x00001804040b7981 */ /* 0x000f68000c1e1900 */ /*02a0*/ LDG.E R12, [R2.64+0x18] ; /* 0x00001804020c7981 */ /* 0x000f68000c1e1900 */ /*02b0*/ LDG.E R13, [R4.64+0x1c] ; /* 0x00001c04040d7981 */ /* 0x000f68000c1e1900 */ /*02c0*/ LDG.E R14, [R2.64+0x1c] ; /* 0x00001c04020e7981 */ /* 0x000f62000c1e1900 */ /*02d0*/ FFMA R17, R16, R17, R15 ; /* 0x0000001110117223 */ /* 0x004fc6000000000f */ /*02e0*/ LDG.E R15, [R4.64+0x20] ; /* 0x00002004040f7981 */ /* 0x0000a8000c1e1900 */ /*02f0*/ LDG.E R16, [R2.64+0x20] ; /* 0x0000200402107981 */ /* 0x0002a2000c1e1900 */ /*0300*/ FFMA R25, R18, R25, R17 ; /* 0x0000001912197223 */ /* 0x008fc60000000011 */ /*0310*/ LDG.E R17, [R4.64+0x24] ; /* 0x0000240404117981 */ /* 0x0000e8000c1e1900 */ /*0320*/ LDG.E R18, [R2.64+0x24] ; /* 0x0000240402127981 */ /* 0x0002e2000c1e1900 */ /*0330*/ FFMA R25, R19, R20, R25 ; /* 0x0000001413197223 */ /* 0x010fc60000000019 */ /*0340*/ LDG.E R19, [R4.64+0x28] ; /* 0x0000280404137981 */ /* 0x000128000c1e1900 */ /*0350*/ LDG.E R20, [R2.64+0x28] ; /* 0x0000280402147981 */ /* 0x000322000c1e1900 */ /*0360*/ FFMA R25, R22, R21, R25 ; /* 0x0000001516197223 */ /* 0x020fc60000000019 */ /*0370*/ LDG.E R21, [R4.64+0x2c] ; /* 0x00002c0404157981 */ /* 0x000168000c1e1900 */ /*0380*/ LDG.E R22, [R2.64+0x2c] ; /* 0x00002c0402167981 */ /* 0x000362000c1e1900 */ /*0390*/ FFMA R25, R23, R24, R25 ; /* 0x0000001817197223 */ /* 0x000fc60000000019 */ /*03a0*/ LDG.E R23, [R4.64+0x30] ; /* 0x0000300404177981 */ /* 0x000168000c1e1900 */ /*03b0*/ LDG.E R24, [R2.64+0x30] ; /* 0x0000300402187981 */ /* 0x000362000c1e1900 */ /*03c0*/ FFMA R25, R9, R10, R25 ; /* 0x0000000a09197223 */ /* 0x000fc60000000019 */ /*03d0*/ LDG.E R9, [R4.64+0x34] ; /* 0x0000340404097981 */ /* 0x000168000c1e1900 */ /*03e0*/ LDG.E R10, [R2.64+0x34] ; /* 0x00003404020a7981 */ /* 0x000362000c1e1900 */ /*03f0*/ FFMA R25, R11, R12, R25 ; /* 0x0000000c0b197223 */ /* 0x000fc60000000019 */ /*0400*/ LDG.E R11, [R4.64+0x38] ; /* 0x00003804040b7981 */ /* 0x000168000c1e1900 */ /*0410*/ LDG.E R12, [R2.64+0x38] ; /* 0x00003804020c7981 */ /* 0x000362000c1e1900 */ /*0420*/ FFMA R25, R13, R14, R25 ; /* 0x0000000e0d197223 */ /* 0x000fc60000000019 */ /*0430*/ LDG.E R13, [R4.64+0x3c] ; /* 0x00003c04040d7981 */ /* 0x000168000c1e1900 */ /*0440*/ LDG.E R14, [R2.64+0x3c] ; /* 0x00003c04020e7981 */ /* 0x000362000c1e1900 */ /*0450*/ IADD3 R8, R8, -0x10, RZ ; /* 0xfffffff008087810 */ /* 0x000fe40007ffe0ff */ /*0460*/ IADD3 R7, R7, 0x10, RZ ; /* 0x0000001007077810 */ /* 0x000fe40007ffe0ff */ /*0470*/ ISETP.GT.AND P1, PT, R8, 0xc, PT ; /* 0x0000000c0800780c */ /* 0x000fe40003f24270 */ /*0480*/ IADD3 R4, P3, R4, 0x40, RZ ; /* 0x0000004004047810 */ /* 0x001fc40007f7e0ff */ /*0490*/ IADD3 R2, P2, R2, 0x40, RZ ; /* 0x0000004002027810 */ /* 0x002fe40007f5e0ff */ /*04a0*/ IADD3.X R5, RZ, R5, RZ, P3, !PT ; /* 0x00000005ff057210 */ /* 0x000fc60001ffe4ff */ /*04b0*/ IMAD.X R3, RZ, RZ, R3, P2 ; /* 0x000000ffff037224 */ /* 0x000fe400010e0603 */ /*04c0*/ FFMA R15, R15, R16, R25 ; /* 0x000000100f0f7223 */ /* 0x004fc80000000019 */ /*04d0*/ FFMA R15, R17, R18, R15 ; /* 0x00000012110f7223 */ /* 0x008fc8000000000f */ /*04e0*/ FFMA R15, R19, R20, R15 ; /* 0x00000014130f7223 */ /* 0x010fc8000000000f */ /*04f0*/ FFMA R15, R21, R22, R15 ; /* 0x00000016150f7223 */ /* 0x020fc8000000000f */ /*0500*/ FFMA R15, R23, R24, R15 ; /* 0x00000018170f7223 */ /* 0x000fc8000000000f */ /*0510*/ FFMA R9, R9, R10, R15 ; /* 0x0000000a09097223 */ /* 0x000fc8000000000f */ /*0520*/ FFMA R9, R11, R12, R9 ; /* 0x0000000c0b097223 */ /* 0x000fc80000000009 */ /*0530*/ FFMA R15, R13, R14, R9 ; /* 0x0000000e0d0f7223 */ /* 0x000fe20000000009 */ /*0540*/ @P1 BRA 0x1d0 ; /* 0xfffffc8000001947 */ /* 0x000fea000383ffff */ /*0550*/ ISETP.GT.AND P1, PT, R8, 0x4, PT ; /* 0x000000040800780c */ /* 0x000fda0003f24270 */ /*0560*/ @!P1 BRA 0x770 ; /* 0x0000020000009947 */ /* 0x000fea0003800000 */ /*0570*/ LDG.E R16, [R4.64] ; /* 0x0000000404107981 */ /* 0x0000a8000c1e1900 */ /*0580*/ LDG.E R17, [R2.64] ; /* 0x0000000402117981 */ /* 0x000ea8000c1e1900 */ /*0590*/ LDG.E R19, [R4.64+0x4] ; /* 0x0000040404137981 */ /* 0x0000e8000c1e1900 */ /*05a0*/ LDG.E R18, [R2.64+0x4] ; /* 0x0000040402127981 */ /* 0x000ee8000c1e1900 */ /*05b0*/ LDG.E R21, [R4.64+0x8] ; /* 0x0000080404157981 */ /* 0x000128000c1e1900 */ /*05c0*/ LDG.E R20, [R2.64+0x8] ; /* 0x0000080402147981 */ /* 0x000f28000c1e1900 */ /*05d0*/ LDG.E R23, [R4.64+0xc] ; /* 0x00000c0404177981 */ /* 0x000168000c1e1900 */ /*05e0*/ LDG.E R22, [R2.64+0xc] ; /* 0x00000c0402167981 */ /* 0x000f68000c1e1900 */ /*05f0*/ LDG.E R25, [R4.64+0x10] ; /* 0x0000100404197981 */ /* 0x000168000c1e1900 */ /*0600*/ LDG.E R24, [R2.64+0x10] ; /* 0x0000100402187981 */ /* 0x000f68000c1e1900 */ /*0610*/ LDG.E R13, [R4.64+0x14] ; /* 0x00001404040d7981 */ /* 0x000168000c1e1900 */ /*0620*/ LDG.E R14, [R2.64+0x14] ; /* 0x00001404020e7981 */ /* 0x000f68000c1e1900 */ /*0630*/ LDG.E R11, [R4.64+0x18] ; /* 0x00001804040b7981 */ /* 0x000168000c1e1900 */ /*0640*/ LDG.E R12, [R2.64+0x18] ; /* 0x00001804020c7981 */ /* 0x000f68000c1e1900 */ /*0650*/ LDG.E R9, [R4.64+0x1c] ; /* 0x00001c0404097981 */ /* 0x000168000c1e1900 */ /*0660*/ LDG.E R10, [R2.64+0x1c] ; /* 0x00001c04020a7981 */ /* 0x000362000c1e1900 */ /*0670*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fc40003f0e170 */ /*0680*/ IADD3 R7, R7, 0x8, RZ ; /* 0x0000000807077810 */ /* 0x000fe40007ffe0ff */ /*0690*/ IADD3 R8, R8, -0x8, RZ ; /* 0xfffffff808087810 */ /* 0x000fe40007ffe0ff */ /*06a0*/ IADD3 R4, P2, R4, 0x20, RZ ; /* 0x0000002004047810 */ /* 0x001fca0007f5e0ff */ /*06b0*/ IMAD.X R5, RZ, RZ, R5, P2 ; /* 0x000000ffff057224 */ /* 0x000fe400010e0605 */ /*06c0*/ FFMA R16, R16, R17, R15 ; /* 0x0000001110107223 */ /* 0x004fc8000000000f */ /*06d0*/ FFMA R16, R19, R18, R16 ; /* 0x0000001213107223 */ /* 0x008fc80000000010 */ /*06e0*/ FFMA R16, R21, R20, R16 ; /* 0x0000001415107223 */ /* 0x010fc80000000010 */ /*06f0*/ FFMA R16, R23, R22, R16 ; /* 0x0000001617107223 */ /* 0x020fc80000000010 */ /*0700*/ FFMA R16, R25, R24, R16 ; /* 0x0000001819107223 */ /* 0x000fc80000000010 */ /*0710*/ FFMA R13, R13, R14, R16 ; /* 0x0000000e0d0d7223 */ /* 0x000fc80000000010 */ /*0720*/ FFMA R11, R11, R12, R13 ; /* 0x0000000c0b0b7223 */ /* 0x000fe2000000000d */ /*0730*/ IADD3 R12, P1, R2, 0x20, RZ ; /* 0x00000020020c7810 */ /* 0x000fc80007f3e0ff */ /*0740*/ IADD3.X R3, RZ, R3, RZ, P1, !PT ; /* 0x00000003ff037210 */ /* 0x002fe40000ffe4ff */ /*0750*/ MOV R2, R12 ; /* 0x0000000c00027202 */ /* 0x000fe20000000f00 */ /*0760*/ FFMA R15, R9, R10, R11 ; /* 0x0000000a090f7223 */ /* 0x000fe4000000000b */ /*0770*/ ISETP.NE.OR P0, PT, R8, RZ, P0 ; /* 0x000000ff0800720c */ /* 0x000fda0000705670 */ /*0780*/ @!P0 BRA 0x910 ; /* 0x0000018000008947 */ /* 0x000fea0003800000 */ /*0790*/ LDG.E R10, [R4.64] ; /* 0x00000004040a7981 */ /* 0x000ea8000c1e1900 */ /*07a0*/ LDG.E R9, [R2.64] ; /* 0x0000000402097981 */ /* 0x000ea8000c1e1900 */ /*07b0*/ LDG.E R12, [R4.64+0x4] ; /* 0x00000404040c7981 */ /* 0x0000e8000c1e1900 */ /*07c0*/ LDG.E R11, [R2.64+0x4] ; /* 0x00000404020b7981 */ /* 0x000ee8000c1e1900 */ /*07d0*/ LDG.E R14, [R4.64+0x8] ; /* 0x00000804040e7981 */ /* 0x000128000c1e1900 */ /*07e0*/ LDG.E R13, [R2.64+0x8] ; /* 0x00000804020d7981 */ /* 0x000328000c1e1900 */ /*07f0*/ LDG.E R16, [R4.64+0xc] ; /* 0x00000c0404107981 */ /* 0x000168000c1e1900 */ /*0800*/ LDG.E R17, [R2.64+0xc] ; /* 0x00000c0402117981 */ /* 0x000362000c1e1900 */ /*0810*/ IADD3 R8, R8, -0x4, RZ ; /* 0xfffffffc08087810 */ /* 0x000fc40007ffe0ff */ /*0820*/ IADD3 R7, R7, 0x4, RZ ; /* 0x0000000407077810 */ /* 0x000fe40007ffe0ff */ /*0830*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fe20003f05270 */ /*0840*/ FFMA R9, R10, R9, R15 ; /* 0x000000090a097223 */ /* 0x004fe2000000000f */ /*0850*/ IADD3 R10, P2, R4, 0x10, RZ ; /* 0x00000010040a7810 */ /* 0x000fca0007f5e0ff */ /*0860*/ IMAD.MOV.U32 R4, RZ, RZ, R10 ; /* 0x000000ffff047224 */ /* 0x001fe400078e000a */ /*0870*/ FFMA R9, R12, R11, R9 ; /* 0x0000000b0c097223 */ /* 0x008fe20000000009 */ /*0880*/ IADD3 R12, P1, R2, 0x10, RZ ; /* 0x00000010020c7810 */ /* 0x000fe20007f3e0ff */ /*0890*/ IMAD.X R11, RZ, RZ, R5, P2 ; /* 0x000000ffff0b7224 */ /* 0x000fc800010e0605 */ /*08a0*/ IMAD.MOV.U32 R2, RZ, RZ, R12 ; /* 0x000000ffff027224 */ /* 0x002fe200078e000c */ /*08b0*/ MOV R5, R11 ; /* 0x0000000b00057202 */ /* 0x000fe20000000f00 */ /*08c0*/ FFMA R9, R14, R13, R9 ; /* 0x0000000d0e097223 */ /* 0x010fe20000000009 */ /*08d0*/ IADD3.X R13, RZ, R3, RZ, P1, !PT ; /* 0x00000003ff0d7210 */ /* 0x000fc80000ffe4ff */ /*08e0*/ MOV R3, R13 ; /* 0x0000000d00037202 */ /* 0x000fe20000000f00 */ /*08f0*/ FFMA R15, R16, R17, R9 ; /* 0x00000011100f7223 */ /* 0x020fe20000000009 */ /*0900*/ @P0 BRA 0x790 ; /* 0xfffffe8000000947 */ /* 0x000fea000383ffff */ /*0910*/ ISETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600720c */ /* 0x000fda0003f05270 */ /*0920*/ @!P0 BRA 0xa50 ; /* 0x0000012000008947 */ /* 0x000fea0003800000 */ /*0930*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe400078e00ff */ /*0940*/ IMAD R4, R0, c[0x0][0x178], R7 ; /* 0x00005e0000047a24 */ /* 0x000fe400078e0207 */ /*0950*/ IMAD.WIDE R2, R7, R5, c[0x0][0x168] ; /* 0x00005a0007027625 */ /* 0x000fc800078e0205 */ /*0960*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0970*/ MOV R8, R2 ; /* 0x0000000200087202 */ /* 0x000fc60000000f00 */ /*0980*/ IMAD.MOV.U32 R7, RZ, RZ, R4 ; /* 0x000000ffff077224 */ /* 0x000fc600078e0004 */ /*0990*/ MOV R2, R8 ; /* 0x0000000800027202 */ /* 0x000fe40000000f00 */ /*09a0*/ MOV R4, R7 ; /* 0x0000000700047202 */ /* 0x000fc80000000f00 */ /*09b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x0000a8000c1e1900 */ /*09c0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x0002a2000c1e1900 */ /*09d0*/ IADD3 R6, R6, -0x1, RZ ; /* 0xffffffff06067810 */ /* 0x000fe40007ffe0ff */ /*09e0*/ IADD3 R8, P1, R8, 0x4, RZ ; /* 0x0000000408087810 */ /* 0x000fe40007f3e0ff */ /*09f0*/ ISETP.NE.AND P0, PT, R6, RZ, PT ; /* 0x000000ff0600720c */ /* 0x000fe40003f05270 */ /*0a00*/ IADD3 R7, P2, R7, 0x4, RZ ; /* 0x0000000407077810 */ /* 0x000fe20007f5e0ff */ /*0a10*/ IMAD.X R3, RZ, RZ, R3, P1 ; /* 0x000000ffff037224 */ /* 0x001fc600008e0603 */ /*0a20*/ IADD3.X R5, RZ, R5, RZ, P2, !PT ; /* 0x00000005ff057210 */ /* 0x002fe200017fe4ff */ /*0a30*/ FFMA R15, R2, R4, R15 ; /* 0x00000004020f7223 */ /* 0x004fcc000000000f */ /*0a40*/ @P0 BRA 0x990 ; /* 0xffffff4000000947 */ /* 0x000fea000383ffff */ /*0a50*/ MOV R3, 0x4 ; /* 0x0000000400037802 */ /* 0x000fca0000000f00 */ /*0a60*/ IMAD.WIDE R2, R0, R3, c[0x0][0x170] ; /* 0x00005c0000027625 */ /* 0x000fca00078e0203 */ /*0a70*/ STG.E [R2.64], R15 ; /* 0x0000000f02007986 */ /* 0x000fe2000c101904 */ /*0a80*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0a90*/ BRA 0xa90; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b00*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b10*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b20*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0b70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z16MatVecMul_KernelPfS_S_i .globl _Z16MatVecMul_KernelPfS_S_i .p2align 8 .type _Z16MatVecMul_KernelPfS_S_i,@function _Z16MatVecMul_KernelPfS_S_i: s_load_b32 s2, s[0:1], 0x18 s_mov_b32 s3, exec_lo s_waitcnt lgkmcnt(0) v_cmpx_gt_i32_e64 s2, v0 s_cbranch_execz .LBB0_6 s_cmp_lt_i32 s2, 1 s_cbranch_scc1 .LBB0_4 s_load_b128 s[4:7], s[0:1], 0x0 v_mul_lo_u32 v1, v0, s2 v_mov_b32_e32 v2, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[3:4], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v3, vcc_lo, s4, v3 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v4, vcc_lo, s5, v4, vcc_lo .LBB0_3: global_load_b32 v1, v[3:4], off s_load_b32 s3, s[6:7], 0x0 v_add_co_u32 v3, vcc_lo, v3, 4 s_add_i32 s2, s2, -1 v_add_co_ci_u32_e32 v4, vcc_lo, 0, v4, vcc_lo s_add_u32 s6, s6, 4 s_addc_u32 s7, s7, 0 s_cmp_eq_u32 s2, 0 s_waitcnt vmcnt(0) lgkmcnt(0) v_fmac_f32_e32 v2, s3, v1 s_cbranch_scc0 .LBB0_3 s_branch .LBB0_5 .LBB0_4: v_mov_b32_e32 v2, 0 .LBB0_5: s_load_b64 s[0:1], s[0:1], 0x10 v_lshlrev_b32_e32 v0, 2, v0 s_waitcnt lgkmcnt(0) global_store_b32 v0, v2, s[0:1] .LBB0_6: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16MatVecMul_KernelPfS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 28 .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 8 .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 _Z16MatVecMul_KernelPfS_S_i, .Lfunc_end0-_Z16MatVecMul_KernelPfS_S_i .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: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 28 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z16MatVecMul_KernelPfS_S_i .private_segment_fixed_size: 0 .sgpr_count: 10 .sgpr_spill_count: 0 .symbol: _Z16MatVecMul_KernelPfS_S_i.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 from CUDA host assembly to AMD host assembly.
.file "tmpxft_0018df27_00000000-6_e2.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3778: .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 .LFE3778: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i .type _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i, @function _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i: .LFB3800: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 _Z16MatVecMul_KernelPfS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3800: .size _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i, .-_Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i .globl _Z16MatVecMul_KernelPfS_S_i .type _Z16MatVecMul_KernelPfS_S_i, @function _Z16MatVecMul_KernelPfS_S_i: .LFB3801: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3801: .size _Z16MatVecMul_KernelPfS_S_i, .-_Z16MatVecMul_KernelPfS_S_i .globl _Z13MatVecMul_GPUPfS_S_i .type _Z13MatVecMul_GPUPfS_S_i, @function _Z13MatVecMul_GPUPfS_S_i: .LFB3768: .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 $64, %rsp .cfi_def_cfa_offset 112 movq %rdi, %r14 movq %rsi, %r13 movq %rdx, %r12 movl %ecx, %ebp movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl %ecx, %ebx imull %ecx, %ebx sall $2, %ebx movslq %ebx, %rbx leaq 8(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r14, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT leal 0(,%rbp,4), %ebx movslq %ebx, %rbx leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r13, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT leaq 24(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r12, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $10, 44(%rsp) movl $1, 48(%rsp) movl $1, 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 .L15 .L12: movl $2, %ecx movq %rbx, %rdx movq 24(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L16 addq $64, %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 .L15: .cfi_restore_state movl %ebp, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z41__device_stub__Z16MatVecMul_KernelPfS_S_iPfS_S_i jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE3768: .size _Z13MatVecMul_GPUPfS_S_i, .-_Z13MatVecMul_GPUPfS_S_i .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "MatVecMul_GPU: " .LC2: .string "ns." .LC3: .string " " .text .globl main .type main, @function main: .LFB3769: .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 $24, %rsp .cfi_def_cfa_offset 80 movl $400, %edi call malloc@PLT movq %rax, %rbx movq %rax, 8(%rsp) movl $40, %edi call malloc@PLT movq %rax, %r12 movl $40, %edi call malloc@PLT movq %rax, %rbp movq %rbx, %rax leaq 400(%rbx), %rdx movss .LC0(%rip), %xmm0 .L18: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L18 movl $0, %eax movss .LC0(%rip), %xmm0 .L19: movss %xmm0, (%r12,%rax) movss %xmm0, 0(%rbp,%rax) addq $4, %rax cmpq $40, %rax jne .L19 call _ZNSt6chrono3_V212system_clock3nowEv@PLT movq %rax, %r13 movl $10, %ecx movq %rbp, %rdx movq %r12, %rsi movq 8(%rsp), %rdi call _Z13MatVecMul_GPUPfS_S_i call _ZNSt6chrono3_V212system_clock3nowEv@PLT movq %rax, %rbx leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %rsi subq %r13, %rsi call _ZNSo9_M_insertIlEERSoT_@PLT movq %rax, %rdi leaq .LC2(%rip), %rsi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq %rbp, %rbx leaq 40(%rbp), %r15 leaq _ZSt4cout(%rip), %r14 leaq .LC3(%rip), %r13 .L20: pxor %xmm0, %xmm0 cvtss2sd (%rbx), %xmm0 movq %r14, %rdi call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi movl $1, %edx movq %r13, %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $4, %rbx cmpq %r15, %rbx jne .L20 leaq _ZSt4cout(%rip), %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 8(%rsp), %rdi call free@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movl $0, %eax addq $24, %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 .cfi_endproc .LFE3769: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z16MatVecMul_KernelPfS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3803: .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 _Z16MatVecMul_KernelPfS_S_i(%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 .LFE3803: .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.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1065353216 .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 "e2.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z31__device_stub__MatVecMul_KernelPfS_S_i # -- Begin function _Z31__device_stub__MatVecMul_KernelPfS_S_i .p2align 4, 0x90 .type _Z31__device_stub__MatVecMul_KernelPfS_S_i,@function _Z31__device_stub__MatVecMul_KernelPfS_S_i: # @_Z31__device_stub__MatVecMul_KernelPfS_S_i .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z16MatVecMul_KernelPfS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z31__device_stub__MatVecMul_KernelPfS_S_i, .Lfunc_end0-_Z31__device_stub__MatVecMul_KernelPfS_S_i .cfi_endproc # -- End function .globl _Z13MatVecMul_GPUPfS_S_i # -- Begin function _Z13MatVecMul_GPUPfS_S_i .p2align 4, 0x90 .type _Z13MatVecMul_GPUPfS_S_i,@function _Z13MatVecMul_GPUPfS_S_i: # @_Z13MatVecMul_GPUPfS_S_i .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 %ecx, %r15d movq %rdx, %rbx movq %rsi, %r12 movq %rdi, %r14 leal (,%r15,4), %ebp movl %ebp, %eax imull %ecx, %eax movslq %eax, %r13 leaq 24(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 24(%rsp), %rdi movq %r14, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movslq %ebp, %r14 leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %r12, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 8(%rsp), %rdi movq %rbx, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy 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: movq 24(%rsp), %rax movq 16(%rsp), %rcx movq 8(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl %r15d, 36(%rsp) leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 36(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z16MatVecMul_KernelPfS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: movq 8(%rsp), %rsi movq %rbx, %rdi 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 $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 .Lfunc_end1: .size _Z13MatVecMul_GPUPfS_S_i, .Lfunc_end1-_Z13MatVecMul_GPUPfS_S_i .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 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $400, %edi # imm = 0x190 callq malloc movq %rax, %rbx movl $40, %edi callq malloc movq %rax, %r14 movl $40, %edi callq malloc movq %rax, %r15 xorl %eax, %eax .p2align 4, 0x90 .LBB2_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%rbx,%rax,4) # imm = 0x3F800000 incq %rax cmpq $100, %rax jne .LBB2_1 # %bb.2: # %.preheader.preheader xorl %eax, %eax .p2align 4, 0x90 .LBB2_3: # %.preheader # =>This Inner Loop Header: Depth=1 movl $1065353216, (%r14,%rax,4) # imm = 0x3F800000 movl $1065353216, (%r15,%rax,4) # imm = 0x3F800000 incq %rax cmpq $10, %rax jne .LBB2_3 # %bb.4: callq _ZNSt6chrono3_V212system_clock3nowEv movq %rax, %r12 movq %rbx, %rdi movq %r14, %rsi movq %r15, %rdx movl $10, %ecx callq _Z13MatVecMul_GPUPfS_S_i callq _ZNSt6chrono3_V212system_clock3nowEv movq %rax, %r13 movl $_ZSt4cout, %edi movl $.L.str, %esi movl $15, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l subq %r12, %r13 movl $_ZSt4cout, %edi movq %r13, %rsi callq _ZNSo9_M_insertIlEERSoT_ movq %rax, %r12 movl $.L.str.1, %esi movl $3, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movq (%r12), %rax movq -24(%rax), %rax movq 240(%r12,%rax), %r13 testq %r13, %r13 je .LBB2_15 # %bb.5: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r13) je .LBB2_7 # %bb.6: movzbl 67(%r13), %eax jmp .LBB2_8 .LBB2_7: movq %r13, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r13), %rax movq %r13, %rdi movl $10, %esi callq *48(%rax) .LBB2_8: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %al, %esi movq %r12, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv xorl %r12d, %r12d .p2align 4, 0x90 .LBB2_9: # =>This Inner Loop Header: Depth=1 movss (%r15,%r12,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movl $.L.str.2, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r12 cmpq $10, %r12 jne .LBB2_9 # %bb.10: movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r12 testq %r12, %r12 je .LBB2_15 # %bb.11: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i32 cmpb $0, 56(%r12) je .LBB2_13 # %bb.12: movzbl 67(%r12), %eax jmp .LBB2_14 .LBB2_13: movq %r12, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r12), %rax movq %r12, %rdi movl $10, %esi callq *48(%rax) .LBB2_14: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit35 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax 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 .LBB2_15: .cfi_def_cfa_offset 48 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 $_Z16MatVecMul_KernelPfS_S_i, %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 _Z16MatVecMul_KernelPfS_S_i,@object # @_Z16MatVecMul_KernelPfS_S_i .section .rodata,"a",@progbits .globl _Z16MatVecMul_KernelPfS_S_i .p2align 3, 0x0 _Z16MatVecMul_KernelPfS_S_i: .quad _Z31__device_stub__MatVecMul_KernelPfS_S_i .size _Z16MatVecMul_KernelPfS_S_i, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "MatVecMul_GPU: " .size .L.str, 16 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "ns." .size .L.str.1, 4 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz " " .size .L.str.2, 2 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z16MatVecMul_KernelPfS_S_i" .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 _Z31__device_stub__MatVecMul_KernelPfS_S_i .addrsig_sym __gxx_personality_v0 .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16MatVecMul_KernelPfS_S_i .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 host assembly.
#include <stdio.h> #include <algorithm> #include <cstdlib> #include <curand.h> #include <curand_kernel.h> // In the following section, define the prob distribution parameters #define N_PARAMS 3 #define PARAM1 50.0f, 3.0f, 0.5f // format: LAMBDA, A, B #define PARAM2 1.5f, 0.8f, 5.0f // parameters saved as constants unsigned int N_BYTES_PRM = N_PARAMS * sizeof(float); // size of parameter unsigned int N_SIMS, N_BLK, N_THRD, N_BYTES_I, N_BYTES_F; const unsigned int MAX_THREADS = 512; // max threads per block // Calculate and return mean of an array of floats float calcMean(float arr[], unsigned int const n) { double sum = 0.0; for (unsigned int i=0; i<n; i++) { sum += (arr[i] / n); } return sum; } __host__ cudaEvent_t get_time(void) { cudaEvent_t time; cudaEventCreate(&time); cudaEventRecord(time); return time; } // Based on parameter, draw poisson random number as frequency of losses __global__ void sim_freq(unsigned int *f_out, float *prm, unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { float lambda = prm[0]; // lambda for poisson curandState_t state; // initialize rand state curand_init(tid, 0, 0, &state); // set seed to thread index f_out[tid] = curand_poisson(&state, lambda); // save loss frequency } } // Based on parameter and freq, draw and sum pareto loss amounts __global__ void sim_severity(float *loss_out, unsigned int *freq, float *prm, const unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { double A = prm[1]; double B = prm[2]; // two pareto parameters curandState_t state; // initialize rand state curand_init(tid, 0, 0, &state); // set seed to thread index double sum = 0.0; double unif = 0.0; // temp var for storing uniform rand for (int f=0; f < freq[tid]; f++) { unif = curand_uniform_double(&state); // draw unif rand as double sum += B / pow(1-unif, 1/A); // quantile function (inverse CDF) } loss_out[tid] = (float) sum; // sum of all losses } } void asynch() { return; } int main(int argc, char* argv[]) { if (argc == 2) { // get number of simulations based on CMDLINE input N_SIMS = atoi(argv[1]); } else { printf("Usage: %s [nSimulations].\n", argv[0]); return EXIT_FAILURE; } N_BLK = N_SIMS / MAX_THREADS + 1; // min of one block N_THRD = std::min(N_SIMS, MAX_THREADS); // num of threads per block N_BYTES_F = N_SIMS * sizeof(float); // size of loss array N_BYTES_I = N_SIMS * sizeof(unsigned int); // size of frequency array printf("Running %u simulations ...\n", N_SIMS); cudaStream_t s1, s2; // Create and initialize streams cudaStreamCreate(&s1); cudaStreamCreate(&s2); // allocate and copy parameter to device float h_prm1 [N_PARAMS] = {PARAM1}; float h_prm2 [N_PARAMS] = {PARAM2}; cudaHostRegister(h_prm1, N_BYTES_PRM, cudaHostRegisterDefault); cudaHostRegister(h_prm2, N_BYTES_PRM, cudaHostRegisterDefault); float *d_prm1, *d_prm2; // parameter on device memory cudaMalloc((void **)&d_prm1, N_BYTES_PRM); // allocate and copy cudaMalloc((void **)&d_prm2, N_BYTES_PRM); cudaMemcpyAsync(d_prm1, h_prm1, N_BYTES_PRM, cudaMemcpyHostToDevice, s1); cudaMemcpyAsync(d_prm2, h_prm2, N_BYTES_PRM, cudaMemcpyHostToDevice, s2); unsigned int *h_freq1, *d_freq1, *h_freq2, *d_freq2; // frequency arrays float *h_loss1, *d_loss1, *h_loss2, *d_loss2; // loss arrays cudaMalloc((void **)&d_freq1, N_BYTES_I); // device mem for freq and loss cudaMalloc((void **)&d_loss1, N_BYTES_F); cudaMalloc((void **)&d_freq2, N_BYTES_I); cudaMalloc((void **)&d_loss2, N_BYTES_F); cudaMallocHost((void**)&h_freq1, N_BYTES_I); // pinned host array cudaMallocHost((void**)&h_loss1, N_BYTES_F); cudaMallocHost((void**)&h_freq2, N_BYTES_I); cudaMallocHost((void**)&h_loss2, N_BYTES_F); float dur, mean1, mean2; // to record duration and averages /****** asynchronus run ****** ******************************/ cudaEvent_t start = get_time(); // start clock cudaEvent_t copyEnd1, copyEnd2; // event to ensure copying loss is finished cudaEventCreate(&copyEnd1); cudaEventCreate(&copyEnd2); // first simulate frequency of losses in two streams sim_freq<<<N_BLK, N_THRD, 0, s1>>>(d_freq1, d_prm1, N_SIMS); sim_freq<<<N_BLK, N_THRD, 0, s2>>>(d_freq2, d_prm2, N_SIMS); // based on frequency, draw pareto loss amounts and sum sim_severity<<<N_BLK, N_THRD, 0, s1>>>(d_loss1, d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD, 0, s2>>>(d_loss2, d_freq2, d_prm2, N_SIMS); cudaMemcpyAsync(h_loss1, d_loss1, N_BYTES_F, cudaMemcpyDeviceToHost, s1); cudaEventRecord(copyEnd1, s1); // finish copying result in stream 1 cudaMemcpyAsync(h_loss2, d_loss2, N_BYTES_F, cudaMemcpyDeviceToHost, s2); cudaEventRecord(copyEnd2, s2); // finish copying result in stream 2 cudaMemcpyAsync(h_freq1, d_freq1, N_BYTES_I, cudaMemcpyDeviceToHost, s1); cudaMemcpyAsync(h_freq2, d_freq2, N_BYTES_I, cudaMemcpyDeviceToHost, s2); cudaEventSynchronize(copyEnd1); // wait for result copy before calculation mean1 = calcMean(h_loss1, N_SIMS); cudaEventSynchronize(copyEnd2); // wait for result copy before calculation mean2 = calcMean(h_loss2, N_SIMS); cudaStreamSynchronize( s1 ); cudaStreamSynchronize( s2 ); cudaEvent_t stop = get_time(); // stop clock cudaEventSynchronize(stop); cudaEventElapsedTime(&dur, start, stop); printf("\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); /****** synchronus run ******* ******************************/ start = get_time(); sim_freq<<<N_BLK, N_THRD>>>(d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss1, d_freq1, d_prm1, N_SIMS); cudaMemcpy(h_freq1, d_freq1, N_BYTES_I, cudaMemcpyDeviceToHost); cudaMemcpy(h_loss1, d_loss1, N_BYTES_F, cudaMemcpyDeviceToHost); sim_freq<<<N_BLK, N_THRD>>>(d_freq2, d_prm2, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss2, d_freq2, d_prm2, N_SIMS); cudaMemcpy(h_freq2, d_freq2, N_BYTES_I, cudaMemcpyDeviceToHost); cudaMemcpy(h_loss2, d_loss2, N_BYTES_F, cudaMemcpyDeviceToHost); mean1 = calcMean(h_loss1, N_SIMS); mean2 = calcMean(h_loss2, N_SIMS); stop = get_time(); // stop time cudaEventSynchronize(stop); cudaEventElapsedTime(&dur, start, stop); printf("\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); return EXIT_SUCCESS; }
.file "tmpxft_0016f86c_00000000-6_stream_registers.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2558: .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 .LFE2558: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z8calcMeanPfj .type _Z8calcMeanPfj, @function _Z8calcMeanPfj: .LFB2551: .cfi_startproc endbr64 testl %esi, %esi je .L8 movq %rdi, %rax movl %esi, %edx leaq (%rdi,%rdx,4), %rdx pxor %xmm0, %xmm0 movl %esi, %esi pxor %xmm2, %xmm2 cvtsi2ssq %rsi, %xmm2 .L10: movss (%rax), %xmm1 divss %xmm2, %xmm1 cvtss2sd %xmm1, %xmm1 addsd %xmm1, %xmm0 addq $4, %rax cmpq %rdx, %rax jne .L10 .L4: cvtsd2ss %xmm0, %xmm0 ret .L8: pxor %xmm0, %xmm0 jmp .L4 .cfi_endproc .LFE2551: .size _Z8calcMeanPfj, .-_Z8calcMeanPfj .globl _Z8get_timev .type _Z8get_timev, @function _Z8get_timev: .LFB2552: .cfi_startproc endbr64 subq $24, %rsp .cfi_def_cfa_offset 32 movq %fs:40, %rax movq %rax, 8(%rsp) xorl %eax, %eax movq %rsp, %rdi call cudaEventCreate@PLT movl $0, %esi movq (%rsp), %rdi call cudaEventRecord@PLT movq (%rsp), %rax movq 8(%rsp), %rdx subq %fs:40, %rdx jne .L14 addq $24, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2552: .size _Z8get_timev, .-_Z8get_timev .globl _Z6asynchv .type _Z6asynchv, @function _Z6asynchv: .LFB2553: .cfi_startproc endbr64 ret .cfi_endproc .LFE2553: .size _Z6asynchv, .-_Z6asynchv .globl _Z30__device_stub__Z8sim_freqPjPfjPjPfj .type _Z30__device_stub__Z8sim_freqPjPfjPjPfj, @function _Z30__device_stub__Z8sim_freqPjPfjPjPfj: .LFB2580: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 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 .L20 .L16: movq 120(%rsp), %rax subq %fs:40, %rax jne .L21 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L20: .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 _Z8sim_freqPjPfj(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L16 .L21: call __stack_chk_fail@PLT .cfi_endproc .LFE2580: .size _Z30__device_stub__Z8sim_freqPjPfjPjPfj, .-_Z30__device_stub__Z8sim_freqPjPfjPjPfj .globl _Z8sim_freqPjPfj .type _Z8sim_freqPjPfj, @function _Z8sim_freqPjPfj: .LFB2581: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z8sim_freqPjPfjPjPfj addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2581: .size _Z8sim_freqPjPfj, .-_Z8sim_freqPjPfj .globl _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j .type _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j, @function _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j: .LFB2582: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L28 .L24: movq 136(%rsp), %rax subq %fs:40, %rax jne .L29 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L28: .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 _Z12sim_severityPfPjS_j(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L24 .L29: call __stack_chk_fail@PLT .cfi_endproc .LFE2582: .size _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j, .-_Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j .globl _Z12sim_severityPfPjS_j .type _Z12sim_severityPfPjS_j, @function _Z12sim_severityPfPjS_j: .LFB2583: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2583: .size _Z12sim_severityPfPjS_j, .-_Z12sim_severityPfPjS_j .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Running %u simulations ...\n" .LC8: .string "Usage: %s [nSimulations].\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC9: .string "\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .align 8 .LC10: .string "\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .text .globl main .type main, @function main: .LFB2554: .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 $200, %rsp .cfi_def_cfa_offset 224 movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax cmpl $2, %edi jne .L33 movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movl %eax, %ecx movl %eax, N_SIMS(%rip) movl %eax, %edx shrl $9, %edx addl $1, %edx movl %edx, N_BLK(%rip) movl $512, %edx cmpl %edx, %eax cmovbe %eax, %edx movl %edx, N_THRD(%rip) sall $2, %eax movl %eax, N_BYTES_F(%rip) movl %eax, N_BYTES_I(%rip) movl %ecx, %edx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 24(%rsp), %rdi call cudaStreamCreate@PLT leaq 32(%rsp), %rdi call cudaStreamCreate@PLT movl $0x42480000, 160(%rsp) movl $0x40400000, 164(%rsp) movl $0x3f000000, 168(%rsp) movl $0x3fc00000, 172(%rsp) movl $0x3f4ccccd, 176(%rsp) movl $0x40a00000, 180(%rsp) movl N_BYTES_PRM(%rip), %esi leaq 160(%rsp), %rbp movl $0, %edx movq %rbp, %rdi call cudaHostRegister@PLT movl N_BYTES_PRM(%rip), %esi leaq 172(%rsp), %rbx movl $0, %edx movq %rbx, %rdi call cudaHostRegister@PLT movl N_BYTES_PRM(%rip), %esi leaq 40(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_PRM(%rip), %esi leaq 48(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_PRM(%rip), %edx movq 24(%rsp), %r8 movl $1, %ecx movq %rbp, %rsi movq 40(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_PRM(%rip), %edx movq 32(%rsp), %r8 movl $1, %ecx movq %rbx, %rsi movq 48(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_I(%rip), %esi leaq 64(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_F(%rip), %esi leaq 96(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_I(%rip), %esi leaq 80(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_F(%rip), %esi leaq 112(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_I(%rip), %esi leaq 56(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_F(%rip), %esi leaq 88(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_I(%rip), %esi leaq 72(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_F(%rip), %esi leaq 104(%rsp), %rdi call cudaMallocHost@PLT call _Z8get_timev movq %rax, %rbp leaq 120(%rsp), %rdi call cudaEventCreate@PLT leaq 128(%rsp), %rdi call cudaEventCreate@PLT movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 24(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L47 .L34: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 32(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L48 .L37: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 24(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L49 .L38: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 32(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L50 .L39: movl N_BYTES_F(%rip), %edx movq 24(%rsp), %r8 movl $2, %ecx movq 96(%rsp), %rsi movq 88(%rsp), %rdi call cudaMemcpyAsync@PLT movq 24(%rsp), %rsi movq 120(%rsp), %rdi call cudaEventRecord@PLT movl N_BYTES_F(%rip), %edx movq 32(%rsp), %r8 movl $2, %ecx movq 112(%rsp), %rsi movq 104(%rsp), %rdi call cudaMemcpyAsync@PLT movq 32(%rsp), %rsi movq 128(%rsp), %rdi call cudaEventRecord@PLT movl N_BYTES_I(%rip), %edx movq 24(%rsp), %r8 movl $2, %ecx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_I(%rip), %edx movq 32(%rsp), %r8 movl $2, %ecx movq 80(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpyAsync@PLT movq 120(%rsp), %rdi call cudaEventSynchronize@PLT movl N_SIMS(%rip), %esi movq 88(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 8(%rsp) movq 128(%rsp), %rdi call cudaEventSynchronize@PLT movl N_SIMS(%rip), %esi movq 104(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 12(%rsp) movq 24(%rsp), %rdi call cudaStreamSynchronize@PLT movq 32(%rsp), %rdi call cudaStreamSynchronize@PLT call _Z8get_timev movq %rax, %rbx movq %rax, %rdi call cudaEventSynchronize@PLT leaq 20(%rsp), %rdi movq %rbx, %rdx movq %rbp, %rsi call cudaEventElapsedTime@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 pxor %xmm2, %xmm2 cvtss2sd 20(%rsp), %xmm2 pxor %xmm1, %xmm1 cvtss2sd 12(%rsp), %xmm1 leaq .LC9(%rip), %rsi movl $2, %edi movl $3, %eax call __printf_chk@PLT call _Z8get_timev movq %rax, %rbp movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L51 .L40: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L52 .L41: movl N_BYTES_I(%rip), %edx movl $2, %ecx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpy@PLT movl N_BYTES_F(%rip), %edx movl $2, %ecx movq 96(%rsp), %rsi movq 88(%rsp), %rdi call cudaMemcpy@PLT movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L53 .L42: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L54 .L43: movl N_BYTES_I(%rip), %edx movl $2, %ecx movq 80(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpy@PLT movl N_BYTES_F(%rip), %edx movl $2, %ecx movq 112(%rsp), %rsi movq 104(%rsp), %rdi call cudaMemcpy@PLT movl N_SIMS(%rip), %ebx movl %ebx, %esi movq 88(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 8(%rsp) movl %ebx, %esi movq 104(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 12(%rsp) call _Z8get_timev movq %rax, %rbx movq %rax, %rdi call cudaEventSynchronize@PLT leaq 20(%rsp), %rdi movq %rbx, %rdx movq %rbp, %rsi call cudaEventElapsedTime@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 pxor %xmm2, %xmm2 cvtss2sd 20(%rsp), %xmm2 pxor %xmm1, %xmm1 cvtss2sd 12(%rsp), %xmm1 leaq .LC10(%rip), %rsi movl $2, %edi movl $3, %eax call __printf_chk@PLT movl $0, %eax jmp .L32 .L33: movq (%rsi), %rdx leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %eax .L32: movq 184(%rsp), %rdx subq %fs:40, %rdx jne .L55 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L47: .cfi_restore_state movl N_SIMS(%rip), %edx movq 40(%rsp), %rsi movq 64(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L34 .L48: movl N_SIMS(%rip), %edx movq 48(%rsp), %rsi movq 80(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L37 .L49: movl N_SIMS(%rip), %ecx movq 40(%rsp), %rdx movq 64(%rsp), %rsi movq 96(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L38 .L50: movl N_SIMS(%rip), %ecx movq 48(%rsp), %rdx movq 80(%rsp), %rsi movq 112(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L39 .L51: movl N_SIMS(%rip), %edx movq 40(%rsp), %rsi movq 64(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L40 .L52: movl N_SIMS(%rip), %ecx movq 40(%rsp), %rdx movq 64(%rsp), %rsi movq 96(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L41 .L53: movl N_SIMS(%rip), %edx movq 48(%rsp), %rsi movq 80(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L42 .L54: movl N_SIMS(%rip), %ecx movq 48(%rsp), %rdx movq 80(%rsp), %rsi movq 112(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L43 .L55: call __stack_chk_fail@PLT .cfi_endproc .LFE2554: .size main, .-main .section .rodata.str1.1 .LC11: .string "_Z12sim_severityPfPjS_j" .LC12: .string "_Z8sim_freqPjPfj" .LC13: .string "precalc_xorwow_matrix" .LC14: .string "precalc_xorwow_offset_matrix" .LC15: .string "mrg32k3aM1" .LC16: .string "mrg32k3aM2" .LC17: .string "mrg32k3aM1SubSeq" .LC18: .string "mrg32k3aM2SubSeq" .LC19: .string "mrg32k3aM1Seq" .LC20: .string "mrg32k3aM2Seq" .LC21: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2585: .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 .LC11(%rip), %rdx movq %rdx, %rcx leaq _Z12sim_severityPfPjS_j(%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 .LC12(%rip), %rdx movq %rdx, %rcx leaq _Z8sim_freqPjPfj(%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 .LC13(%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 .LC14(%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 .LC15(%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 .LC16(%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 .LC17(%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 .LC18(%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 .LC19(%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 .LC20(%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 .LC21(%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 .LFE2585: .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 N_BYTES_F .bss .align 4 .type N_BYTES_F, @object .size N_BYTES_F, 4 N_BYTES_F: .zero 4 .globl N_BYTES_I .align 4 .type N_BYTES_I, @object .size N_BYTES_I, 4 N_BYTES_I: .zero 4 .globl N_THRD .align 4 .type N_THRD, @object .size N_THRD, 4 N_THRD: .zero 4 .globl N_BLK .align 4 .type N_BLK, @object .size N_BLK, 4 N_BLK: .zero 4 .globl N_SIMS .align 4 .type N_SIMS, @object .size N_SIMS, 4 N_SIMS: .zero 4 .globl N_BYTES_PRM .data .align 4 .type N_BYTES_PRM, @object .size N_BYTES_PRM, 4 N_BYTES_PRM: .long 12 .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 <stdio.h> #include <algorithm> #include <cstdlib> #include <curand.h> #include <curand_kernel.h> // In the following section, define the prob distribution parameters #define N_PARAMS 3 #define PARAM1 50.0f, 3.0f, 0.5f // format: LAMBDA, A, B #define PARAM2 1.5f, 0.8f, 5.0f // parameters saved as constants unsigned int N_BYTES_PRM = N_PARAMS * sizeof(float); // size of parameter unsigned int N_SIMS, N_BLK, N_THRD, N_BYTES_I, N_BYTES_F; const unsigned int MAX_THREADS = 512; // max threads per block // Calculate and return mean of an array of floats float calcMean(float arr[], unsigned int const n) { double sum = 0.0; for (unsigned int i=0; i<n; i++) { sum += (arr[i] / n); } return sum; } __host__ cudaEvent_t get_time(void) { cudaEvent_t time; cudaEventCreate(&time); cudaEventRecord(time); return time; } // Based on parameter, draw poisson random number as frequency of losses __global__ void sim_freq(unsigned int *f_out, float *prm, unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { float lambda = prm[0]; // lambda for poisson curandState_t state; // initialize rand state curand_init(tid, 0, 0, &state); // set seed to thread index f_out[tid] = curand_poisson(&state, lambda); // save loss frequency } } // Based on parameter and freq, draw and sum pareto loss amounts __global__ void sim_severity(float *loss_out, unsigned int *freq, float *prm, const unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { double A = prm[1]; double B = prm[2]; // two pareto parameters curandState_t state; // initialize rand state curand_init(tid, 0, 0, &state); // set seed to thread index double sum = 0.0; double unif = 0.0; // temp var for storing uniform rand for (int f=0; f < freq[tid]; f++) { unif = curand_uniform_double(&state); // draw unif rand as double sum += B / pow(1-unif, 1/A); // quantile function (inverse CDF) } loss_out[tid] = (float) sum; // sum of all losses } } void asynch() { return; } int main(int argc, char* argv[]) { if (argc == 2) { // get number of simulations based on CMDLINE input N_SIMS = atoi(argv[1]); } else { printf("Usage: %s [nSimulations].\n", argv[0]); return EXIT_FAILURE; } N_BLK = N_SIMS / MAX_THREADS + 1; // min of one block N_THRD = std::min(N_SIMS, MAX_THREADS); // num of threads per block N_BYTES_F = N_SIMS * sizeof(float); // size of loss array N_BYTES_I = N_SIMS * sizeof(unsigned int); // size of frequency array printf("Running %u simulations ...\n", N_SIMS); cudaStream_t s1, s2; // Create and initialize streams cudaStreamCreate(&s1); cudaStreamCreate(&s2); // allocate and copy parameter to device float h_prm1 [N_PARAMS] = {PARAM1}; float h_prm2 [N_PARAMS] = {PARAM2}; cudaHostRegister(h_prm1, N_BYTES_PRM, cudaHostRegisterDefault); cudaHostRegister(h_prm2, N_BYTES_PRM, cudaHostRegisterDefault); float *d_prm1, *d_prm2; // parameter on device memory cudaMalloc((void **)&d_prm1, N_BYTES_PRM); // allocate and copy cudaMalloc((void **)&d_prm2, N_BYTES_PRM); cudaMemcpyAsync(d_prm1, h_prm1, N_BYTES_PRM, cudaMemcpyHostToDevice, s1); cudaMemcpyAsync(d_prm2, h_prm2, N_BYTES_PRM, cudaMemcpyHostToDevice, s2); unsigned int *h_freq1, *d_freq1, *h_freq2, *d_freq2; // frequency arrays float *h_loss1, *d_loss1, *h_loss2, *d_loss2; // loss arrays cudaMalloc((void **)&d_freq1, N_BYTES_I); // device mem for freq and loss cudaMalloc((void **)&d_loss1, N_BYTES_F); cudaMalloc((void **)&d_freq2, N_BYTES_I); cudaMalloc((void **)&d_loss2, N_BYTES_F); cudaMallocHost((void**)&h_freq1, N_BYTES_I); // pinned host array cudaMallocHost((void**)&h_loss1, N_BYTES_F); cudaMallocHost((void**)&h_freq2, N_BYTES_I); cudaMallocHost((void**)&h_loss2, N_BYTES_F); float dur, mean1, mean2; // to record duration and averages /****** asynchronus run ****** ******************************/ cudaEvent_t start = get_time(); // start clock cudaEvent_t copyEnd1, copyEnd2; // event to ensure copying loss is finished cudaEventCreate(&copyEnd1); cudaEventCreate(&copyEnd2); // first simulate frequency of losses in two streams sim_freq<<<N_BLK, N_THRD, 0, s1>>>(d_freq1, d_prm1, N_SIMS); sim_freq<<<N_BLK, N_THRD, 0, s2>>>(d_freq2, d_prm2, N_SIMS); // based on frequency, draw pareto loss amounts and sum sim_severity<<<N_BLK, N_THRD, 0, s1>>>(d_loss1, d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD, 0, s2>>>(d_loss2, d_freq2, d_prm2, N_SIMS); cudaMemcpyAsync(h_loss1, d_loss1, N_BYTES_F, cudaMemcpyDeviceToHost, s1); cudaEventRecord(copyEnd1, s1); // finish copying result in stream 1 cudaMemcpyAsync(h_loss2, d_loss2, N_BYTES_F, cudaMemcpyDeviceToHost, s2); cudaEventRecord(copyEnd2, s2); // finish copying result in stream 2 cudaMemcpyAsync(h_freq1, d_freq1, N_BYTES_I, cudaMemcpyDeviceToHost, s1); cudaMemcpyAsync(h_freq2, d_freq2, N_BYTES_I, cudaMemcpyDeviceToHost, s2); cudaEventSynchronize(copyEnd1); // wait for result copy before calculation mean1 = calcMean(h_loss1, N_SIMS); cudaEventSynchronize(copyEnd2); // wait for result copy before calculation mean2 = calcMean(h_loss2, N_SIMS); cudaStreamSynchronize( s1 ); cudaStreamSynchronize( s2 ); cudaEvent_t stop = get_time(); // stop clock cudaEventSynchronize(stop); cudaEventElapsedTime(&dur, start, stop); printf("\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); /****** synchronus run ******* ******************************/ start = get_time(); sim_freq<<<N_BLK, N_THRD>>>(d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss1, d_freq1, d_prm1, N_SIMS); cudaMemcpy(h_freq1, d_freq1, N_BYTES_I, cudaMemcpyDeviceToHost); cudaMemcpy(h_loss1, d_loss1, N_BYTES_F, cudaMemcpyDeviceToHost); sim_freq<<<N_BLK, N_THRD>>>(d_freq2, d_prm2, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss2, d_freq2, d_prm2, N_SIMS); cudaMemcpy(h_freq2, d_freq2, N_BYTES_I, cudaMemcpyDeviceToHost); cudaMemcpy(h_loss2, d_loss2, N_BYTES_F, cudaMemcpyDeviceToHost); mean1 = calcMean(h_loss1, N_SIMS); mean2 = calcMean(h_loss2, N_SIMS); stop = get_time(); // stop time cudaEventSynchronize(stop); cudaEventElapsedTime(&dur, start, stop); printf("\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); return EXIT_SUCCESS; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <algorithm> #include <cstdlib> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> // In the following section, define the prob distribution parameters #define N_PARAMS 3 #define PARAM1 50.0f, 3.0f, 0.5f // format: LAMBDA, A, B #define PARAM2 1.5f, 0.8f, 5.0f // parameters saved as constants unsigned int N_BYTES_PRM = N_PARAMS * sizeof(float); // size of parameter unsigned int N_SIMS, N_BLK, N_THRD, N_BYTES_I, N_BYTES_F; const unsigned int MAX_THREADS = 512; // max threads per block // Calculate and return mean of an array of floats float calcMean(float arr[], unsigned int const n) { double sum = 0.0; for (unsigned int i=0; i<n; i++) { sum += (arr[i] / n); } return sum; } __host__ hipEvent_t get_time(void) { hipEvent_t time; hipEventCreate(&time); hipEventRecord(time); return time; } // Based on parameter, draw poisson random number as frequency of losses __global__ void sim_freq(unsigned int *f_out, float *prm, unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { float lambda = prm[0]; // lambda for poisson hiprandState_t state; // initialize rand state hiprand_init(tid, 0, 0, &state); // set seed to thread index f_out[tid] = hiprand_poisson(&state, lambda); // save loss frequency } } // Based on parameter and freq, draw and sum pareto loss amounts __global__ void sim_severity(float *loss_out, unsigned int *freq, float *prm, const unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { double A = prm[1]; double B = prm[2]; // two pareto parameters hiprandState_t state; // initialize rand state hiprand_init(tid, 0, 0, &state); // set seed to thread index double sum = 0.0; double unif = 0.0; // temp var for storing uniform rand for (int f=0; f < freq[tid]; f++) { unif = hiprand_uniform_double(&state); // draw unif rand as double sum += B / pow(1-unif, 1/A); // quantile function (inverse CDF) } loss_out[tid] = (float) sum; // sum of all losses } } void asynch() { return; } int main(int argc, char* argv[]) { if (argc == 2) { // get number of simulations based on CMDLINE input N_SIMS = atoi(argv[1]); } else { printf("Usage: %s [nSimulations].\n", argv[0]); return EXIT_FAILURE; } N_BLK = N_SIMS / MAX_THREADS + 1; // min of one block N_THRD = std::min(N_SIMS, MAX_THREADS); // num of threads per block N_BYTES_F = N_SIMS * sizeof(float); // size of loss array N_BYTES_I = N_SIMS * sizeof(unsigned int); // size of frequency array printf("Running %u simulations ...\n", N_SIMS); hipStream_t s1, s2; // Create and initialize streams hipStreamCreate(&s1); hipStreamCreate(&s2); // allocate and copy parameter to device float h_prm1 [N_PARAMS] = {PARAM1}; float h_prm2 [N_PARAMS] = {PARAM2}; hipHostRegister(h_prm1, N_BYTES_PRM, hipHostRegisterDefault); hipHostRegister(h_prm2, N_BYTES_PRM, hipHostRegisterDefault); float *d_prm1, *d_prm2; // parameter on device memory hipMalloc((void **)&d_prm1, N_BYTES_PRM); // allocate and copy hipMalloc((void **)&d_prm2, N_BYTES_PRM); hipMemcpyAsync(d_prm1, h_prm1, N_BYTES_PRM, hipMemcpyHostToDevice, s1); hipMemcpyAsync(d_prm2, h_prm2, N_BYTES_PRM, hipMemcpyHostToDevice, s2); unsigned int *h_freq1, *d_freq1, *h_freq2, *d_freq2; // frequency arrays float *h_loss1, *d_loss1, *h_loss2, *d_loss2; // loss arrays hipMalloc((void **)&d_freq1, N_BYTES_I); // device mem for freq and loss hipMalloc((void **)&d_loss1, N_BYTES_F); hipMalloc((void **)&d_freq2, N_BYTES_I); hipMalloc((void **)&d_loss2, N_BYTES_F); hipHostMalloc((void**)&h_freq1, N_BYTES_I, hipHostMallocDefault); // pinned host array hipHostMalloc((void**)&h_loss1, N_BYTES_F, hipHostMallocDefault); hipHostMalloc((void**)&h_freq2, N_BYTES_I, hipHostMallocDefault); hipHostMalloc((void**)&h_loss2, N_BYTES_F, hipHostMallocDefault); float dur, mean1, mean2; // to record duration and averages /****** asynchronus run ****** ******************************/ hipEvent_t start = get_time(); // start clock hipEvent_t copyEnd1, copyEnd2; // event to ensure copying loss is finished hipEventCreate(&copyEnd1); hipEventCreate(&copyEnd2); // first simulate frequency of losses in two streams sim_freq<<<N_BLK, N_THRD, 0, s1>>>(d_freq1, d_prm1, N_SIMS); sim_freq<<<N_BLK, N_THRD, 0, s2>>>(d_freq2, d_prm2, N_SIMS); // based on frequency, draw pareto loss amounts and sum sim_severity<<<N_BLK, N_THRD, 0, s1>>>(d_loss1, d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD, 0, s2>>>(d_loss2, d_freq2, d_prm2, N_SIMS); hipMemcpyAsync(h_loss1, d_loss1, N_BYTES_F, hipMemcpyDeviceToHost, s1); hipEventRecord(copyEnd1, s1); // finish copying result in stream 1 hipMemcpyAsync(h_loss2, d_loss2, N_BYTES_F, hipMemcpyDeviceToHost, s2); hipEventRecord(copyEnd2, s2); // finish copying result in stream 2 hipMemcpyAsync(h_freq1, d_freq1, N_BYTES_I, hipMemcpyDeviceToHost, s1); hipMemcpyAsync(h_freq2, d_freq2, N_BYTES_I, hipMemcpyDeviceToHost, s2); hipEventSynchronize(copyEnd1); // wait for result copy before calculation mean1 = calcMean(h_loss1, N_SIMS); hipEventSynchronize(copyEnd2); // wait for result copy before calculation mean2 = calcMean(h_loss2, N_SIMS); hipStreamSynchronize( s1 ); hipStreamSynchronize( s2 ); hipEvent_t stop = get_time(); // stop clock hipEventSynchronize(stop); hipEventElapsedTime(&dur, start, stop); printf("\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); /****** synchronus run ******* ******************************/ start = get_time(); sim_freq<<<N_BLK, N_THRD>>>(d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss1, d_freq1, d_prm1, N_SIMS); hipMemcpy(h_freq1, d_freq1, N_BYTES_I, hipMemcpyDeviceToHost); hipMemcpy(h_loss1, d_loss1, N_BYTES_F, hipMemcpyDeviceToHost); sim_freq<<<N_BLK, N_THRD>>>(d_freq2, d_prm2, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss2, d_freq2, d_prm2, N_SIMS); hipMemcpy(h_freq2, d_freq2, N_BYTES_I, hipMemcpyDeviceToHost); hipMemcpy(h_loss2, d_loss2, N_BYTES_F, hipMemcpyDeviceToHost); mean1 = calcMean(h_loss1, N_SIMS); mean2 = calcMean(h_loss2, N_SIMS); stop = get_time(); // stop time hipEventSynchronize(stop); hipEventElapsedTime(&dur, start, stop); printf("\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); return EXIT_SUCCESS; }
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 <algorithm> #include <cstdlib> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> // In the following section, define the prob distribution parameters #define N_PARAMS 3 #define PARAM1 50.0f, 3.0f, 0.5f // format: LAMBDA, A, B #define PARAM2 1.5f, 0.8f, 5.0f // parameters saved as constants unsigned int N_BYTES_PRM = N_PARAMS * sizeof(float); // size of parameter unsigned int N_SIMS, N_BLK, N_THRD, N_BYTES_I, N_BYTES_F; const unsigned int MAX_THREADS = 512; // max threads per block // Calculate and return mean of an array of floats float calcMean(float arr[], unsigned int const n) { double sum = 0.0; for (unsigned int i=0; i<n; i++) { sum += (arr[i] / n); } return sum; } __host__ hipEvent_t get_time(void) { hipEvent_t time; hipEventCreate(&time); hipEventRecord(time); return time; } // Based on parameter, draw poisson random number as frequency of losses __global__ void sim_freq(unsigned int *f_out, float *prm, unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { float lambda = prm[0]; // lambda for poisson hiprandState_t state; // initialize rand state hiprand_init(tid, 0, 0, &state); // set seed to thread index f_out[tid] = hiprand_poisson(&state, lambda); // save loss frequency } } // Based on parameter and freq, draw and sum pareto loss amounts __global__ void sim_severity(float *loss_out, unsigned int *freq, float *prm, const unsigned int N) { unsigned int const tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < N) { double A = prm[1]; double B = prm[2]; // two pareto parameters hiprandState_t state; // initialize rand state hiprand_init(tid, 0, 0, &state); // set seed to thread index double sum = 0.0; double unif = 0.0; // temp var for storing uniform rand for (int f=0; f < freq[tid]; f++) { unif = hiprand_uniform_double(&state); // draw unif rand as double sum += B / pow(1-unif, 1/A); // quantile function (inverse CDF) } loss_out[tid] = (float) sum; // sum of all losses } } void asynch() { return; } int main(int argc, char* argv[]) { if (argc == 2) { // get number of simulations based on CMDLINE input N_SIMS = atoi(argv[1]); } else { printf("Usage: %s [nSimulations].\n", argv[0]); return EXIT_FAILURE; } N_BLK = N_SIMS / MAX_THREADS + 1; // min of one block N_THRD = std::min(N_SIMS, MAX_THREADS); // num of threads per block N_BYTES_F = N_SIMS * sizeof(float); // size of loss array N_BYTES_I = N_SIMS * sizeof(unsigned int); // size of frequency array printf("Running %u simulations ...\n", N_SIMS); hipStream_t s1, s2; // Create and initialize streams hipStreamCreate(&s1); hipStreamCreate(&s2); // allocate and copy parameter to device float h_prm1 [N_PARAMS] = {PARAM1}; float h_prm2 [N_PARAMS] = {PARAM2}; hipHostRegister(h_prm1, N_BYTES_PRM, hipHostRegisterDefault); hipHostRegister(h_prm2, N_BYTES_PRM, hipHostRegisterDefault); float *d_prm1, *d_prm2; // parameter on device memory hipMalloc((void **)&d_prm1, N_BYTES_PRM); // allocate and copy hipMalloc((void **)&d_prm2, N_BYTES_PRM); hipMemcpyAsync(d_prm1, h_prm1, N_BYTES_PRM, hipMemcpyHostToDevice, s1); hipMemcpyAsync(d_prm2, h_prm2, N_BYTES_PRM, hipMemcpyHostToDevice, s2); unsigned int *h_freq1, *d_freq1, *h_freq2, *d_freq2; // frequency arrays float *h_loss1, *d_loss1, *h_loss2, *d_loss2; // loss arrays hipMalloc((void **)&d_freq1, N_BYTES_I); // device mem for freq and loss hipMalloc((void **)&d_loss1, N_BYTES_F); hipMalloc((void **)&d_freq2, N_BYTES_I); hipMalloc((void **)&d_loss2, N_BYTES_F); hipHostMalloc((void**)&h_freq1, N_BYTES_I, hipHostMallocDefault); // pinned host array hipHostMalloc((void**)&h_loss1, N_BYTES_F, hipHostMallocDefault); hipHostMalloc((void**)&h_freq2, N_BYTES_I, hipHostMallocDefault); hipHostMalloc((void**)&h_loss2, N_BYTES_F, hipHostMallocDefault); float dur, mean1, mean2; // to record duration and averages /****** asynchronus run ****** ******************************/ hipEvent_t start = get_time(); // start clock hipEvent_t copyEnd1, copyEnd2; // event to ensure copying loss is finished hipEventCreate(&copyEnd1); hipEventCreate(&copyEnd2); // first simulate frequency of losses in two streams sim_freq<<<N_BLK, N_THRD, 0, s1>>>(d_freq1, d_prm1, N_SIMS); sim_freq<<<N_BLK, N_THRD, 0, s2>>>(d_freq2, d_prm2, N_SIMS); // based on frequency, draw pareto loss amounts and sum sim_severity<<<N_BLK, N_THRD, 0, s1>>>(d_loss1, d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD, 0, s2>>>(d_loss2, d_freq2, d_prm2, N_SIMS); hipMemcpyAsync(h_loss1, d_loss1, N_BYTES_F, hipMemcpyDeviceToHost, s1); hipEventRecord(copyEnd1, s1); // finish copying result in stream 1 hipMemcpyAsync(h_loss2, d_loss2, N_BYTES_F, hipMemcpyDeviceToHost, s2); hipEventRecord(copyEnd2, s2); // finish copying result in stream 2 hipMemcpyAsync(h_freq1, d_freq1, N_BYTES_I, hipMemcpyDeviceToHost, s1); hipMemcpyAsync(h_freq2, d_freq2, N_BYTES_I, hipMemcpyDeviceToHost, s2); hipEventSynchronize(copyEnd1); // wait for result copy before calculation mean1 = calcMean(h_loss1, N_SIMS); hipEventSynchronize(copyEnd2); // wait for result copy before calculation mean2 = calcMean(h_loss2, N_SIMS); hipStreamSynchronize( s1 ); hipStreamSynchronize( s2 ); hipEvent_t stop = get_time(); // stop clock hipEventSynchronize(stop); hipEventElapsedTime(&dur, start, stop); printf("\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); /****** synchronus run ******* ******************************/ start = get_time(); sim_freq<<<N_BLK, N_THRD>>>(d_freq1, d_prm1, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss1, d_freq1, d_prm1, N_SIMS); hipMemcpy(h_freq1, d_freq1, N_BYTES_I, hipMemcpyDeviceToHost); hipMemcpy(h_loss1, d_loss1, N_BYTES_F, hipMemcpyDeviceToHost); sim_freq<<<N_BLK, N_THRD>>>(d_freq2, d_prm2, N_SIMS); sim_severity<<<N_BLK, N_THRD>>>(d_loss2, d_freq2, d_prm2, N_SIMS); hipMemcpy(h_freq2, d_freq2, N_BYTES_I, hipMemcpyDeviceToHost); hipMemcpy(h_loss2, d_loss2, N_BYTES_F, hipMemcpyDeviceToHost); mean1 = calcMean(h_loss1, N_SIMS); mean2 = calcMean(h_loss2, N_SIMS); stop = get_time(); // stop time hipEventSynchronize(stop); hipEventElapsedTime(&dur, start, stop); printf("\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n", mean1, mean2, dur); return EXIT_SUCCESS; }
.text .file "stream_registers.hip" .globl _Z8calcMeanPfj # -- Begin function _Z8calcMeanPfj .p2align 4, 0x90 .type _Z8calcMeanPfj,@function _Z8calcMeanPfj: # @_Z8calcMeanPfj .cfi_startproc # %bb.0: testl %esi, %esi je .LBB0_1 # %bb.4: # %.lr.ph movl %esi, %eax cvtsi2ss %rax, %xmm1 xorpd %xmm0, %xmm0 xorl %ecx, %ecx .p2align 4, 0x90 .LBB0_5: # =>This Inner Loop Header: Depth=1 movss (%rdi,%rcx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm1, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm0 incq %rcx cmpq %rcx, %rax jne .LBB0_5 # %bb.2: # %._crit_edge.loopexit cvtsd2ss %xmm0, %xmm0 retq .LBB0_1: xorps %xmm0, %xmm0 retq .Lfunc_end0: .size _Z8calcMeanPfj, .Lfunc_end0-_Z8calcMeanPfj .cfi_endproc # -- End function .globl _Z8get_timev # -- Begin function _Z8get_timev .p2align 4, 0x90 .type _Z8get_timev,@function _Z8get_timev: # @_Z8get_timev .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movq %rsp, %rdi callq hipEventCreate movq (%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq (%rsp), %rax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z8get_timev, .Lfunc_end1-_Z8get_timev .cfi_endproc # -- End function .globl _Z23__device_stub__sim_freqPjPfj # -- Begin function _Z23__device_stub__sim_freqPjPfj .p2align 4, 0x90 .type _Z23__device_stub__sim_freqPjPfj,@function _Z23__device_stub__sim_freqPjPfj: # @_Z23__device_stub__sim_freqPjPfj .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%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 $_Z8sim_freqPjPfj, %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_end2: .size _Z23__device_stub__sim_freqPjPfj, .Lfunc_end2-_Z23__device_stub__sim_freqPjPfj .cfi_endproc # -- End function .globl _Z27__device_stub__sim_severityPfPjS_j # -- Begin function _Z27__device_stub__sim_severityPfPjS_j .p2align 4, 0x90 .type _Z27__device_stub__sim_severityPfPjS_j,@function _Z27__device_stub__sim_severityPfPjS_j: # @_Z27__device_stub__sim_severityPfPjS_j .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z12sim_severityPfPjS_j, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end3: .size _Z27__device_stub__sim_severityPfPjS_j, .Lfunc_end3-_Z27__device_stub__sim_severityPfPjS_j .cfi_endproc # -- End function .globl _Z6asynchv # -- Begin function _Z6asynchv .p2align 4, 0x90 .type _Z6asynchv,@function _Z6asynchv: # @_Z6asynchv .cfi_startproc # %bb.0: retq .Lfunc_end4: .size _Z6asynchv, .Lfunc_end4-_Z6asynchv .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 $272, %rsp # imm = 0x110 .cfi_def_cfa_offset 304 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 cmpl $2, %edi jne .LBB5_35 # %bb.1: movabsq $4294967296, %r15 # imm = 0x100000000 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movl %eax, N_SIMS(%rip) movl %eax, %ecx shrl $9, %ecx incl %ecx movl %ecx, N_BLK(%rip) cmpl $512, %eax # imm = 0x200 movl $512, %ecx # imm = 0x200 cmovbl %eax, %ecx movl %ecx, N_THRD(%rip) leal (,%rax,4), %ecx movl %ecx, N_BYTES_F(%rip) movl %ecx, N_BYTES_I(%rip) movl $.L.str.1, %edi movl %eax, %esi xorl %eax, %eax callq printf leaq 120(%rsp), %rdi callq hipStreamCreate leaq 112(%rsp), %rdi callq hipStreamCreate movabsq $4629700418048884736, %rax # imm = 0x4040000042480000 movq %rax, 260(%rsp) movl $1056964608, 268(%rsp) # imm = 0x3F000000 movabsq $4561245704529379328, %rax # imm = 0x3F4CCCCD3FC00000 movq %rax, 248(%rsp) movl $1084227584, 256(%rsp) # imm = 0x40A00000 movl N_BYTES_PRM(%rip), %esi leaq 260(%rsp), %r14 movq %r14, %rdi xorl %edx, %edx callq hipHostRegister movl N_BYTES_PRM(%rip), %esi leaq 248(%rsp), %rbx movq %rbx, %rdi xorl %edx, %edx callq hipHostRegister movl N_BYTES_PRM(%rip), %esi leaq 160(%rsp), %rdi callq hipMalloc movl N_BYTES_PRM(%rip), %esi leaq 152(%rsp), %rdi callq hipMalloc movq 160(%rsp), %rdi movl N_BYTES_PRM(%rip), %edx movq 120(%rsp), %r8 movq %r14, %rsi movl $1, %ecx callq hipMemcpyAsync movq 152(%rsp), %rdi movl N_BYTES_PRM(%rip), %edx movq 112(%rsp), %r8 movq %rbx, %rsi movl $1, %ecx callq hipMemcpyAsync movl N_BYTES_I(%rip), %esi leaq 144(%rsp), %rdi callq hipMalloc movl N_BYTES_F(%rip), %esi leaq 192(%rsp), %rdi callq hipMalloc movl N_BYTES_I(%rip), %esi leaq 136(%rsp), %rdi callq hipMalloc movl N_BYTES_F(%rip), %esi leaq 176(%rsp), %rdi callq hipMalloc movl N_BYTES_I(%rip), %esi leaq 240(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_F(%rip), %esi leaq 200(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_I(%rip), %esi leaq 232(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_F(%rip), %esi leaq 184(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %rbx leaq 224(%rsp), %rdi callq hipEventCreate leaq 216(%rsp), %rdi callq hipEventCreate movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 120(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_3 # %bb.2: movq 144(%rsp), %rax movq 160(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_3: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 112(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_5 # %bb.4: movq 136(%rsp), %rax movq 152(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_5: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 120(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_7 # %bb.6: movq 192(%rsp), %rax movq 144(%rsp), %rcx movq 160(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_7: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 112(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_9 # %bb.8: movq 176(%rsp), %rax movq 136(%rsp), %rcx movq 152(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_9: movq 200(%rsp), %rdi movq 192(%rsp), %rsi movl N_BYTES_F(%rip), %edx movq 120(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 224(%rsp), %rdi movq 120(%rsp), %rsi callq hipEventRecord movq 184(%rsp), %rdi movq 176(%rsp), %rsi movl N_BYTES_F(%rip), %edx movq 112(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 216(%rsp), %rdi movq 112(%rsp), %rsi callq hipEventRecord movq 240(%rsp), %rdi movq 144(%rsp), %rsi movl N_BYTES_I(%rip), %edx movq 120(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 232(%rsp), %rdi movq 136(%rsp), %rsi movl N_BYTES_I(%rip), %edx movq 112(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 224(%rsp), %rdi callq hipEventSynchronize movl N_SIMS(%rip), %eax xorps %xmm0, %xmm0 movsd %xmm0, 168(%rsp) # 8-byte Spill testq %rax, %rax je .LBB5_13 # %bb.10: # %.lr.ph.i xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 200(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_11: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_11 # %bb.12: # %._crit_edge.loopexit.i xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 .LBB5_13: # %_Z8calcMeanPfj.exit movsd %xmm0, 208(%rsp) # 8-byte Spill movq 216(%rsp), %rdi callq hipEventSynchronize movl N_SIMS(%rip), %eax testq %rax, %rax je .LBB5_17 # %bb.14: # %.lr.ph.i69 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 184(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_15: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_15 # %bb.16: # %._crit_edge.loopexit.i75 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 movsd %xmm0, 168(%rsp) # 8-byte Spill .LBB5_17: # %_Z8calcMeanPfj.exit77 movq 120(%rsp), %rdi callq hipStreamSynchronize movq 112(%rsp), %rdi callq hipStreamSynchronize leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %r14 movq %r14, %rdi callq hipEventSynchronize leaq 132(%rsp), %rdi movq %rbx, %rsi movq %r14, %rdx callq hipEventElapsedTime movss 132(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm0, %xmm2 movl $.L.str.2, %edi movsd 208(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movsd 168(%rsp), %xmm1 # 8-byte Reload # xmm1 = mem[0],zero movb $3, %al callq printf leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %rbx movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_19 # %bb.18: movq 144(%rsp), %rax movq 160(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_19: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_21 # %bb.20: movq 192(%rsp), %rax movq 144(%rsp), %rcx movq 160(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_21: movq 240(%rsp), %rdi movq 144(%rsp), %rsi movl N_BYTES_I(%rip), %edx movl $2, %ecx callq hipMemcpy movq 200(%rsp), %rdi movq 192(%rsp), %rsi movl N_BYTES_F(%rip), %edx movl $2, %ecx callq hipMemcpy movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_23 # %bb.22: movq 136(%rsp), %rax movq 152(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_23: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_25 # %bb.24: movq 176(%rsp), %rax movq 136(%rsp), %rcx movq 152(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_25: movq 232(%rsp), %rdi movq 136(%rsp), %rsi movl N_BYTES_I(%rip), %edx movl $2, %ecx callq hipMemcpy movq 184(%rsp), %rdi movq 176(%rsp), %rsi movl N_BYTES_F(%rip), %edx movl $2, %ecx callq hipMemcpy movl N_SIMS(%rip), %eax xorps %xmm3, %xmm3 xorps %xmm0, %xmm0 testq %rax, %rax je .LBB5_29 # %bb.26: # %.lr.ph.i103 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 200(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_27: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_27 # %bb.28: # %._crit_edge.loopexit.i109 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 .LBB5_29: # %_Z8calcMeanPfj.exit111 movsd %xmm0, 168(%rsp) # 8-byte Spill testq %rax, %rax je .LBB5_33 # %bb.30: # %.lr.ph.i113 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 184(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_31: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_31 # %bb.32: # %._crit_edge.loopexit.i119 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 xorps %xmm3, %xmm3 cvtss2sd %xmm0, %xmm3 .LBB5_33: # %_Z8calcMeanPfj.exit121 movsd %xmm3, 208(%rsp) # 8-byte Spill leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %r14d, %r14d xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %r15 movq %r15, %rdi callq hipEventSynchronize leaq 132(%rsp), %rdi movq %rbx, %rsi movq %r15, %rdx callq hipEventElapsedTime movss 132(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm0, %xmm2 movl $.L.str.3, %edi movsd 168(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movsd 208(%rsp), %xmm1 # 8-byte Reload # xmm1 = mem[0],zero movb $3, %al callq printf jmp .LBB5_34 .LBB5_35: movq (%rsi), %rsi movl $.L.str, %edi xorl %eax, %eax callq printf movl $1, %r14d .LBB5_34: movl %r14d, %eax addq $272, %rsp # imm = 0x110 .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_end5: .size main, .Lfunc_end5-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 .LBB6_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB6_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8sim_freqPjPfj, %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 $_Z12sim_severityPfPjS_j, %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_end6: .size __hip_module_ctor, .Lfunc_end6-__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 .LBB7_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 .LBB7_2: retq .Lfunc_end7: .size __hip_module_dtor, .Lfunc_end7-__hip_module_dtor .cfi_endproc # -- End function .type N_BYTES_PRM,@object # @N_BYTES_PRM .data .globl N_BYTES_PRM .p2align 2, 0x0 N_BYTES_PRM: .long 12 # 0xc .size N_BYTES_PRM, 4 .type N_SIMS,@object # @N_SIMS .bss .globl N_SIMS .p2align 2, 0x0 N_SIMS: .long 0 # 0x0 .size N_SIMS, 4 .type N_BLK,@object # @N_BLK .globl N_BLK .p2align 2, 0x0 N_BLK: .long 0 # 0x0 .size N_BLK, 4 .type N_THRD,@object # @N_THRD .globl N_THRD .p2align 2, 0x0 N_THRD: .long 0 # 0x0 .size N_THRD, 4 .type N_BYTES_I,@object # @N_BYTES_I .globl N_BYTES_I .p2align 2, 0x0 N_BYTES_I: .long 0 # 0x0 .size N_BYTES_I, 4 .type N_BYTES_F,@object # @N_BYTES_F .globl N_BYTES_F .p2align 2, 0x0 N_BYTES_F: .long 0 # 0x0 .size N_BYTES_F, 4 .type _Z8sim_freqPjPfj,@object # @_Z8sim_freqPjPfj .section .rodata,"a",@progbits .globl _Z8sim_freqPjPfj .p2align 3, 0x0 _Z8sim_freqPjPfj: .quad _Z23__device_stub__sim_freqPjPfj .size _Z8sim_freqPjPfj, 8 .type _Z12sim_severityPfPjS_j,@object # @_Z12sim_severityPfPjS_j .globl _Z12sim_severityPfPjS_j .p2align 3, 0x0 _Z12sim_severityPfPjS_j: .quad _Z27__device_stub__sim_severityPfPjS_j .size _Z12sim_severityPfPjS_j, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Usage: %s [nSimulations].\n" .size .L.str, 27 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Running %u simulations ...\n" .size .L.str.1, 28 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .size .L.str.2, 59 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .size .L.str.3, 58 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z8sim_freqPjPfj" .size .L__unnamed_1, 17 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z12sim_severityPfPjS_j" .size .L__unnamed_2, 24 .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 _Z23__device_stub__sim_freqPjPfj .addrsig_sym _Z27__device_stub__sim_severityPfPjS_j .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8sim_freqPjPfj .addrsig_sym _Z12sim_severityPfPjS_j .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_0016f86c_00000000-6_stream_registers.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2558: .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 .LFE2558: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z8calcMeanPfj .type _Z8calcMeanPfj, @function _Z8calcMeanPfj: .LFB2551: .cfi_startproc endbr64 testl %esi, %esi je .L8 movq %rdi, %rax movl %esi, %edx leaq (%rdi,%rdx,4), %rdx pxor %xmm0, %xmm0 movl %esi, %esi pxor %xmm2, %xmm2 cvtsi2ssq %rsi, %xmm2 .L10: movss (%rax), %xmm1 divss %xmm2, %xmm1 cvtss2sd %xmm1, %xmm1 addsd %xmm1, %xmm0 addq $4, %rax cmpq %rdx, %rax jne .L10 .L4: cvtsd2ss %xmm0, %xmm0 ret .L8: pxor %xmm0, %xmm0 jmp .L4 .cfi_endproc .LFE2551: .size _Z8calcMeanPfj, .-_Z8calcMeanPfj .globl _Z8get_timev .type _Z8get_timev, @function _Z8get_timev: .LFB2552: .cfi_startproc endbr64 subq $24, %rsp .cfi_def_cfa_offset 32 movq %fs:40, %rax movq %rax, 8(%rsp) xorl %eax, %eax movq %rsp, %rdi call cudaEventCreate@PLT movl $0, %esi movq (%rsp), %rdi call cudaEventRecord@PLT movq (%rsp), %rax movq 8(%rsp), %rdx subq %fs:40, %rdx jne .L14 addq $24, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2552: .size _Z8get_timev, .-_Z8get_timev .globl _Z6asynchv .type _Z6asynchv, @function _Z6asynchv: .LFB2553: .cfi_startproc endbr64 ret .cfi_endproc .LFE2553: .size _Z6asynchv, .-_Z6asynchv .globl _Z30__device_stub__Z8sim_freqPjPfjPjPfj .type _Z30__device_stub__Z8sim_freqPjPfjPjPfj, @function _Z30__device_stub__Z8sim_freqPjPfjPjPfj: .LFB2580: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 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 .L20 .L16: movq 120(%rsp), %rax subq %fs:40, %rax jne .L21 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L20: .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 _Z8sim_freqPjPfj(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L16 .L21: call __stack_chk_fail@PLT .cfi_endproc .LFE2580: .size _Z30__device_stub__Z8sim_freqPjPfjPjPfj, .-_Z30__device_stub__Z8sim_freqPjPfjPjPfj .globl _Z8sim_freqPjPfj .type _Z8sim_freqPjPfj, @function _Z8sim_freqPjPfj: .LFB2581: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z8sim_freqPjPfjPjPfj addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2581: .size _Z8sim_freqPjPfj, .-_Z8sim_freqPjPfj .globl _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j .type _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j, @function _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j: .LFB2582: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%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) leaq 4(%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 .L28 .L24: movq 136(%rsp), %rax subq %fs:40, %rax jne .L29 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L28: .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 _Z12sim_severityPfPjS_j(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L24 .L29: call __stack_chk_fail@PLT .cfi_endproc .LFE2582: .size _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j, .-_Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j .globl _Z12sim_severityPfPjS_j .type _Z12sim_severityPfPjS_j, @function _Z12sim_severityPfPjS_j: .LFB2583: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2583: .size _Z12sim_severityPfPjS_j, .-_Z12sim_severityPfPjS_j .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Running %u simulations ...\n" .LC8: .string "Usage: %s [nSimulations].\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC9: .string "\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .align 8 .LC10: .string "\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .text .globl main .type main, @function main: .LFB2554: .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 $200, %rsp .cfi_def_cfa_offset 224 movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax cmpl $2, %edi jne .L33 movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movl %eax, %ecx movl %eax, N_SIMS(%rip) movl %eax, %edx shrl $9, %edx addl $1, %edx movl %edx, N_BLK(%rip) movl $512, %edx cmpl %edx, %eax cmovbe %eax, %edx movl %edx, N_THRD(%rip) sall $2, %eax movl %eax, N_BYTES_F(%rip) movl %eax, N_BYTES_I(%rip) movl %ecx, %edx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 24(%rsp), %rdi call cudaStreamCreate@PLT leaq 32(%rsp), %rdi call cudaStreamCreate@PLT movl $0x42480000, 160(%rsp) movl $0x40400000, 164(%rsp) movl $0x3f000000, 168(%rsp) movl $0x3fc00000, 172(%rsp) movl $0x3f4ccccd, 176(%rsp) movl $0x40a00000, 180(%rsp) movl N_BYTES_PRM(%rip), %esi leaq 160(%rsp), %rbp movl $0, %edx movq %rbp, %rdi call cudaHostRegister@PLT movl N_BYTES_PRM(%rip), %esi leaq 172(%rsp), %rbx movl $0, %edx movq %rbx, %rdi call cudaHostRegister@PLT movl N_BYTES_PRM(%rip), %esi leaq 40(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_PRM(%rip), %esi leaq 48(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_PRM(%rip), %edx movq 24(%rsp), %r8 movl $1, %ecx movq %rbp, %rsi movq 40(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_PRM(%rip), %edx movq 32(%rsp), %r8 movl $1, %ecx movq %rbx, %rsi movq 48(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_I(%rip), %esi leaq 64(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_F(%rip), %esi leaq 96(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_I(%rip), %esi leaq 80(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_F(%rip), %esi leaq 112(%rsp), %rdi call cudaMalloc@PLT movl N_BYTES_I(%rip), %esi leaq 56(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_F(%rip), %esi leaq 88(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_I(%rip), %esi leaq 72(%rsp), %rdi call cudaMallocHost@PLT movl N_BYTES_F(%rip), %esi leaq 104(%rsp), %rdi call cudaMallocHost@PLT call _Z8get_timev movq %rax, %rbp leaq 120(%rsp), %rdi call cudaEventCreate@PLT leaq 128(%rsp), %rdi call cudaEventCreate@PLT movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 24(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L47 .L34: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 32(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L48 .L37: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 24(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L49 .L38: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movq 32(%rsp), %r9 movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L50 .L39: movl N_BYTES_F(%rip), %edx movq 24(%rsp), %r8 movl $2, %ecx movq 96(%rsp), %rsi movq 88(%rsp), %rdi call cudaMemcpyAsync@PLT movq 24(%rsp), %rsi movq 120(%rsp), %rdi call cudaEventRecord@PLT movl N_BYTES_F(%rip), %edx movq 32(%rsp), %r8 movl $2, %ecx movq 112(%rsp), %rsi movq 104(%rsp), %rdi call cudaMemcpyAsync@PLT movq 32(%rsp), %rsi movq 128(%rsp), %rdi call cudaEventRecord@PLT movl N_BYTES_I(%rip), %edx movq 24(%rsp), %r8 movl $2, %ecx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpyAsync@PLT movl N_BYTES_I(%rip), %edx movq 32(%rsp), %r8 movl $2, %ecx movq 80(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpyAsync@PLT movq 120(%rsp), %rdi call cudaEventSynchronize@PLT movl N_SIMS(%rip), %esi movq 88(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 8(%rsp) movq 128(%rsp), %rdi call cudaEventSynchronize@PLT movl N_SIMS(%rip), %esi movq 104(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 12(%rsp) movq 24(%rsp), %rdi call cudaStreamSynchronize@PLT movq 32(%rsp), %rdi call cudaStreamSynchronize@PLT call _Z8get_timev movq %rax, %rbx movq %rax, %rdi call cudaEventSynchronize@PLT leaq 20(%rsp), %rdi movq %rbx, %rdx movq %rbp, %rsi call cudaEventElapsedTime@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 pxor %xmm2, %xmm2 cvtss2sd 20(%rsp), %xmm2 pxor %xmm1, %xmm1 cvtss2sd 12(%rsp), %xmm1 leaq .LC9(%rip), %rsi movl $2, %edi movl $3, %eax call __printf_chk@PLT call _Z8get_timev movq %rax, %rbp movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L51 .L40: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L52 .L41: movl N_BYTES_I(%rip), %edx movl $2, %ecx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpy@PLT movl N_BYTES_F(%rip), %edx movl $2, %ecx movq 96(%rsp), %rsi movq 88(%rsp), %rdi call cudaMemcpy@PLT movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L53 .L42: movl N_THRD(%rip), %eax movl %eax, 148(%rsp) movl $1, 152(%rsp) movl $1, 156(%rsp) movl N_BLK(%rip), %eax movl %eax, 136(%rsp) movl $1, 140(%rsp) movl $1, 144(%rsp) movl $0, %r9d movl $0, %r8d movq 148(%rsp), %rdx movl $1, %ecx movq 136(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L54 .L43: movl N_BYTES_I(%rip), %edx movl $2, %ecx movq 80(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpy@PLT movl N_BYTES_F(%rip), %edx movl $2, %ecx movq 112(%rsp), %rsi movq 104(%rsp), %rdi call cudaMemcpy@PLT movl N_SIMS(%rip), %ebx movl %ebx, %esi movq 88(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 8(%rsp) movl %ebx, %esi movq 104(%rsp), %rdi call _Z8calcMeanPfj movss %xmm0, 12(%rsp) call _Z8get_timev movq %rax, %rbx movq %rax, %rdi call cudaEventSynchronize@PLT leaq 20(%rsp), %rdi movq %rbx, %rdx movq %rbp, %rsi call cudaEventElapsedTime@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 pxor %xmm2, %xmm2 cvtss2sd 20(%rsp), %xmm2 pxor %xmm1, %xmm1 cvtss2sd 12(%rsp), %xmm1 leaq .LC10(%rip), %rsi movl $2, %edi movl $3, %eax call __printf_chk@PLT movl $0, %eax jmp .L32 .L33: movq (%rsi), %rdx leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %eax .L32: movq 184(%rsp), %rdx subq %fs:40, %rdx jne .L55 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L47: .cfi_restore_state movl N_SIMS(%rip), %edx movq 40(%rsp), %rsi movq 64(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L34 .L48: movl N_SIMS(%rip), %edx movq 48(%rsp), %rsi movq 80(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L37 .L49: movl N_SIMS(%rip), %ecx movq 40(%rsp), %rdx movq 64(%rsp), %rsi movq 96(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L38 .L50: movl N_SIMS(%rip), %ecx movq 48(%rsp), %rdx movq 80(%rsp), %rsi movq 112(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L39 .L51: movl N_SIMS(%rip), %edx movq 40(%rsp), %rsi movq 64(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L40 .L52: movl N_SIMS(%rip), %ecx movq 40(%rsp), %rdx movq 64(%rsp), %rsi movq 96(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L41 .L53: movl N_SIMS(%rip), %edx movq 48(%rsp), %rsi movq 80(%rsp), %rdi call _Z30__device_stub__Z8sim_freqPjPfjPjPfj jmp .L42 .L54: movl N_SIMS(%rip), %ecx movq 48(%rsp), %rdx movq 80(%rsp), %rsi movq 112(%rsp), %rdi call _Z37__device_stub__Z12sim_severityPfPjS_jPfPjS_j jmp .L43 .L55: call __stack_chk_fail@PLT .cfi_endproc .LFE2554: .size main, .-main .section .rodata.str1.1 .LC11: .string "_Z12sim_severityPfPjS_j" .LC12: .string "_Z8sim_freqPjPfj" .LC13: .string "precalc_xorwow_matrix" .LC14: .string "precalc_xorwow_offset_matrix" .LC15: .string "mrg32k3aM1" .LC16: .string "mrg32k3aM2" .LC17: .string "mrg32k3aM1SubSeq" .LC18: .string "mrg32k3aM2SubSeq" .LC19: .string "mrg32k3aM1Seq" .LC20: .string "mrg32k3aM2Seq" .LC21: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2585: .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 .LC11(%rip), %rdx movq %rdx, %rcx leaq _Z12sim_severityPfPjS_j(%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 .LC12(%rip), %rdx movq %rdx, %rcx leaq _Z8sim_freqPjPfj(%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 .LC13(%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 .LC14(%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 .LC15(%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 .LC16(%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 .LC17(%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 .LC18(%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 .LC19(%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 .LC20(%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 .LC21(%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 .LFE2585: .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 N_BYTES_F .bss .align 4 .type N_BYTES_F, @object .size N_BYTES_F, 4 N_BYTES_F: .zero 4 .globl N_BYTES_I .align 4 .type N_BYTES_I, @object .size N_BYTES_I, 4 N_BYTES_I: .zero 4 .globl N_THRD .align 4 .type N_THRD, @object .size N_THRD, 4 N_THRD: .zero 4 .globl N_BLK .align 4 .type N_BLK, @object .size N_BLK, 4 N_BLK: .zero 4 .globl N_SIMS .align 4 .type N_SIMS, @object .size N_SIMS, 4 N_SIMS: .zero 4 .globl N_BYTES_PRM .data .align 4 .type N_BYTES_PRM, @object .size N_BYTES_PRM, 4 N_BYTES_PRM: .long 12 .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 "stream_registers.hip" .globl _Z8calcMeanPfj # -- Begin function _Z8calcMeanPfj .p2align 4, 0x90 .type _Z8calcMeanPfj,@function _Z8calcMeanPfj: # @_Z8calcMeanPfj .cfi_startproc # %bb.0: testl %esi, %esi je .LBB0_1 # %bb.4: # %.lr.ph movl %esi, %eax cvtsi2ss %rax, %xmm1 xorpd %xmm0, %xmm0 xorl %ecx, %ecx .p2align 4, 0x90 .LBB0_5: # =>This Inner Loop Header: Depth=1 movss (%rdi,%rcx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm1, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm0 incq %rcx cmpq %rcx, %rax jne .LBB0_5 # %bb.2: # %._crit_edge.loopexit cvtsd2ss %xmm0, %xmm0 retq .LBB0_1: xorps %xmm0, %xmm0 retq .Lfunc_end0: .size _Z8calcMeanPfj, .Lfunc_end0-_Z8calcMeanPfj .cfi_endproc # -- End function .globl _Z8get_timev # -- Begin function _Z8get_timev .p2align 4, 0x90 .type _Z8get_timev,@function _Z8get_timev: # @_Z8get_timev .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movq %rsp, %rdi callq hipEventCreate movq (%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq (%rsp), %rax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z8get_timev, .Lfunc_end1-_Z8get_timev .cfi_endproc # -- End function .globl _Z23__device_stub__sim_freqPjPfj # -- Begin function _Z23__device_stub__sim_freqPjPfj .p2align 4, 0x90 .type _Z23__device_stub__sim_freqPjPfj,@function _Z23__device_stub__sim_freqPjPfj: # @_Z23__device_stub__sim_freqPjPfj .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%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 $_Z8sim_freqPjPfj, %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_end2: .size _Z23__device_stub__sim_freqPjPfj, .Lfunc_end2-_Z23__device_stub__sim_freqPjPfj .cfi_endproc # -- End function .globl _Z27__device_stub__sim_severityPfPjS_j # -- Begin function _Z27__device_stub__sim_severityPfPjS_j .p2align 4, 0x90 .type _Z27__device_stub__sim_severityPfPjS_j,@function _Z27__device_stub__sim_severityPfPjS_j: # @_Z27__device_stub__sim_severityPfPjS_j .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%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 4(%rsp), %rax movq %rax, 104(%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 $_Z12sim_severityPfPjS_j, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end3: .size _Z27__device_stub__sim_severityPfPjS_j, .Lfunc_end3-_Z27__device_stub__sim_severityPfPjS_j .cfi_endproc # -- End function .globl _Z6asynchv # -- Begin function _Z6asynchv .p2align 4, 0x90 .type _Z6asynchv,@function _Z6asynchv: # @_Z6asynchv .cfi_startproc # %bb.0: retq .Lfunc_end4: .size _Z6asynchv, .Lfunc_end4-_Z6asynchv .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 $272, %rsp # imm = 0x110 .cfi_def_cfa_offset 304 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 cmpl $2, %edi jne .LBB5_35 # %bb.1: movabsq $4294967296, %r15 # imm = 0x100000000 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movl %eax, N_SIMS(%rip) movl %eax, %ecx shrl $9, %ecx incl %ecx movl %ecx, N_BLK(%rip) cmpl $512, %eax # imm = 0x200 movl $512, %ecx # imm = 0x200 cmovbl %eax, %ecx movl %ecx, N_THRD(%rip) leal (,%rax,4), %ecx movl %ecx, N_BYTES_F(%rip) movl %ecx, N_BYTES_I(%rip) movl $.L.str.1, %edi movl %eax, %esi xorl %eax, %eax callq printf leaq 120(%rsp), %rdi callq hipStreamCreate leaq 112(%rsp), %rdi callq hipStreamCreate movabsq $4629700418048884736, %rax # imm = 0x4040000042480000 movq %rax, 260(%rsp) movl $1056964608, 268(%rsp) # imm = 0x3F000000 movabsq $4561245704529379328, %rax # imm = 0x3F4CCCCD3FC00000 movq %rax, 248(%rsp) movl $1084227584, 256(%rsp) # imm = 0x40A00000 movl N_BYTES_PRM(%rip), %esi leaq 260(%rsp), %r14 movq %r14, %rdi xorl %edx, %edx callq hipHostRegister movl N_BYTES_PRM(%rip), %esi leaq 248(%rsp), %rbx movq %rbx, %rdi xorl %edx, %edx callq hipHostRegister movl N_BYTES_PRM(%rip), %esi leaq 160(%rsp), %rdi callq hipMalloc movl N_BYTES_PRM(%rip), %esi leaq 152(%rsp), %rdi callq hipMalloc movq 160(%rsp), %rdi movl N_BYTES_PRM(%rip), %edx movq 120(%rsp), %r8 movq %r14, %rsi movl $1, %ecx callq hipMemcpyAsync movq 152(%rsp), %rdi movl N_BYTES_PRM(%rip), %edx movq 112(%rsp), %r8 movq %rbx, %rsi movl $1, %ecx callq hipMemcpyAsync movl N_BYTES_I(%rip), %esi leaq 144(%rsp), %rdi callq hipMalloc movl N_BYTES_F(%rip), %esi leaq 192(%rsp), %rdi callq hipMalloc movl N_BYTES_I(%rip), %esi leaq 136(%rsp), %rdi callq hipMalloc movl N_BYTES_F(%rip), %esi leaq 176(%rsp), %rdi callq hipMalloc movl N_BYTES_I(%rip), %esi leaq 240(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_F(%rip), %esi leaq 200(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_I(%rip), %esi leaq 232(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc movl N_BYTES_F(%rip), %esi leaq 184(%rsp), %rdi xorl %edx, %edx callq hipHostMalloc leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %rbx leaq 224(%rsp), %rdi callq hipEventCreate leaq 216(%rsp), %rdi callq hipEventCreate movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 120(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_3 # %bb.2: movq 144(%rsp), %rax movq 160(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_3: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 112(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_5 # %bb.4: movq 136(%rsp), %rax movq 152(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_5: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 120(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_7 # %bb.6: movq 192(%rsp), %rax movq 144(%rsp), %rcx movq 160(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_7: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx movq 112(%rsp), %r9 orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_9 # %bb.8: movq 176(%rsp), %rax movq 136(%rsp), %rcx movq 152(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_9: movq 200(%rsp), %rdi movq 192(%rsp), %rsi movl N_BYTES_F(%rip), %edx movq 120(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 224(%rsp), %rdi movq 120(%rsp), %rsi callq hipEventRecord movq 184(%rsp), %rdi movq 176(%rsp), %rsi movl N_BYTES_F(%rip), %edx movq 112(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 216(%rsp), %rdi movq 112(%rsp), %rsi callq hipEventRecord movq 240(%rsp), %rdi movq 144(%rsp), %rsi movl N_BYTES_I(%rip), %edx movq 120(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 232(%rsp), %rdi movq 136(%rsp), %rsi movl N_BYTES_I(%rip), %edx movq 112(%rsp), %r8 movl $2, %ecx callq hipMemcpyAsync movq 224(%rsp), %rdi callq hipEventSynchronize movl N_SIMS(%rip), %eax xorps %xmm0, %xmm0 movsd %xmm0, 168(%rsp) # 8-byte Spill testq %rax, %rax je .LBB5_13 # %bb.10: # %.lr.ph.i xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 200(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_11: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_11 # %bb.12: # %._crit_edge.loopexit.i xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 .LBB5_13: # %_Z8calcMeanPfj.exit movsd %xmm0, 208(%rsp) # 8-byte Spill movq 216(%rsp), %rdi callq hipEventSynchronize movl N_SIMS(%rip), %eax testq %rax, %rax je .LBB5_17 # %bb.14: # %.lr.ph.i69 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 184(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_15: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_15 # %bb.16: # %._crit_edge.loopexit.i75 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 movsd %xmm0, 168(%rsp) # 8-byte Spill .LBB5_17: # %_Z8calcMeanPfj.exit77 movq 120(%rsp), %rdi callq hipStreamSynchronize movq 112(%rsp), %rdi callq hipStreamSynchronize leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %r14 movq %r14, %rdi callq hipEventSynchronize leaq 132(%rsp), %rdi movq %rbx, %rsi movq %r14, %rdx callq hipEventElapsedTime movss 132(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm0, %xmm2 movl $.L.str.2, %edi movsd 208(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movsd 168(%rsp), %xmm1 # 8-byte Reload # xmm1 = mem[0],zero movb $3, %al callq printf leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %rbx movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_19 # %bb.18: movq 144(%rsp), %rax movq 160(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_19: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_21 # %bb.20: movq 192(%rsp), %rax movq 144(%rsp), %rcx movq 160(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_21: movq 240(%rsp), %rdi movq 144(%rsp), %rsi movl N_BYTES_I(%rip), %edx movl $2, %ecx callq hipMemcpy movq 200(%rsp), %rdi movq 192(%rsp), %rsi movl N_BYTES_F(%rip), %edx movl $2, %ecx callq hipMemcpy movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_23 # %bb.22: movq 136(%rsp), %rax movq 152(%rsp), %rcx movl N_SIMS(%rip), %edx movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movl %edx, (%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 48(%rsp), %rdx leaq 8(%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 $_Z8sim_freqPjPfj, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB5_23: movl N_BLK(%rip), %edi movl N_THRD(%rip), %edx orq %r15, %rdi orq %r15, %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB5_25 # %bb.24: movq 176(%rsp), %rax movq 136(%rsp), %rcx movq 152(%rsp), %rdx movl N_SIMS(%rip), %esi movq %rax, 64(%rsp) movq %rcx, 56(%rsp) movq %rdx, 48(%rsp) movl %esi, 76(%rsp) leaq 64(%rsp), %rax movq %rax, 80(%rsp) leaq 56(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 76(%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 $_Z12sim_severityPfPjS_j, %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 .LBB5_25: movq 232(%rsp), %rdi movq 136(%rsp), %rsi movl N_BYTES_I(%rip), %edx movl $2, %ecx callq hipMemcpy movq 184(%rsp), %rdi movq 176(%rsp), %rsi movl N_BYTES_F(%rip), %edx movl $2, %ecx callq hipMemcpy movl N_SIMS(%rip), %eax xorps %xmm3, %xmm3 xorps %xmm0, %xmm0 testq %rax, %rax je .LBB5_29 # %bb.26: # %.lr.ph.i103 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 200(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_27: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_27 # %bb.28: # %._crit_edge.loopexit.i109 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 cvtss2sd %xmm0, %xmm0 .LBB5_29: # %_Z8calcMeanPfj.exit111 movsd %xmm0, 168(%rsp) # 8-byte Spill testq %rax, %rax je .LBB5_33 # %bb.30: # %.lr.ph.i113 xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 movq 184(%rsp), %rcx xorpd %xmm1, %xmm1 xorl %edx, %edx .p2align 4, 0x90 .LBB5_31: # =>This Inner Loop Header: Depth=1 movss (%rcx,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero divss %xmm0, %xmm2 cvtss2sd %xmm2, %xmm2 addsd %xmm2, %xmm1 incq %rdx cmpq %rdx, %rax jne .LBB5_31 # %bb.32: # %._crit_edge.loopexit.i119 xorps %xmm0, %xmm0 cvtsd2ss %xmm1, %xmm0 xorps %xmm3, %xmm3 cvtss2sd %xmm0, %xmm3 .LBB5_33: # %_Z8calcMeanPfj.exit121 movsd %xmm3, 208(%rsp) # 8-byte Spill leaq 80(%rsp), %rdi callq hipEventCreate movq 80(%rsp), %rdi xorl %r14d, %r14d xorl %esi, %esi callq hipEventRecord movq 80(%rsp), %r15 movq %r15, %rdi callq hipEventSynchronize leaq 132(%rsp), %rdi movq %rbx, %rsi movq %r15, %rdx callq hipEventElapsedTime movss 132(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm0, %xmm2 movl $.L.str.3, %edi movsd 168(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero movsd 208(%rsp), %xmm1 # 8-byte Reload # xmm1 = mem[0],zero movb $3, %al callq printf jmp .LBB5_34 .LBB5_35: movq (%rsi), %rsi movl $.L.str, %edi xorl %eax, %eax callq printf movl $1, %r14d .LBB5_34: movl %r14d, %eax addq $272, %rsp # imm = 0x110 .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_end5: .size main, .Lfunc_end5-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 .LBB6_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB6_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8sim_freqPjPfj, %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 $_Z12sim_severityPfPjS_j, %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_end6: .size __hip_module_ctor, .Lfunc_end6-__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 .LBB7_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 .LBB7_2: retq .Lfunc_end7: .size __hip_module_dtor, .Lfunc_end7-__hip_module_dtor .cfi_endproc # -- End function .type N_BYTES_PRM,@object # @N_BYTES_PRM .data .globl N_BYTES_PRM .p2align 2, 0x0 N_BYTES_PRM: .long 12 # 0xc .size N_BYTES_PRM, 4 .type N_SIMS,@object # @N_SIMS .bss .globl N_SIMS .p2align 2, 0x0 N_SIMS: .long 0 # 0x0 .size N_SIMS, 4 .type N_BLK,@object # @N_BLK .globl N_BLK .p2align 2, 0x0 N_BLK: .long 0 # 0x0 .size N_BLK, 4 .type N_THRD,@object # @N_THRD .globl N_THRD .p2align 2, 0x0 N_THRD: .long 0 # 0x0 .size N_THRD, 4 .type N_BYTES_I,@object # @N_BYTES_I .globl N_BYTES_I .p2align 2, 0x0 N_BYTES_I: .long 0 # 0x0 .size N_BYTES_I, 4 .type N_BYTES_F,@object # @N_BYTES_F .globl N_BYTES_F .p2align 2, 0x0 N_BYTES_F: .long 0 # 0x0 .size N_BYTES_F, 4 .type _Z8sim_freqPjPfj,@object # @_Z8sim_freqPjPfj .section .rodata,"a",@progbits .globl _Z8sim_freqPjPfj .p2align 3, 0x0 _Z8sim_freqPjPfj: .quad _Z23__device_stub__sim_freqPjPfj .size _Z8sim_freqPjPfj, 8 .type _Z12sim_severityPfPjS_j,@object # @_Z12sim_severityPfPjS_j .globl _Z12sim_severityPfPjS_j .p2align 3, 0x0 _Z12sim_severityPfPjS_j: .quad _Z27__device_stub__sim_severityPfPjS_j .size _Z12sim_severityPfPjS_j, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Usage: %s [nSimulations].\n" .size .L.str, 27 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Running %u simulations ...\n" .size .L.str.1, 28 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "\tasynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .size .L.str.2, 59 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "\tsynchronously:\t loss1=%.3f, loss2=%.3f, %.3f ms taken, \n" .size .L.str.3, 58 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z8sim_freqPjPfj" .size .L__unnamed_1, 17 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z12sim_severityPfPjS_j" .size .L__unnamed_2, 24 .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 _Z23__device_stub__sim_freqPjPfj .addrsig_sym _Z27__device_stub__sim_severityPfPjS_j .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8sim_freqPjPfj .addrsig_sym _Z12sim_severityPfPjS_j .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.
#define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
code for sm_80 Function : _Z14select_matchesPKjPKijji .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*/ ULDC UR4, c[0x0][0x4] ; /* 0x0000010000047ab9 */ /* 0x000fe40000000800 */ /*0020*/ USHF.R.U32.HI UR4, URZ, 0x1, UR4 ; /* 0x000000013f047899 */ /* 0x000fcc0008011604 */ /*0030*/ ISETP.NE.AND P0, PT, RZ, UR4, PT ; /* 0x00000004ff007c0c */ /* 0x000fda000bf05270 */ /*0040*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0050*/ S2R R7, SR_TID.Y ; /* 0x0000000000077919 */ /* 0x000e220000002200 */ /*0060*/ IMAD.U32 R0, RZ, RZ, UR4 ; /* 0x00000004ff007e24 */ /* 0x000fc6000f8e00ff */ /*0070*/ S2R R2, SR_TID.X ; /* 0x0000000000027919 */ /* 0x000e240000002100 */ /*0080*/ IMAD R2, R2, c[0x0][0x4], R7 ; /* 0x0000010002027a24 */ /* 0x001fc800078e0207 */ /*0090*/ IMAD.SHL.U32 R3, R2, 0x4, RZ ; /* 0x0000000402037824 */ /* 0x000fe400078e00ff */ /*00a0*/ ISETP.GE.U32.AND P0, PT, R7, R0, PT ; /* 0x000000000700720c */ /* 0x000fda0003f06070 */ /*00b0*/ @P0 BRA 0x150 ; /* 0x0000009000000947 */ /* 0x001fea0003800000 */ /*00c0*/ IMAD R6, R0, 0x4, R3 ; /* 0x0000000400067824 */ /* 0x000fe200078e0203 */ /*00d0*/ LDS R4, [R2.X4] ; /* 0x0000000002047984 */ /* 0x000fe20000004800 */ /*00e0*/ WARPSYNC 0xffffffff ; /* 0xffffffff00007948 */ /* 0x000fe60003800000 */ /*00f0*/ LDS R9, [R6] ; /* 0x0000000006097984 */ /* 0x000e240000000800 */ /*0100*/ ISETP.GE.AND P0, PT, R9, R4, PT ; /* 0x000000040900720c */ /* 0x001fda0003f06270 */ /*0110*/ @!P0 STS [R2.X4], R9 ; /* 0x0000000902008388 */ /* 0x000fe80000004800 */ /*0120*/ @!P0 LDS R5, [R6+0x400] ; /* 0x0004000006058984 */ /* 0x000e280000000800 */ /*0130*/ @!P0 STS [R2.X4+0x400], R5 ; /* 0x0004000502008388 */ /* 0x0011e80000004800 */ /*0140*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0150*/ SHF.R.U32.HI R0, RZ, 0x1, R0 ; /* 0x00000001ff007819 */ /* 0x000fc80000011600 */ /*0160*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f05270 */ /*0170*/ @P0 BRA 0xa0 ; /* 0xffffff2000000947 */ /* 0x000fea000383ffff */ /*0180*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0190*/ BRA 0x190; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
.file "tmpxft_0014f26e_00000000-6_select_matches.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__Z14select_matchesPKjPKijjiPKjPKijji .type _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji, @function _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%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 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%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 _Z14select_matchesPKjPKijji(%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__Z14select_matchesPKjPKijjiPKjPKijji, .-_Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji .globl _Z14select_matchesPKjPKijji .type _Z14select_matchesPKjPKijji, @function _Z14select_matchesPKjPKijji: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z14select_matchesPKjPKijji, .-_Z14select_matchesPKjPKijji .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14select_matchesPKjPKijji" .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 _Z14select_matchesPKjPKijji(%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.
#define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
#include <hip/hip_runtime.h> #define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14select_matchesPKjPKijji .globl _Z14select_matchesPKjPKijji .p2align 8 .type _Z14select_matchesPKjPKijji,@function _Z14select_matchesPKjPKijji: s_load_b32 s0, s[0:1], 0x2c s_waitcnt lgkmcnt(0) s_lshr_b32 s0, s0, 16 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_lt_u32 s0, 2 s_cbranch_scc1 .LBB0_7 v_bfe_u32 v1, v0, 10, 10 v_and_b32_e32 v0, 0x3ff, v0 s_and_b32 s0, 0xffff, s0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_mad_u32_u24 v0, v0, s0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b32_e32 v2, 2, v0 v_add_nc_u32_e32 v3, 0x400, v2 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_4 .p2align 6 .LBB0_2: s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_3: s_or_b32 exec_lo, exec_lo, s2 s_cmp_lt_u32 s0, 4 s_mov_b32 s0, s1 s_cbranch_scc1 .LBB0_7 .LBB0_4: s_lshr_b32 s1, s0, 1 s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e64 s1, v1 s_cbranch_execz .LBB0_3 v_add_nc_u32_e32 v4, s1, v0 s_delay_alu instid0(VALU_DEP_1) v_lshlrev_b32_e32 v5, 2, v4 ds_load_b32 v4, v5 offset:1024 ds_load_b32 v6, v3 s_waitcnt lgkmcnt(0) v_cmp_lt_i32_e32 vcc_lo, v4, v6 s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_2 ds_load_b32 v5, v5 ds_store_b32 v3, v4 s_waitcnt lgkmcnt(1) ds_store_b32 v2, v5 s_branch .LBB0_2 .LBB0_7: s_set_inst_prefetch_distance 0x2 s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14select_matchesPKjPKijji .amdhsa_group_segment_fixed_size 2048 .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 1 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 3 .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 _Z14select_matchesPKjPKijji, .Lfunc_end0-_Z14select_matchesPKjPKijji .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: by_value - .offset: 20 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .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: 2048 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14select_matchesPKjPKijji .private_segment_fixed_size: 0 .sgpr_count: 5 .sgpr_spill_count: 0 .symbol: _Z14select_matchesPKjPKijji.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> #define THREADS 256 __global__ void select_matches( const unsigned* in_idx, const int* in_dist, const unsigned nfeat, const unsigned nelem, const int max_dist) { unsigned f = blockIdx.x * blockDim.x + threadIdx.x; unsigned sid = threadIdx.x * blockDim.y + threadIdx.y; __shared__ int s_dist[THREADS]; __shared__ unsigned s_idx[THREADS]; // Reduce best matches and find the best of them all for (unsigned i = blockDim.y / 2; i > 0; i >>= 1) { if (threadIdx.y < i) { int dist = s_dist[sid + i]; if (dist < s_dist[sid]) { s_dist[sid] = dist; s_idx[sid] = s_idx[sid + i]; } __syncthreads(); } } }
.text .file "select_matches.hip" .globl _Z29__device_stub__select_matchesPKjPKijji # -- Begin function _Z29__device_stub__select_matchesPKjPKijji .p2align 4, 0x90 .type _Z29__device_stub__select_matchesPKjPKijji,@function _Z29__device_stub__select_matchesPKjPKijji: # @_Z29__device_stub__select_matchesPKjPKijji .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%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 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 $_Z14select_matchesPKjPKijji, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z29__device_stub__select_matchesPKjPKijji, .Lfunc_end0-_Z29__device_stub__select_matchesPKjPKijji .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 $_Z14select_matchesPKjPKijji, %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 _Z14select_matchesPKjPKijji,@object # @_Z14select_matchesPKjPKijji .section .rodata,"a",@progbits .globl _Z14select_matchesPKjPKijji .p2align 3, 0x0 _Z14select_matchesPKjPKijji: .quad _Z29__device_stub__select_matchesPKjPKijji .size _Z14select_matchesPKjPKijji, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14select_matchesPKjPKijji" .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 _Z29__device_stub__select_matchesPKjPKijji .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14select_matchesPKjPKijji .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 : _Z14select_matchesPKjPKijji .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*/ ULDC UR4, c[0x0][0x4] ; /* 0x0000010000047ab9 */ /* 0x000fe40000000800 */ /*0020*/ USHF.R.U32.HI UR4, URZ, 0x1, UR4 ; /* 0x000000013f047899 */ /* 0x000fcc0008011604 */ /*0030*/ ISETP.NE.AND P0, PT, RZ, UR4, PT ; /* 0x00000004ff007c0c */ /* 0x000fda000bf05270 */ /*0040*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0050*/ S2R R7, SR_TID.Y ; /* 0x0000000000077919 */ /* 0x000e220000002200 */ /*0060*/ IMAD.U32 R0, RZ, RZ, UR4 ; /* 0x00000004ff007e24 */ /* 0x000fc6000f8e00ff */ /*0070*/ S2R R2, SR_TID.X ; /* 0x0000000000027919 */ /* 0x000e240000002100 */ /*0080*/ IMAD R2, R2, c[0x0][0x4], R7 ; /* 0x0000010002027a24 */ /* 0x001fc800078e0207 */ /*0090*/ IMAD.SHL.U32 R3, R2, 0x4, RZ ; /* 0x0000000402037824 */ /* 0x000fe400078e00ff */ /*00a0*/ ISETP.GE.U32.AND P0, PT, R7, R0, PT ; /* 0x000000000700720c */ /* 0x000fda0003f06070 */ /*00b0*/ @P0 BRA 0x150 ; /* 0x0000009000000947 */ /* 0x001fea0003800000 */ /*00c0*/ IMAD R6, R0, 0x4, R3 ; /* 0x0000000400067824 */ /* 0x000fe200078e0203 */ /*00d0*/ LDS R4, [R2.X4] ; /* 0x0000000002047984 */ /* 0x000fe20000004800 */ /*00e0*/ WARPSYNC 0xffffffff ; /* 0xffffffff00007948 */ /* 0x000fe60003800000 */ /*00f0*/ LDS R9, [R6] ; /* 0x0000000006097984 */ /* 0x000e240000000800 */ /*0100*/ ISETP.GE.AND P0, PT, R9, R4, PT ; /* 0x000000040900720c */ /* 0x001fda0003f06270 */ /*0110*/ @!P0 STS [R2.X4], R9 ; /* 0x0000000902008388 */ /* 0x000fe80000004800 */ /*0120*/ @!P0 LDS R5, [R6+0x400] ; /* 0x0004000006058984 */ /* 0x000e280000000800 */ /*0130*/ @!P0 STS [R2.X4+0x400], R5 ; /* 0x0004000502008388 */ /* 0x0011e80000004800 */ /*0140*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0150*/ SHF.R.U32.HI R0, RZ, 0x1, R0 ; /* 0x00000001ff007819 */ /* 0x000fc80000011600 */ /*0160*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f05270 */ /*0170*/ @P0 BRA 0xa0 ; /* 0xffffff2000000947 */ /* 0x000fea000383ffff */ /*0180*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0190*/ BRA 0x190; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14select_matchesPKjPKijji .globl _Z14select_matchesPKjPKijji .p2align 8 .type _Z14select_matchesPKjPKijji,@function _Z14select_matchesPKjPKijji: s_load_b32 s0, s[0:1], 0x2c s_waitcnt lgkmcnt(0) s_lshr_b32 s0, s0, 16 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_lt_u32 s0, 2 s_cbranch_scc1 .LBB0_7 v_bfe_u32 v1, v0, 10, 10 v_and_b32_e32 v0, 0x3ff, v0 s_and_b32 s0, 0xffff, s0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_mad_u32_u24 v0, v0, s0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b32_e32 v2, 2, v0 v_add_nc_u32_e32 v3, 0x400, v2 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_4 .p2align 6 .LBB0_2: s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_3: s_or_b32 exec_lo, exec_lo, s2 s_cmp_lt_u32 s0, 4 s_mov_b32 s0, s1 s_cbranch_scc1 .LBB0_7 .LBB0_4: s_lshr_b32 s1, s0, 1 s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e64 s1, v1 s_cbranch_execz .LBB0_3 v_add_nc_u32_e32 v4, s1, v0 s_delay_alu instid0(VALU_DEP_1) v_lshlrev_b32_e32 v5, 2, v4 ds_load_b32 v4, v5 offset:1024 ds_load_b32 v6, v3 s_waitcnt lgkmcnt(0) v_cmp_lt_i32_e32 vcc_lo, v4, v6 s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_2 ds_load_b32 v5, v5 ds_store_b32 v3, v4 s_waitcnt lgkmcnt(1) ds_store_b32 v2, v5 s_branch .LBB0_2 .LBB0_7: s_set_inst_prefetch_distance 0x2 s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14select_matchesPKjPKijji .amdhsa_group_segment_fixed_size 2048 .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 1 .amdhsa_next_free_vgpr 7 .amdhsa_next_free_sgpr 3 .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 _Z14select_matchesPKjPKijji, .Lfunc_end0-_Z14select_matchesPKjPKijji .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: by_value - .offset: 20 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .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: 2048 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14select_matchesPKjPKijji .private_segment_fixed_size: 0 .sgpr_count: 5 .sgpr_spill_count: 0 .symbol: _Z14select_matchesPKjPKijji.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_0014f26e_00000000-6_select_matches.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__Z14select_matchesPKjPKijjiPKjPKijji .type _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji, @function _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%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 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%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 _Z14select_matchesPKjPKijji(%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__Z14select_matchesPKjPKijjiPKjPKijji, .-_Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji .globl _Z14select_matchesPKjPKijji .type _Z14select_matchesPKjPKijji, @function _Z14select_matchesPKjPKijji: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z14select_matchesPKjPKijjiPKjPKijji addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z14select_matchesPKjPKijji, .-_Z14select_matchesPKjPKijji .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14select_matchesPKjPKijji" .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 _Z14select_matchesPKjPKijji(%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 "select_matches.hip" .globl _Z29__device_stub__select_matchesPKjPKijji # -- Begin function _Z29__device_stub__select_matchesPKjPKijji .p2align 4, 0x90 .type _Z29__device_stub__select_matchesPKjPKijji,@function _Z29__device_stub__select_matchesPKjPKijji: # @_Z29__device_stub__select_matchesPKjPKijji .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%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 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 $_Z14select_matchesPKjPKijji, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z29__device_stub__select_matchesPKjPKijji, .Lfunc_end0-_Z29__device_stub__select_matchesPKjPKijji .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 $_Z14select_matchesPKjPKijji, %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 _Z14select_matchesPKjPKijji,@object # @_Z14select_matchesPKjPKijji .section .rodata,"a",@progbits .globl _Z14select_matchesPKjPKijji .p2align 3, 0x0 _Z14select_matchesPKjPKijji: .quad _Z29__device_stub__select_matchesPKjPKijji .size _Z14select_matchesPKjPKijji, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14select_matchesPKjPKijji" .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 _Z29__device_stub__select_matchesPKjPKijji .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14select_matchesPKjPKijji .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.
// Copyright (C) 2018 NVIDIA CORPORATION. All rights reserved. // // Licensed 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. // ============================================================================= #define TO_NEXT_MULT_P2(x,p) (((x)+((p)-1)) & ~(p-1)) __host__ __device__ ulonglong2 operator<<(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(0ull, a.x << (l-64)); } else { b = make_ulonglong2(a.x << l, (a.y << l) | (a.x >> (8*sizeof(a.x)-l))); } return b; } __host__ __device__ ulonglong2 operator>>(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(a.y >> (l-64), 0ull); } else { b = make_ulonglong2((a.x >> l) | (a.y << (8*sizeof(a.y)-l)), a.y >> l); } return b; } __host__ __device__ ulonglong2 operator|(ulonglong2 a, ulonglong2 b) { return make_ulonglong2(a.x | b.x, a.y | b.y); } template<int BDIM_X, int MAXIOB, int SH_BYTE_X_BL, typename LDST_T> __device__ void memcpy_d(const size_t n, const unsigned char *__restrict__ src, unsigned char *__restrict__ dst, unsigned char *__restrict__ __sh) { const int tid = threadIdx.x; const unsigned long long srcULL = reinterpret_cast<unsigned long long>(src); const unsigned long long dstULL = reinterpret_cast<unsigned long long>(dst); int srcOff = (MAXIOB - srcULL) & (MAXIOB-1); int dstOff = (MAXIOB - dstULL) & (MAXIOB-1); const int ELXTH = SH_BYTE_X_BL/(BDIM_X*MAXIOB); LDST_T *__ptrSH = reinterpret_cast<LDST_T *>(__sh); if (srcOff == dstOff) { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = (n-srcOff) / sizeof(*__ptrLDG); int remBytes = (n-srcOff) % sizeof(*__ptrLDG); LDST_T __loc[ELXTH]; #pragma unroll for(int j = 0; j < ELXTH; j++) { if (j*BDIM_X+tid < nread) { __loc[j] = __ptrLDG[j*BDIM_X+tid]; } } for(int i = 0; i < nread; i += BDIM_X*ELXTH) { #pragma unroll for(int j = 0; j < ELXTH; j++) { __ptrSH[j*BDIM_X+tid] = __loc[j]; } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + BDIM_X*ELXTH + j*BDIM_X + tid < nread) { __loc[j] = __ptrLDG[i + BDIM_X*ELXTH + j*BDIM_X + tid]; } } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + j*BDIM_X + tid < nread) { __ptrSTG[i + j*BDIM_X + tid] = __ptrSH[j*BDIM_X+tid]; } } } if (tid < srcOff+remBytes) { const int off = (tid < srcOff) ? tid : n-remBytes+tid-srcOff; dst[off] = src[off]; } } else { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = ((n-srcOff) / sizeof(*__ptrLDG)); int remBytes = ((n-srcOff) % sizeof(*__ptrLDG)); int lowShft, uppShft; if (srcOff > dstOff) { uppShft = (srcOff-dstOff)*8; lowShft = (8*sizeof(*__ptrLDG)) - uppShft; __ptrSTG++; } else { lowShft = (dstOff-srcOff)*8; uppShft = (8*sizeof(*__ptrLDG)) - lowShft; } for(int i = 0; i < nread-1; i += BDIM_X) { if (i+tid < nread-1) { const LDST_T low = __ptrLDG[i+tid]; const LDST_T upp = __ptrLDG[i+tid+1]; __ptrSTG[i+tid] = (low >> lowShft) | (upp << uppShft); } } remBytes += sizeof(*__ptrLDG); if (srcOff > dstOff) { dstOff += sizeof(*__ptrLDG); if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } else { if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } } } template<int BDIM_X, int MAXIOB> __global__ void memcpy_k(const size_t *sizes, const unsigned char *const __restrict__ *__restrict__ in, unsigned char *__restrict__ *__restrict__ out) { const int SH_BYTE_X_BL = 32768; __shared__ unsigned char __sh[SH_BYTE_X_BL]; switch(MAXIOB) { case 4: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned int>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 8: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned long long>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 16: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, ulonglong2>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; } return; } #define NTHREADS 1024 void batched_d2d_memcpy(void** out_ptrs, void** in_ptrs, size_t* sizes, int num_copies, cudaStream_t stream) { memcpy_k<NTHREADS, 16><<<num_copies, NTHREADS, 0, stream>>>(sizes, (unsigned char**) in_ptrs, (unsigned char**) out_ptrs); }
.file "tmpxft_001837f3_00000000-6_batched_memcpy.cudafe1.cpp" .text #APP #NO_APP .type _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh, @function _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh: .LFB2058: .cfi_startproc subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 96(%rsp) movq %rsi, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) movq %rdx, 24(%rsp) leaq 24(%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 .L5 .L1: movq 120(%rsp), %rax subq %fs:40, %rax jne .L6 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L5: .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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L1 .L6: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh, .-_ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh .section .text._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .type _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, @function _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: .LFB2107: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2107: .size _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, .-_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .text .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2036: .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 .LFE2036: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Zls10ulonglong2i .type _Zls10ulonglong2i, @function _Zls10ulonglong2i: .LFB2027: .cfi_startproc endbr64 cmpl $64, %edx jle .L12 leal -64(%rdx), %ecx salq %cl, %rdi movq %rdi, %rsi movl $0, %eax .L13: movq %rsi, %rdx ret .L12: movl %edx, %ecx salq %cl, %rsi movl $64, %ecx subl %edx, %ecx movq %rdi, %rax shrq %cl, %rax orq %rax, %rsi movq %rdi, %rax movl %edx, %ecx salq %cl, %rax jmp .L13 .cfi_endproc .LFE2027: .size _Zls10ulonglong2i, .-_Zls10ulonglong2i .globl _Zrs10ulonglong2i .type _Zrs10ulonglong2i, @function _Zrs10ulonglong2i: .LFB2028: .cfi_startproc endbr64 cmpl $64, %edx jle .L15 leal -64(%rdx), %ecx movq %rsi, %rax shrq %cl, %rax movl $0, %r8d .L16: movq %r8, %rdx ret .L15: movq %rsi, %r8 movl %edx, %ecx shrq %cl, %r8 shrq %cl, %rdi movl $64, %ecx subl %edx, %ecx salq %cl, %rsi movq %rdi, %rax orq %rsi, %rax jmp .L16 .cfi_endproc .LFE2028: .size _Zrs10ulonglong2i, .-_Zrs10ulonglong2i .globl _Zor10ulonglong2S_ .type _Zor10ulonglong2S_, @function _Zor10ulonglong2S_: .LFB2029: .cfi_startproc endbr64 movq %rdx, %rax orq %rcx, %rsi movq %rsi, %rdx orq %rdi, %rax ret .cfi_endproc .LFE2029: .size _Zor10ulonglong2S_, .-_Zor10ulonglong2S_ .globl _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st .type _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st, @function _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st: .LFB2033: .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 $32, %rsp .cfi_def_cfa_offset 64 movq %rdi, %r12 movq %rsi, %rbp movq %rdx, %rbx movl $1024, 20(%rsp) movl $1, 24(%rsp) movl %ecx, 8(%rsp) movl $1, 12(%rsp) movq %r8, %r9 movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L21 .L18: addq $32, %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 .L21: .cfi_restore_state movq %r12, %rdx movq %rbp, %rsi movq %rbx, %rdi call _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh jmp .L18 .cfi_endproc .LFE2033: .size _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st, .-_Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2061: .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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh(%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 .LFE2061: .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.
// Copyright (C) 2018 NVIDIA CORPORATION. All rights reserved. // // Licensed 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. // ============================================================================= #define TO_NEXT_MULT_P2(x,p) (((x)+((p)-1)) & ~(p-1)) __host__ __device__ ulonglong2 operator<<(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(0ull, a.x << (l-64)); } else { b = make_ulonglong2(a.x << l, (a.y << l) | (a.x >> (8*sizeof(a.x)-l))); } return b; } __host__ __device__ ulonglong2 operator>>(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(a.y >> (l-64), 0ull); } else { b = make_ulonglong2((a.x >> l) | (a.y << (8*sizeof(a.y)-l)), a.y >> l); } return b; } __host__ __device__ ulonglong2 operator|(ulonglong2 a, ulonglong2 b) { return make_ulonglong2(a.x | b.x, a.y | b.y); } template<int BDIM_X, int MAXIOB, int SH_BYTE_X_BL, typename LDST_T> __device__ void memcpy_d(const size_t n, const unsigned char *__restrict__ src, unsigned char *__restrict__ dst, unsigned char *__restrict__ __sh) { const int tid = threadIdx.x; const unsigned long long srcULL = reinterpret_cast<unsigned long long>(src); const unsigned long long dstULL = reinterpret_cast<unsigned long long>(dst); int srcOff = (MAXIOB - srcULL) & (MAXIOB-1); int dstOff = (MAXIOB - dstULL) & (MAXIOB-1); const int ELXTH = SH_BYTE_X_BL/(BDIM_X*MAXIOB); LDST_T *__ptrSH = reinterpret_cast<LDST_T *>(__sh); if (srcOff == dstOff) { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = (n-srcOff) / sizeof(*__ptrLDG); int remBytes = (n-srcOff) % sizeof(*__ptrLDG); LDST_T __loc[ELXTH]; #pragma unroll for(int j = 0; j < ELXTH; j++) { if (j*BDIM_X+tid < nread) { __loc[j] = __ptrLDG[j*BDIM_X+tid]; } } for(int i = 0; i < nread; i += BDIM_X*ELXTH) { #pragma unroll for(int j = 0; j < ELXTH; j++) { __ptrSH[j*BDIM_X+tid] = __loc[j]; } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + BDIM_X*ELXTH + j*BDIM_X + tid < nread) { __loc[j] = __ptrLDG[i + BDIM_X*ELXTH + j*BDIM_X + tid]; } } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + j*BDIM_X + tid < nread) { __ptrSTG[i + j*BDIM_X + tid] = __ptrSH[j*BDIM_X+tid]; } } } if (tid < srcOff+remBytes) { const int off = (tid < srcOff) ? tid : n-remBytes+tid-srcOff; dst[off] = src[off]; } } else { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = ((n-srcOff) / sizeof(*__ptrLDG)); int remBytes = ((n-srcOff) % sizeof(*__ptrLDG)); int lowShft, uppShft; if (srcOff > dstOff) { uppShft = (srcOff-dstOff)*8; lowShft = (8*sizeof(*__ptrLDG)) - uppShft; __ptrSTG++; } else { lowShft = (dstOff-srcOff)*8; uppShft = (8*sizeof(*__ptrLDG)) - lowShft; } for(int i = 0; i < nread-1; i += BDIM_X) { if (i+tid < nread-1) { const LDST_T low = __ptrLDG[i+tid]; const LDST_T upp = __ptrLDG[i+tid+1]; __ptrSTG[i+tid] = (low >> lowShft) | (upp << uppShft); } } remBytes += sizeof(*__ptrLDG); if (srcOff > dstOff) { dstOff += sizeof(*__ptrLDG); if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } else { if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } } } template<int BDIM_X, int MAXIOB> __global__ void memcpy_k(const size_t *sizes, const unsigned char *const __restrict__ *__restrict__ in, unsigned char *__restrict__ *__restrict__ out) { const int SH_BYTE_X_BL = 32768; __shared__ unsigned char __sh[SH_BYTE_X_BL]; switch(MAXIOB) { case 4: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned int>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 8: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned long long>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 16: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, ulonglong2>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; } return; } #define NTHREADS 1024 void batched_d2d_memcpy(void** out_ptrs, void** in_ptrs, size_t* sizes, int num_copies, cudaStream_t stream) { memcpy_k<NTHREADS, 16><<<num_copies, NTHREADS, 0, stream>>>(sizes, (unsigned char**) in_ptrs, (unsigned char**) out_ptrs); }
#include <hip/hip_runtime.h> // Copyright (C) 2018 NVIDIA CORPORATION. All rights reserved. // // Licensed 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. // ============================================================================= #define TO_NEXT_MULT_P2(x,p) (((x)+((p)-1)) & ~(p-1)) __host__ __device__ ulonglong2 operator<<(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(0ull, a.x << (l-64)); } else { b = make_ulonglong2(a.x << l, (a.y << l) | (a.x >> (8*sizeof(a.x)-l))); } return b; } __host__ __device__ ulonglong2 operator>>(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(a.y >> (l-64), 0ull); } else { b = make_ulonglong2((a.x >> l) | (a.y << (8*sizeof(a.y)-l)), a.y >> l); } return b; } __host__ __device__ ulonglong2 operator|(ulonglong2 a, ulonglong2 b) { return make_ulonglong2(a.x | b.x, a.y | b.y); } template<int BDIM_X, int MAXIOB, int SH_BYTE_X_BL, typename LDST_T> __device__ void memcpy_d(const size_t n, const unsigned char *__restrict__ src, unsigned char *__restrict__ dst, unsigned char *__restrict__ __sh) { const int tid = threadIdx.x; const unsigned long long srcULL = reinterpret_cast<unsigned long long>(src); const unsigned long long dstULL = reinterpret_cast<unsigned long long>(dst); int srcOff = (MAXIOB - srcULL) & (MAXIOB-1); int dstOff = (MAXIOB - dstULL) & (MAXIOB-1); const int ELXTH = SH_BYTE_X_BL/(BDIM_X*MAXIOB); LDST_T *__ptrSH = reinterpret_cast<LDST_T *>(__sh); if (srcOff == dstOff) { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = (n-srcOff) / sizeof(*__ptrLDG); int remBytes = (n-srcOff) % sizeof(*__ptrLDG); LDST_T __loc[ELXTH]; #pragma unroll for(int j = 0; j < ELXTH; j++) { if (j*BDIM_X+tid < nread) { __loc[j] = __ptrLDG[j*BDIM_X+tid]; } } for(int i = 0; i < nread; i += BDIM_X*ELXTH) { #pragma unroll for(int j = 0; j < ELXTH; j++) { __ptrSH[j*BDIM_X+tid] = __loc[j]; } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + BDIM_X*ELXTH + j*BDIM_X + tid < nread) { __loc[j] = __ptrLDG[i + BDIM_X*ELXTH + j*BDIM_X + tid]; } } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + j*BDIM_X + tid < nread) { __ptrSTG[i + j*BDIM_X + tid] = __ptrSH[j*BDIM_X+tid]; } } } if (tid < srcOff+remBytes) { const int off = (tid < srcOff) ? tid : n-remBytes+tid-srcOff; dst[off] = src[off]; } } else { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = ((n-srcOff) / sizeof(*__ptrLDG)); int remBytes = ((n-srcOff) % sizeof(*__ptrLDG)); int lowShft, uppShft; if (srcOff > dstOff) { uppShft = (srcOff-dstOff)*8; lowShft = (8*sizeof(*__ptrLDG)) - uppShft; __ptrSTG++; } else { lowShft = (dstOff-srcOff)*8; uppShft = (8*sizeof(*__ptrLDG)) - lowShft; } for(int i = 0; i < nread-1; i += BDIM_X) { if (i+tid < nread-1) { const LDST_T low = __ptrLDG[i+tid]; const LDST_T upp = __ptrLDG[i+tid+1]; __ptrSTG[i+tid] = (low >> lowShft) | (upp << uppShft); } } remBytes += sizeof(*__ptrLDG); if (srcOff > dstOff) { dstOff += sizeof(*__ptrLDG); if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } else { if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } } } template<int BDIM_X, int MAXIOB> __global__ void memcpy_k(const size_t *sizes, const unsigned char *const __restrict__ *__restrict__ in, unsigned char *__restrict__ *__restrict__ out) { const int SH_BYTE_X_BL = 32768; __shared__ unsigned char __sh[SH_BYTE_X_BL]; switch(MAXIOB) { case 4: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned int>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 8: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned long long>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 16: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, ulonglong2>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; } return; } #define NTHREADS 1024 void batched_d2d_memcpy(void** out_ptrs, void** in_ptrs, size_t* sizes, int num_copies, hipStream_t stream) { memcpy_k<NTHREADS, 16><<<num_copies, NTHREADS, 0, stream>>>(sizes, (unsigned char**) in_ptrs, (unsigned char**) out_ptrs); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> // Copyright (C) 2018 NVIDIA CORPORATION. All rights reserved. // // Licensed 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. // ============================================================================= #define TO_NEXT_MULT_P2(x,p) (((x)+((p)-1)) & ~(p-1)) __host__ __device__ ulonglong2 operator<<(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(0ull, a.x << (l-64)); } else { b = make_ulonglong2(a.x << l, (a.y << l) | (a.x >> (8*sizeof(a.x)-l))); } return b; } __host__ __device__ ulonglong2 operator>>(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(a.y >> (l-64), 0ull); } else { b = make_ulonglong2((a.x >> l) | (a.y << (8*sizeof(a.y)-l)), a.y >> l); } return b; } __host__ __device__ ulonglong2 operator|(ulonglong2 a, ulonglong2 b) { return make_ulonglong2(a.x | b.x, a.y | b.y); } template<int BDIM_X, int MAXIOB, int SH_BYTE_X_BL, typename LDST_T> __device__ void memcpy_d(const size_t n, const unsigned char *__restrict__ src, unsigned char *__restrict__ dst, unsigned char *__restrict__ __sh) { const int tid = threadIdx.x; const unsigned long long srcULL = reinterpret_cast<unsigned long long>(src); const unsigned long long dstULL = reinterpret_cast<unsigned long long>(dst); int srcOff = (MAXIOB - srcULL) & (MAXIOB-1); int dstOff = (MAXIOB - dstULL) & (MAXIOB-1); const int ELXTH = SH_BYTE_X_BL/(BDIM_X*MAXIOB); LDST_T *__ptrSH = reinterpret_cast<LDST_T *>(__sh); if (srcOff == dstOff) { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = (n-srcOff) / sizeof(*__ptrLDG); int remBytes = (n-srcOff) % sizeof(*__ptrLDG); LDST_T __loc[ELXTH]; #pragma unroll for(int j = 0; j < ELXTH; j++) { if (j*BDIM_X+tid < nread) { __loc[j] = __ptrLDG[j*BDIM_X+tid]; } } for(int i = 0; i < nread; i += BDIM_X*ELXTH) { #pragma unroll for(int j = 0; j < ELXTH; j++) { __ptrSH[j*BDIM_X+tid] = __loc[j]; } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + BDIM_X*ELXTH + j*BDIM_X + tid < nread) { __loc[j] = __ptrLDG[i + BDIM_X*ELXTH + j*BDIM_X + tid]; } } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + j*BDIM_X + tid < nread) { __ptrSTG[i + j*BDIM_X + tid] = __ptrSH[j*BDIM_X+tid]; } } } if (tid < srcOff+remBytes) { const int off = (tid < srcOff) ? tid : n-remBytes+tid-srcOff; dst[off] = src[off]; } } else { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = ((n-srcOff) / sizeof(*__ptrLDG)); int remBytes = ((n-srcOff) % sizeof(*__ptrLDG)); int lowShft, uppShft; if (srcOff > dstOff) { uppShft = (srcOff-dstOff)*8; lowShft = (8*sizeof(*__ptrLDG)) - uppShft; __ptrSTG++; } else { lowShft = (dstOff-srcOff)*8; uppShft = (8*sizeof(*__ptrLDG)) - lowShft; } for(int i = 0; i < nread-1; i += BDIM_X) { if (i+tid < nread-1) { const LDST_T low = __ptrLDG[i+tid]; const LDST_T upp = __ptrLDG[i+tid+1]; __ptrSTG[i+tid] = (low >> lowShft) | (upp << uppShft); } } remBytes += sizeof(*__ptrLDG); if (srcOff > dstOff) { dstOff += sizeof(*__ptrLDG); if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } else { if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } } } template<int BDIM_X, int MAXIOB> __global__ void memcpy_k(const size_t *sizes, const unsigned char *const __restrict__ *__restrict__ in, unsigned char *__restrict__ *__restrict__ out) { const int SH_BYTE_X_BL = 32768; __shared__ unsigned char __sh[SH_BYTE_X_BL]; switch(MAXIOB) { case 4: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned int>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 8: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned long long>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 16: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, ulonglong2>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; } return; } #define NTHREADS 1024 void batched_d2d_memcpy(void** out_ptrs, void** in_ptrs, size_t* sizes, int num_copies, hipStream_t stream) { memcpy_k<NTHREADS, 16><<<num_copies, NTHREADS, 0, stream>>>(sizes, (unsigned char**) in_ptrs, (unsigned char**) out_ptrs); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .section .text._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .protected _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .globl _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .p2align 8 .type _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,@function _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: s_clause 0x1 s_load_b128 s[4:7], s[2:3], 0x0 s_load_b64 s[2:3], s[2:3], 0x10 s_mov_b32 s8, s15 s_mov_b32 s9, 0 v_and_b32_e32 v15, 0x3ff, v0 s_lshl_b64 s[8:9], s[8:9], 3 s_waitcnt lgkmcnt(0) s_add_u32 s10, s4, s8 s_addc_u32 s11, s5, s9 s_add_u32 s4, s6, s8 s_addc_u32 s5, s7, s9 s_add_u32 s2, s2, s8 s_addc_u32 s3, s3, s9 s_load_b64 s[6:7], s[4:5], 0x0 s_load_b64 s[4:5], s[2:3], 0x0 s_load_b64 s[8:9], s[10:11], 0x0 s_waitcnt lgkmcnt(0) s_sub_i32 s2, 0, s6 s_sub_i32 s10, 0, s4 s_and_b32 s3, s2, 15 s_and_b32 s10, s10, 15 s_mov_b32 s2, -1 s_cmp_lg_u32 s10, s3 s_cbranch_scc0 .LBB0_27 s_add_u32 s2, s4, s10 s_addc_u32 s21, s5, 0 s_cmp_ge_u32 s10, s3 s_cselect_b32 s11, -1, 0 s_delay_alu instid0(SALU_CYCLE_1) s_and_b32 vcc_lo, exec_lo, s11 s_cbranch_vccz .LBB0_3 s_sub_i32 s12, s10, s3 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_lshl_b32 s13, s12, 3 s_sub_i32 s12, 0x80, s13 s_cbranch_execz .LBB0_4 s_branch .LBB0_5 .LBB0_3: .LBB0_4: s_sub_i32 s12, s3, s10 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_lshl_b32 s12, s12, 3 s_sub_i32 s13, 0x80, s12 s_add_u32 s2, s2, 16 s_addc_u32 s21, s21, 0 .LBB0_5: s_sub_u32 s14, s8, s3 s_subb_u32 s15, s9, 0 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_alignbit_b32 v1, s15, s14, 4 v_cmp_gt_i32_e32 vcc_lo, 2, v1 s_cbranch_vccnz .LBB0_18 v_add_nc_u32_e32 v16, -1, v1 v_lshlrev_b32_e32 v1, 4, v15 s_cmpk_lt_i32 s13, 0x41 s_cselect_b32 s15, -1, 0 s_sub_i32 s16, 64, s13 s_sub_i32 s17, s13, 64 s_cmpk_lt_i32 s12, 0x41 v_add_co_u32 v9, s2, s2, v1 s_cselect_b32 s18, -1, 0 s_sub_i32 s19, 64, s12 s_sub_i32 s20, s12, 64 v_add_co_ci_u32_e64 v10, null, s21, 0, s2 s_add_u32 s2, s6, s3 s_addc_u32 s21, s7, 0 v_add_co_u32 v11, s2, s2, v1 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v12, null, s21, 0, s2 s_mov_b32 s21, 0 s_branch .LBB0_9 .LBB0_7: s_waitcnt vmcnt(0) lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) v_or_b32_e32 v2, v8, v14 v_or_b32_e32 v1, v7, v13 v_or_b32_e32 v4, v4, v6 v_or_b32_e32 v3, v3, v5 flat_store_b128 v[9:10], v[1:4] .LBB0_8: s_or_b32 exec_lo, exec_lo, s2 v_add_co_u32 v9, vcc_lo, v9, 0x4000 s_addk_i32 s21, 0x400 v_add_co_ci_u32_e32 v10, vcc_lo, 0, v10, vcc_lo v_cmp_ge_i32_e32 vcc_lo, s21, v16 v_add_co_u32 v11, s2, v11, 0x4000 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v12, s2, 0, v12, s2 s_cbranch_vccnz .LBB0_18 .LBB0_9: v_add_nc_u32_e32 v1, s21, v15 s_mov_b32 s2, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_lt_i32_e64 v1, v16 s_cbranch_execz .LBB0_8 s_clause 0x1 flat_load_b128 v[5:8], v[11:12] flat_load_b128 v[1:4], v[11:12] offset:16 s_and_b32 vcc_lo, exec_lo, s15 s_cbranch_vccz .LBB0_12 s_waitcnt vmcnt(1) lgkmcnt(1) v_lshrrev_b64 v[13:14], s13, v[5:6] v_lshlrev_b64 v[17:18], s16, v[7:8] v_lshrrev_b64 v[5:6], s13, v[7:8] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_3) v_or_b32_e32 v14, v14, v18 v_or_b32_e32 v13, v13, v17 s_cbranch_execz .LBB0_13 s_branch .LBB0_14 .LBB0_12: .LBB0_13: s_waitcnt vmcnt(1) lgkmcnt(1) v_lshrrev_b64 v[13:14], s17, v[7:8] v_mov_b32_e32 v5, 0 v_mov_b32_e32 v6, 0 .LBB0_14: s_and_b32 vcc_lo, exec_lo, s18 s_cbranch_vccz .LBB0_16 s_waitcnt vmcnt(0) lgkmcnt(0) v_lshlrev_b64 v[3:4], s12, v[3:4] v_lshrrev_b64 v[17:18], s19, v[1:2] v_lshlrev_b64 v[7:8], s12, v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_3) v_or_b32_e32 v4, v4, v18 v_or_b32_e32 v3, v3, v17 s_cbranch_execnz .LBB0_7 s_branch .LBB0_17 .LBB0_16: .LBB0_17: s_waitcnt vmcnt(0) lgkmcnt(0) v_lshlrev_b64 v[3:4], s20, v[1:2] v_mov_b32_e32 v7, 0 v_mov_b32_e32 v8, 0 s_branch .LBB0_7 .LBB0_18: s_and_b32 s2, s14, 15 s_and_b32 vcc_lo, exec_lo, s11 s_or_b32 s2, s2, 16 s_mov_b32 s11, -1 s_cbranch_vccz .LBB0_22 s_add_i32 s12, s2, s10 s_mov_b32 s11, exec_lo v_cmpx_gt_u32_e64 s12, v15 s_cbranch_execz .LBB0_21 v_add_nc_u32_e32 v1, s8, v15 v_cmp_gt_u32_e32 vcc_lo, s10, v15 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_subrev_nc_u32_e32 v1, s12, v1 v_cndmask_b32_e32 v3, v1, v15, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v4, 31, v3 v_add_co_u32 v1, vcc_lo, s6, v3 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v4, vcc_lo flat_load_u8 v5, v[1:2] v_add_co_u32 v1, vcc_lo, s4, v3 v_add_co_ci_u32_e32 v2, vcc_lo, s5, v4, vcc_lo s_waitcnt vmcnt(0) lgkmcnt(0) flat_store_b8 v[1:2], v5 .LBB0_21: s_or_b32 exec_lo, exec_lo, s11 s_mov_b32 s11, 0 .LBB0_22: s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 vcc_lo, exec_lo, s11 s_cbranch_vccnz .LBB0_26 s_or_b32 s11, s10, 16 s_delay_alu instid0(SALU_CYCLE_1) s_add_i32 s12, s2, s11 s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e64 s12, v15 s_cbranch_execz .LBB0_25 v_add_nc_u32_e32 v1, s8, v15 v_cmp_gt_u32_e32 vcc_lo, s11, v15 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_subrev_nc_u32_e32 v1, s12, v1 v_cndmask_b32_e32 v3, v1, v15, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v4, 31, v3 v_add_co_u32 v1, vcc_lo, s6, v3 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v4, vcc_lo flat_load_u8 v5, v[1:2] v_add_co_u32 v1, vcc_lo, s4, v3 v_add_co_ci_u32_e32 v2, vcc_lo, s5, v4, vcc_lo s_waitcnt vmcnt(0) lgkmcnt(0) flat_store_b8 v[1:2], v5 .LBB0_25: s_or_b32 exec_lo, exec_lo, s2 .LBB0_26: s_mov_b32 s2, 0 .LBB0_27: s_delay_alu instid0(SALU_CYCLE_1) s_and_b32 vcc_lo, exec_lo, s2 s_cbranch_vccz .LBB0_47 s_load_b64 s[0:1], s[0:1], 0x4 v_and_b32_e32 v1, 0x3ff, v0 v_bfe_u32 v2, v0, 10, 10 v_bfe_u32 v0, v0, 20, 10 s_waitcnt lgkmcnt(0) s_lshr_b32 s0, s0, 16 s_delay_alu instid0(VALU_DEP_2) v_mul_u32_u24_e32 v2, s1, v2 s_mul_i32 s0, s0, s1 s_add_u32 s2, s6, s3 v_mul_lo_u32 v1, s0, v1 s_addc_u32 s11, s7, 0 s_sub_u32 s0, s8, s3 s_subb_u32 s1, s9, 0 s_mov_b32 s9, exec_lo v_alignbit_b32 v3, s1, s0, 4 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add3_u32 v0, v1, v2, v0 v_readfirstlane_b32 s1, v3 s_delay_alu instid0(VALU_DEP_2) v_lshl_add_u32 v4, v0, 5, 0x8000 v_cmpx_lt_i32_e64 v15, v3 s_cbranch_execz .LBB0_30 v_lshlrev_b32_e32 v0, 4, v15 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v0, s12, s2, v0 v_add_co_ci_u32_e64 v1, null, s11, 0, s12 flat_load_b128 v[0:3], v[0:1] s_waitcnt vmcnt(0) lgkmcnt(0) ds_store_b128 v4, v[0:3] .LBB0_30: s_or_b32 exec_lo, exec_lo, s9 v_or_b32_e32 v0, 0x400, v15 s_mov_b32 s9, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s1, v0 s_cbranch_execz .LBB0_32 v_lshlrev_b32_e32 v1, 4, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v1, s2, s2, v1 v_add_co_ci_u32_e64 v2, null, s11, 0, s2 flat_load_b128 v[5:8], v[1:2] s_waitcnt vmcnt(0) lgkmcnt(0) ds_store_b128 v4, v[5:8] offset:16 .LBB0_32: s_or_b32 exec_lo, exec_lo, s9 s_cmp_lt_i32 s1, 1 s_cbranch_scc1 .LBB0_43 v_lshlrev_b32_e32 v5, 4, v15 s_add_u32 s2, s4, s10 s_addc_u32 s9, s5, 0 s_add_u32 s10, s6, s3 s_addc_u32 s11, s7, 0 v_add_co_u32 v2, s10, s10, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v3, null, s11, 0, s10 v_lshlrev_b32_e32 v7, 4, v0 v_add_co_u32 v0, s2, s2, v5 v_add_co_u32 v2, vcc_lo, v2, 0xc000 v_add_nc_u32_e32 v6, 16, v4 v_add_co_ci_u32_e64 v1, null, s9, 0, s2 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo s_mov_b32 s2, 0 s_branch .LBB0_35 .LBB0_34: s_or_b32 exec_lo, exec_lo, s9 v_add_co_u32 v0, vcc_lo, v0, 0x8000 v_add_co_ci_u32_e32 v1, vcc_lo, 0, v1, vcc_lo v_add_co_u32 v2, vcc_lo, v2, 0x8000 v_add_co_ci_u32_e32 v3, vcc_lo, 0, v3, vcc_lo s_addk_i32 s2, 0x800 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_ge_i32 s2, s1 s_cbranch_scc1 .LBB0_43 .LBB0_35: ds_load_b128 v[8:11], v4 ds_load_b128 v[16:19], v6 v_add3_u32 v12, v15, s2, 0x800 s_mov_b32 s9, exec_lo s_waitcnt lgkmcnt(1) ds_store_b128 v5, v[8:11] s_waitcnt lgkmcnt(1) ds_store_b128 v7, v[16:19] v_cmpx_gt_i32_e64 s1, v12 s_cbranch_execz .LBB0_37 v_add_co_u32 v8, vcc_lo, 0xffffc000, v2 v_add_co_ci_u32_e32 v9, vcc_lo, -1, v3, vcc_lo flat_load_b128 v[8:11], v[8:9] s_waitcnt vmcnt(0) lgkmcnt(0) ds_store_b128 v4, v[8:11] .LBB0_37: s_or_b32 exec_lo, exec_lo, s9 v_add3_u32 v8, v15, s2, 0xc00 s_mov_b32 s9, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s1, v8 s_cbranch_execz .LBB0_39 flat_load_b128 v[8:11], v[2:3] s_waitcnt vmcnt(0) lgkmcnt(0) ds_store_b128 v6, v[8:11] .LBB0_39: s_or_b32 exec_lo, exec_lo, s9 v_add_nc_u32_e32 v8, s2, v15 s_mov_b32 s9, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s1, v8 s_cbranch_execz .LBB0_41 ds_load_b128 v[8:11], v5 s_waitcnt lgkmcnt(0) flat_store_b128 v[0:1], v[8:11] .LBB0_41: s_or_b32 exec_lo, exec_lo, s9 v_add3_u32 v8, v15, s2, 0x400 s_mov_b32 s9, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s1, v8 s_cbranch_execz .LBB0_34 ds_load_b128 v[8:11], v7 v_add_co_u32 v12, vcc_lo, 0x4000, v0 v_add_co_ci_u32_e32 v13, vcc_lo, 0, v1, vcc_lo s_waitcnt lgkmcnt(0) flat_store_b128 v[12:13], v[8:11] s_branch .LBB0_34 .LBB0_43: s_and_b32 s1, s0, 15 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s1, s1, s3 v_cmp_gt_u32_e32 vcc_lo, s1, v15 s_and_saveexec_b32 s1, vcc_lo s_cbranch_execz .LBB0_47 s_mov_b32 s1, exec_lo v_cmpx_le_u32_e64 s3, v15 v_add_nc_u32_e32 v0, s8, v15 s_and_b32 s0, s0, 15 s_delay_alu instid0(SALU_CYCLE_1) s_add_i32 s0, s0, s3 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_subrev_nc_u32_e32 v15, s0, v0 s_or_b32 exec_lo, exec_lo, s1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v2, 31, v15 v_add_co_u32 v0, vcc_lo, s6, v15 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v2, vcc_lo flat_load_u8 v3, v[0:1] v_add_co_u32 v0, vcc_lo, s4, v15 v_add_co_ci_u32_e32 v1, vcc_lo, s5, v2, vcc_lo s_waitcnt vmcnt(0) lgkmcnt(0) flat_store_b8 v[0:1], v3 .LBB0_47: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .amdhsa_group_segment_fixed_size 65536 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 1 .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 2 .amdhsa_next_free_vgpr 20 .amdhsa_next_free_sgpr 22 .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 .section .text._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .Lfunc_end0: .size _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, .Lfunc_end0-_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .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 - .actual_access: read_only .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .actual_access: read_only .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 65536 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .private_segment_fixed_size: 0 .sgpr_count: 24 .sgpr_spill_count: 0 .symbol: _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 20 .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> // Copyright (C) 2018 NVIDIA CORPORATION. All rights reserved. // // Licensed 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. // ============================================================================= #define TO_NEXT_MULT_P2(x,p) (((x)+((p)-1)) & ~(p-1)) __host__ __device__ ulonglong2 operator<<(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(0ull, a.x << (l-64)); } else { b = make_ulonglong2(a.x << l, (a.y << l) | (a.x >> (8*sizeof(a.x)-l))); } return b; } __host__ __device__ ulonglong2 operator>>(ulonglong2 a, int l) { ulonglong2 b; if (l > 64) { b = make_ulonglong2(a.y >> (l-64), 0ull); } else { b = make_ulonglong2((a.x >> l) | (a.y << (8*sizeof(a.y)-l)), a.y >> l); } return b; } __host__ __device__ ulonglong2 operator|(ulonglong2 a, ulonglong2 b) { return make_ulonglong2(a.x | b.x, a.y | b.y); } template<int BDIM_X, int MAXIOB, int SH_BYTE_X_BL, typename LDST_T> __device__ void memcpy_d(const size_t n, const unsigned char *__restrict__ src, unsigned char *__restrict__ dst, unsigned char *__restrict__ __sh) { const int tid = threadIdx.x; const unsigned long long srcULL = reinterpret_cast<unsigned long long>(src); const unsigned long long dstULL = reinterpret_cast<unsigned long long>(dst); int srcOff = (MAXIOB - srcULL) & (MAXIOB-1); int dstOff = (MAXIOB - dstULL) & (MAXIOB-1); const int ELXTH = SH_BYTE_X_BL/(BDIM_X*MAXIOB); LDST_T *__ptrSH = reinterpret_cast<LDST_T *>(__sh); if (srcOff == dstOff) { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = (n-srcOff) / sizeof(*__ptrLDG); int remBytes = (n-srcOff) % sizeof(*__ptrLDG); LDST_T __loc[ELXTH]; #pragma unroll for(int j = 0; j < ELXTH; j++) { if (j*BDIM_X+tid < nread) { __loc[j] = __ptrLDG[j*BDIM_X+tid]; } } for(int i = 0; i < nread; i += BDIM_X*ELXTH) { #pragma unroll for(int j = 0; j < ELXTH; j++) { __ptrSH[j*BDIM_X+tid] = __loc[j]; } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + BDIM_X*ELXTH + j*BDIM_X + tid < nread) { __loc[j] = __ptrLDG[i + BDIM_X*ELXTH + j*BDIM_X + tid]; } } #pragma unroll for(int j = 0; j < ELXTH; j++) { if (i + j*BDIM_X + tid < nread) { __ptrSTG[i + j*BDIM_X + tid] = __ptrSH[j*BDIM_X+tid]; } } } if (tid < srcOff+remBytes) { const int off = (tid < srcOff) ? tid : n-remBytes+tid-srcOff; dst[off] = src[off]; } } else { const LDST_T *__restrict__ __ptrLDG = reinterpret_cast<const LDST_T *>(src + srcOff); LDST_T *__restrict__ __ptrSTG = reinterpret_cast< LDST_T *>(dst + dstOff); int nread = ((n-srcOff) / sizeof(*__ptrLDG)); int remBytes = ((n-srcOff) % sizeof(*__ptrLDG)); int lowShft, uppShft; if (srcOff > dstOff) { uppShft = (srcOff-dstOff)*8; lowShft = (8*sizeof(*__ptrLDG)) - uppShft; __ptrSTG++; } else { lowShft = (dstOff-srcOff)*8; uppShft = (8*sizeof(*__ptrLDG)) - lowShft; } for(int i = 0; i < nread-1; i += BDIM_X) { if (i+tid < nread-1) { const LDST_T low = __ptrLDG[i+tid]; const LDST_T upp = __ptrLDG[i+tid+1]; __ptrSTG[i+tid] = (low >> lowShft) | (upp << uppShft); } } remBytes += sizeof(*__ptrLDG); if (srcOff > dstOff) { dstOff += sizeof(*__ptrLDG); if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } else { if (tid < dstOff+remBytes) { const int off = (tid < dstOff) ? tid : n-remBytes + tid-dstOff; dst[off] = src[off]; } } } } template<int BDIM_X, int MAXIOB> __global__ void memcpy_k(const size_t *sizes, const unsigned char *const __restrict__ *__restrict__ in, unsigned char *__restrict__ *__restrict__ out) { const int SH_BYTE_X_BL = 32768; __shared__ unsigned char __sh[SH_BYTE_X_BL]; switch(MAXIOB) { case 4: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned int>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 8: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, unsigned long long>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; case 16: memcpy_d<BDIM_X, MAXIOB, SH_BYTE_X_BL, ulonglong2>(sizes[blockIdx.x], in[blockIdx.x], out[blockIdx.x], __sh); break; } return; } #define NTHREADS 1024 void batched_d2d_memcpy(void** out_ptrs, void** in_ptrs, size_t* sizes, int num_copies, hipStream_t stream) { memcpy_k<NTHREADS, 16><<<num_copies, NTHREADS, 0, stream>>>(sizes, (unsigned char**) in_ptrs, (unsigned char**) out_ptrs); }
.text .file "batched_memcpy.hip" .globl _Zls15HIP_vector_typeIyLj2EEi # -- Begin function _Zls15HIP_vector_typeIyLj2EEi .p2align 4, 0x90 .type _Zls15HIP_vector_typeIyLj2EEi,@function _Zls15HIP_vector_typeIyLj2EEi: # @_Zls15HIP_vector_typeIyLj2EEi .cfi_startproc # %bb.0: movl %edx, %ecx cmpl $65, %edx jl .LBB0_2 # %bb.1: # kill: def $cl killed $cl killed $ecx shlq %cl, %rdi xorl %eax, %eax movq %rdi, %rdx retq .LBB0_2: movq %rdi, %rax shlq %cl, %rax # kill: def $cl killed $cl killed $ecx shldq %cl, %rdi, %rsi movq %rsi, %rdx retq .Lfunc_end0: .size _Zls15HIP_vector_typeIyLj2EEi, .Lfunc_end0-_Zls15HIP_vector_typeIyLj2EEi .cfi_endproc # -- End function .globl _Zrs15HIP_vector_typeIyLj2EEi # -- Begin function _Zrs15HIP_vector_typeIyLj2EEi .p2align 4, 0x90 .type _Zrs15HIP_vector_typeIyLj2EEi,@function _Zrs15HIP_vector_typeIyLj2EEi: # @_Zrs15HIP_vector_typeIyLj2EEi .cfi_startproc # %bb.0: movl %edx, %ecx movq %rsi, %rax cmpl $65, %edx jl .LBB1_2 # %bb.1: # kill: def $cl killed $cl killed $ecx shrq %cl, %rax xorl %edx, %edx retq .LBB1_2: shrdq %cl, %rax, %rdi movq %rax, %rdx # kill: def $cl killed $cl killed $ecx shrq %cl, %rdx movq %rdi, %rax retq .Lfunc_end1: .size _Zrs15HIP_vector_typeIyLj2EEi, .Lfunc_end1-_Zrs15HIP_vector_typeIyLj2EEi .cfi_endproc # -- End function .globl _Zor15HIP_vector_typeIyLj2EES0_ # -- Begin function _Zor15HIP_vector_typeIyLj2EES0_ .p2align 4, 0x90 .type _Zor15HIP_vector_typeIyLj2EES0_,@function _Zor15HIP_vector_typeIyLj2EES0_: # @_Zor15HIP_vector_typeIyLj2EES0_ .cfi_startproc # %bb.0: movq %rdi, %rax orq %rdx, %rax orq %rcx, %rsi movq %rsi, %rdx retq .Lfunc_end2: .size _Zor15HIP_vector_typeIyLj2EES0_, .Lfunc_end2-_Zor15HIP_vector_typeIyLj2EES0_ .cfi_endproc # -- End function .globl _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t # -- Begin function _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .p2align 4, 0x90 .type _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t,@function _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t: # @_Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .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 $112, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %r8, %r9 movq %rdx, %r15 movq %rsi, %r14 movq %rdi, %rbx movl %ecx, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $1024, %rdx # imm = 0x400 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_2 # %bb.1: movq %r15, 72(%rsp) movq %r14, 64(%rsp) movq %rbx, 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 $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_2: addq $112, %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_end3: .size _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t, .Lfunc_end3-_Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .cfi_endproc # -- End function .section .text._Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh # -- Begin function _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .p2align 4, 0x90 .type _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,@function _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: # @_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .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 $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %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_end4: .size _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, .Lfunc_end4-_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .cfi_endproc # -- End function .text .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 .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %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_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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,@object # @_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .section .rodata._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"aG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .p2align 3, 0x0 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: .quad _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .size _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh" .size .L__unnamed_1, 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 _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .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_001837f3_00000000-6_batched_memcpy.cudafe1.cpp" .text #APP #NO_APP .type _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh, @function _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh: .LFB2058: .cfi_startproc subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 96(%rsp) movq %rsi, 16(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) movq %rdx, 24(%rsp) leaq 24(%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 .L5 .L1: movq 120(%rsp), %rax subq %fs:40, %rax jne .L6 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L5: .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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L1 .L6: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh, .-_ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh .section .text._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .type _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, @function _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: .LFB2107: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2107: .size _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, .-_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .text .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2036: .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 .LFE2036: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Zls10ulonglong2i .type _Zls10ulonglong2i, @function _Zls10ulonglong2i: .LFB2027: .cfi_startproc endbr64 cmpl $64, %edx jle .L12 leal -64(%rdx), %ecx salq %cl, %rdi movq %rdi, %rsi movl $0, %eax .L13: movq %rsi, %rdx ret .L12: movl %edx, %ecx salq %cl, %rsi movl $64, %ecx subl %edx, %ecx movq %rdi, %rax shrq %cl, %rax orq %rax, %rsi movq %rdi, %rax movl %edx, %ecx salq %cl, %rax jmp .L13 .cfi_endproc .LFE2027: .size _Zls10ulonglong2i, .-_Zls10ulonglong2i .globl _Zrs10ulonglong2i .type _Zrs10ulonglong2i, @function _Zrs10ulonglong2i: .LFB2028: .cfi_startproc endbr64 cmpl $64, %edx jle .L15 leal -64(%rdx), %ecx movq %rsi, %rax shrq %cl, %rax movl $0, %r8d .L16: movq %r8, %rdx ret .L15: movq %rsi, %r8 movl %edx, %ecx shrq %cl, %r8 shrq %cl, %rdi movl $64, %ecx subl %edx, %ecx salq %cl, %rsi movq %rdi, %rax orq %rsi, %rax jmp .L16 .cfi_endproc .LFE2028: .size _Zrs10ulonglong2i, .-_Zrs10ulonglong2i .globl _Zor10ulonglong2S_ .type _Zor10ulonglong2S_, @function _Zor10ulonglong2S_: .LFB2029: .cfi_startproc endbr64 movq %rdx, %rax orq %rcx, %rsi movq %rsi, %rdx orq %rdi, %rax ret .cfi_endproc .LFE2029: .size _Zor10ulonglong2S_, .-_Zor10ulonglong2S_ .globl _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st .type _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st, @function _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st: .LFB2033: .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 $32, %rsp .cfi_def_cfa_offset 64 movq %rdi, %r12 movq %rsi, %rbp movq %rdx, %rbx movl $1024, 20(%rsp) movl $1, 24(%rsp) movl %ecx, 8(%rsp) movl $1, 12(%rsp) movq %r8, %r9 movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L21 .L18: addq $32, %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 .L21: .cfi_restore_state movq %r12, %rdx movq %rbp, %rsi movq %rbx, %rdi call _ZL53__device_stub__Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPhPKmPrKPKhPrPh jmp .L18 .cfi_endproc .LFE2033: .size _Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st, .-_Z18batched_d2d_memcpyPPvS0_PmiP11CUstream_st .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2061: .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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh(%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 .LFE2061: .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 "batched_memcpy.hip" .globl _Zls15HIP_vector_typeIyLj2EEi # -- Begin function _Zls15HIP_vector_typeIyLj2EEi .p2align 4, 0x90 .type _Zls15HIP_vector_typeIyLj2EEi,@function _Zls15HIP_vector_typeIyLj2EEi: # @_Zls15HIP_vector_typeIyLj2EEi .cfi_startproc # %bb.0: movl %edx, %ecx cmpl $65, %edx jl .LBB0_2 # %bb.1: # kill: def $cl killed $cl killed $ecx shlq %cl, %rdi xorl %eax, %eax movq %rdi, %rdx retq .LBB0_2: movq %rdi, %rax shlq %cl, %rax # kill: def $cl killed $cl killed $ecx shldq %cl, %rdi, %rsi movq %rsi, %rdx retq .Lfunc_end0: .size _Zls15HIP_vector_typeIyLj2EEi, .Lfunc_end0-_Zls15HIP_vector_typeIyLj2EEi .cfi_endproc # -- End function .globl _Zrs15HIP_vector_typeIyLj2EEi # -- Begin function _Zrs15HIP_vector_typeIyLj2EEi .p2align 4, 0x90 .type _Zrs15HIP_vector_typeIyLj2EEi,@function _Zrs15HIP_vector_typeIyLj2EEi: # @_Zrs15HIP_vector_typeIyLj2EEi .cfi_startproc # %bb.0: movl %edx, %ecx movq %rsi, %rax cmpl $65, %edx jl .LBB1_2 # %bb.1: # kill: def $cl killed $cl killed $ecx shrq %cl, %rax xorl %edx, %edx retq .LBB1_2: shrdq %cl, %rax, %rdi movq %rax, %rdx # kill: def $cl killed $cl killed $ecx shrq %cl, %rdx movq %rdi, %rax retq .Lfunc_end1: .size _Zrs15HIP_vector_typeIyLj2EEi, .Lfunc_end1-_Zrs15HIP_vector_typeIyLj2EEi .cfi_endproc # -- End function .globl _Zor15HIP_vector_typeIyLj2EES0_ # -- Begin function _Zor15HIP_vector_typeIyLj2EES0_ .p2align 4, 0x90 .type _Zor15HIP_vector_typeIyLj2EES0_,@function _Zor15HIP_vector_typeIyLj2EES0_: # @_Zor15HIP_vector_typeIyLj2EES0_ .cfi_startproc # %bb.0: movq %rdi, %rax orq %rdx, %rax orq %rcx, %rsi movq %rsi, %rdx retq .Lfunc_end2: .size _Zor15HIP_vector_typeIyLj2EES0_, .Lfunc_end2-_Zor15HIP_vector_typeIyLj2EES0_ .cfi_endproc # -- End function .globl _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t # -- Begin function _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .p2align 4, 0x90 .type _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t,@function _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t: # @_Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .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 $112, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %r8, %r9 movq %rdx, %r15 movq %rsi, %r14 movq %rdi, %rbx movl %ecx, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $1024, %rdx # imm = 0x400 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_2 # %bb.1: movq %r15, 72(%rsp) movq %r14, 64(%rsp) movq %rbx, 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 $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_2: addq $112, %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_end3: .size _Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t, .Lfunc_end3-_Z18batched_d2d_memcpyPPvS0_PmiP12ihipStream_t .cfi_endproc # -- End function .section .text._Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"axG",@progbits,_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh # -- Begin function _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .p2align 4, 0x90 .type _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,@function _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: # @_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .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 $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %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_end4: .size _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, .Lfunc_end4-_Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .cfi_endproc # -- End function .text .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 .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, %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_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 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,@object # @_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .section .rodata._Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,"aG",@progbits,_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh,comdat .weak _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .p2align 3, 0x0 _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh: .quad _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .size _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh" .size .L__unnamed_1, 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 _Z23__device_stub__memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8memcpy_kILi1024ELi16EEvPKmPrKPKhPrPh .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> constexpr int LEN_INITAL = 8; constexpr int STEPS = 6; constexpr int LEN_W = 1 + 2 * STEPS; constexpr int LEN_Z = LEN_W; constexpr int LEN_Y = LEN_INITAL + 2 * STEPS; constexpr int LEN_X = LEN_INITAL + 2 * STEPS; constexpr int LEN_TOTAL = LEN_W * LEN_Z * LEN_Y * LEN_X; constexpr int SIZE_W = LEN_Z * LEN_Y * LEN_X; constexpr int SIZE_Z = LEN_Y * LEN_X; constexpr int SIZE_Y = LEN_X; constexpr int NUM_THREADS = 512; constexpr int NUM_BLOCKS = LEN_TOTAL / NUM_THREADS; inline int coord_to_idx(const int w, const int z, const int y, const int x) { return w * SIZE_W + z * SIZE_Z + y * SIZE_Y + x; } inline void print_slice(const int* grid, const int w, const int z) { for (int row = 0; row < LEN_Y; ++row) { for (int col = 0; col < LEN_X; ++col) { std::cout << grid[coord_to_idx(w, z, row, col)] << " "; } std::cout << "\n"; } } __device__ int coord_to_idx_dev(const int w, const int z, const int y, const int x) { return w * SIZE_W + z * SIZE_Z + y * SIZE_Y + x; } __global__ void step(const int* grid, int* grid_next) { // Find out where we are. const int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < LEN_TOTAL) { int left = idx; int w = idx / SIZE_W; left = idx - w * SIZE_W; int z = left / SIZE_Z; left = left - z * SIZE_Z; int y = left / SIZE_Y; int x = left - y * SIZE_Y; // TODO: for loop here? const int active = grid[idx]; // Count active neighbors. int active_neighbors = 0; int min_nw = max(0, w - 1); int max_nw = min(LEN_W, w + 2); int min_nz = max(0, z - 1); int max_nz = min(LEN_Z, z + 2); int min_ny = max(0, y - 1); int max_ny = min(LEN_Y, y + 2); int min_nx = max(0, x - 1); int max_nx = min(LEN_X, x + 2); for (int nw = min_nw; nw < max_nw; ++nw) { for (int nz = min_nz; nz < max_nz; ++nz) { for (int ny = min_ny; ny < max_ny; ++ny) { for (int nx = min_nx; nx < max_nx; ++nx) { active_neighbors += grid[coord_to_idx_dev(nw, nz, ny, nx)]; } } } } active_neighbors -= active; // Rules int active_next = active; if (active == 1 && (active_neighbors < 2 || active_neighbors > 3)) { active_next = 0; } else if (active == 0 && active_neighbors == 3) { active_next = 1; } //active_next = idx; grid_next[idx] = active_next; } } int main() { // Initialize grid. int* grid; cudaMallocManaged(&grid, LEN_TOTAL * sizeof(int)); cudaMemset(grid, 0, LEN_TOTAL); int initial_grid[LEN_INITAL][LEN_INITAL] = { {1, 1, 0, 0, 1, 0, 1, 0}, {1, 1, 1, 0, 1, 0, 1, 1}, {0, 0, 1, 1, 1, 0, 0, 1}, {0, 1, 0, 0, 0, 0, 1, 1}, {0, 1, 0, 0, 1, 1, 1, 1}, {1, 1, 1, 1, 1, 0, 0, 0}, {1, 1, 1, 1, 1, 1, 1, 0}, {1, 0, 1, 1, 0, 1, 0, 1} }; for (int row = 0; row < LEN_INITAL; ++row) { for (int col = 0; col < LEN_INITAL; ++col) { grid[coord_to_idx(STEPS, STEPS, STEPS + row, STEPS + col)] = initial_grid[row][col]; } } //print_slice(grid, STEPS, STEPS); int* grid_next; cudaMallocManaged(&grid_next, LEN_TOTAL * sizeof(int)); for (int i = 0; i < STEPS; ++i) { //std::cout << "Step " << i << "\n"; step<<<NUM_BLOCKS, NUM_THREADS>>>(grid, grid_next); cudaDeviceSynchronize(); std::swap(grid, grid_next); //print_slice(grid, STEPS, STEPS); } // Count actives. int count = 0; for (int i = 0; i < LEN_TOTAL; ++i) { count += grid[i]; } std::cout << "Active: " << count << "\n"; cudaFree(grid); cudaFree(grid_next); return 0; }
code for sm_80 Function : _Z4stepPKiPi .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 R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R2, c[0x0][0x0], R3 ; /* 0x0000000002007a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GT.AND P0, PT, R0, 0x1080f, PT ; /* 0x0001080f0000780c */ /* 0x000fda0003f04270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ IMAD.HI R6, R0, 0x64d319ff, RZ ; /* 0x64d319ff00067827 */ /* 0x000fe200078e02ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0080*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe200078e00ff */ /*0090*/ SHF.R.U32.HI R7, RZ, 0x1f, R6 ; /* 0x0000001fff077819 */ /* 0x000fc60000011606 */ /*00a0*/ IMAD.WIDE R4, R0, R5, c[0x0][0x160] ; /* 0x0000580000047625 */ /* 0x000fe200078e0205 */ /*00b0*/ LEA.HI.SX32 R15, R6, R7, 0x15 ; /* 0x00000007060f7211 */ /* 0x000fca00078faaff */ /*00c0*/ IMAD R7, R15.reuse, -0x1450, R0 ; /* 0xffffebb00f077824 */ /* 0x040fe200078e0200 */ /*00d0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000162000c1e1900 */ /*00e0*/ IADD3 R10, R15, -0x1, RZ ; /* 0xffffffff0f0a7810 */ /* 0x000fe20007ffe0ff */ /*00f0*/ BSSY B3, 0xb60 ; /* 0x00000a6000037945 */ /* 0x000fe20003800000 */ /*0100*/ IMAD.HI R12, R7, 0x51eb851f, RZ ; /* 0x51eb851f070c7827 */ /* 0x000fe400078e02ff */ /*0110*/ IMNMX R10, RZ, R10, !PT ; /* 0x0000000aff0a7217 */ /* 0x000fe40007800200 */ /*0120*/ IMAD.MOV.U32 R11, RZ, RZ, RZ ; /* 0x000000ffff0b7224 */ /* 0x000fe200078e00ff */ /*0130*/ SHF.R.U32.HI R9, RZ, 0x1f, R12 ; /* 0x0000001fff097819 */ /* 0x000fe4000001160c */ /*0140*/ IADD3 R5, R15, 0x2, RZ ; /* 0x000000020f057810 */ /* 0x001fc40007ffe0ff */ /*0150*/ LEA.HI.SX32 R12, R12, R9, 0x19 ; /* 0x000000090c0c7211 */ /* 0x000fe400078fcaff */ /*0160*/ IMNMX R5, R5, 0xd, PT ; /* 0x0000000d05057817 */ /* 0x000fc60003800200 */ /*0170*/ IMAD R6, R12, -0x190, R7 ; /* 0xfffffe700c067824 */ /* 0x000fe200078e0207 */ /*0180*/ ISETP.GE.AND P0, PT, R10, R5, PT ; /* 0x000000050a00720c */ /* 0x000fe40003f06270 */ /*0190*/ SHF.R.S32.HI R7, RZ, 0x1f, R0 ; /* 0x0000001fff077819 */ /* 0x000fe20000011400 */ /*01a0*/ IMAD.HI R13, R6, 0x66666667, RZ ; /* 0x66666667060d7827 */ /* 0x000fca00078e02ff */ /*01b0*/ SHF.R.U32.HI R8, RZ, 0x1f, R13 ; /* 0x0000001fff087819 */ /* 0x000fc8000001160d */ /*01c0*/ LEA.HI.SX32 R13, R13, R8, 0x1d ; /* 0x000000080d0d7211 */ /* 0x000fc800078feaff */ /*01d0*/ IADD3 R8, R13.reuse, -0x1, RZ ; /* 0xffffffff0d087810 */ /* 0x040fe20007ffe0ff */ /*01e0*/ IMAD R14, R13, -0x14, R6 ; /* 0xffffffec0d0e7824 */ /* 0x000fe200078e0206 */ /*01f0*/ IADD3 R6, R12, -0x1, RZ ; /* 0xffffffff0c067810 */ /* 0x000fe40007ffe0ff */ /*0200*/ IMNMX R8, RZ, R8, !PT ; /* 0x00000008ff087217 */ /* 0x000fe40007800200 */ /*0210*/ IADD3 R9, R14, -0x1, RZ ; /* 0xffffffff0e097810 */ /* 0x000fe40007ffe0ff */ /*0220*/ IMNMX R6, RZ, R6, !PT ; /* 0x00000006ff067217 */ /* 0x000fe40007800200 */ /*0230*/ IMNMX R9, RZ, R9, !PT ; /* 0x00000009ff097217 */ /* 0x000fe20007800200 */ /*0240*/ @P0 BRA 0xb50 ; /* 0x0000090000000947 */ /* 0x000fea0003800000 */ /*0250*/ IMAD R16, R15, 0xd, R12 ; /* 0x0000000d0f107824 */ /* 0x000fe200078e020c */ /*0260*/ IADD3 R14, R14, 0x2, RZ ; /* 0x000000020e0e7810 */ /* 0x000fe20007ffe0ff */ /*0270*/ IMAD.MOV.U32 R15, RZ, RZ, 0x14 ; /* 0x00000014ff0f7424 */ /* 0x000fe200078e00ff */ /*0280*/ IADD3 R17, R13, 0x2, RZ ; /* 0x000000020d117810 */ /* 0x000fe20007ffe0ff */ /*0290*/ IMAD R16, R16, 0x14, R13 ; /* 0x0000001410107824 */ /* 0x000fe200078e020d */ /*02a0*/ IADD3 R13, R9.reuse, 0x1, RZ ; /* 0x00000001090d7810 */ /* 0x040fe20007ffe0ff */ /*02b0*/ IMAD.MOV R11, RZ, RZ, -R2 ; /* 0x000000ffff0b7224 */ /* 0x000fe200078e0a02 */ /*02c0*/ IADD3 R23, R9, 0x3, RZ ; /* 0x0000000309177810 */ /* 0x000fe20007ffe0ff */ /*02d0*/ IMAD R2, R16, R15, -0x3 ; /* 0xfffffffd10027424 */ /* 0x000fe200078e020f */ /*02e0*/ IADD3 R15, R12, 0x2, RZ ; /* 0x000000020c0f7810 */ /* 0x000fc40007ffe0ff */ /*02f0*/ IADD3 R12, R9, 0x2, RZ ; /* 0x00000002090c7810 */ /* 0x000fe20007ffe0ff */ /*0300*/ IMAD R2, R11, c[0x0][0x0], R2 ; /* 0x000000000b027a24 */ /* 0x000fe200078e0202 */ /*0310*/ IMNMX R14, R14, 0x14, PT ; /* 0x000000140e0e7817 */ /* 0x000fe20003800200 */ /*0320*/ IMAD.MOV.U32 R11, RZ, RZ, RZ ; /* 0x000000ffff0b7224 */ /* 0x000fe200078e00ff */ /*0330*/ IMNMX R15, R15, 0xd, PT ; /* 0x0000000d0f0f7817 */ /* 0x000fe20003800200 */ /*0340*/ IMAD.IADD R2, R2, 0x1, -R3 ; /* 0x0000000102027824 */ /* 0x000fe200078e0a03 */ /*0350*/ IMNMX R17, R17, 0x14, PT ; /* 0x0000001411117817 */ /* 0x000fc80003800200 */ /*0360*/ IMNMX R2, R2, -0x15, !PT ; /* 0xffffffeb02027817 */ /* 0x000fc80007800200 */ /*0370*/ IADD3 R3, -R9, -0x2, -R2 ; /* 0xfffffffe09037810 */ /* 0x000fe40007ffe902 */ /*0380*/ LOP3.LUT R2, RZ, R2, RZ, 0x33, !PT ; /* 0x00000002ff027212 */ /* 0x000fe400078e33ff */ /*0390*/ ISETP.GE.U32.AND P1, PT, R3, 0x3, PT ; /* 0x000000030300780c */ /* 0x000fc60003f26070 */ /*03a0*/ IMAD.IADD R2, R2, 0x1, -R9 ; /* 0x0000000102027824 */ /* 0x000fca00078e0a09 */ /*03b0*/ LOP3.LUT R16, R2, 0x3, RZ, 0xc0, !PT ; /* 0x0000000302107812 */ /* 0x000fe400078ec0ff */ /*03c0*/ ISETP.GE.AND P0, PT, R6, R15, PT ; /* 0x0000000f0600720c */ /* 0x000fe20003f06270 */ /*03d0*/ BSSY B2, 0xb20 ; /* 0x0000074000027945 */ /* 0x000fd80003800000 */ /*03e0*/ @P0 BRA 0xb10 ; /* 0x0000072000000947 */ /* 0x000fea0003800000 */ /*03f0*/ IMAD.MOV.U32 R18, RZ, RZ, R6 ; /* 0x000000ffff127224 */ /* 0x000fe400078e0006 */ /*0400*/ ISETP.GE.AND P0, PT, R8, R17, PT ; /* 0x000000110800720c */ /* 0x000fe20003f06270 */ /*0410*/ BSSY B1, 0xae0 ; /* 0x000006c000017945 */ /* 0x000fd80003800000 */ /*0420*/ @P0 BRA 0xad0 ; /* 0x000006a000000947 */ /* 0x000fea0003800000 */ /*0430*/ IMAD R19, R10, 0xd, R18 ; /* 0x0000000d0a137824 */ /* 0x000fe400078e0212 */ /*0440*/ IMAD.MOV.U32 R20, RZ, RZ, R8 ; /* 0x000000ffff147224 */ /* 0x000fe400078e0008 */ /*0450*/ IMAD R19, R19, 0x190, RZ ; /* 0x0000019013137824 */ /* 0x000fe400078e02ff */ /*0460*/ ISETP.GE.AND P0, PT, R9, R14, PT ; /* 0x0000000e0900720c */ /* 0x000fe20003f06270 */ /*0470*/ BSSY B0, 0xaa0 ; /* 0x0000062000007945 */ /* 0x000fd80003800000 */ /*0480*/ @P0 BRA 0xa90 ; /* 0x0000060000000947 */ /* 0x000fea0003800000 */ /*0490*/ ISETP.NE.AND P0, PT, R16, RZ, PT ; /* 0x000000ff1000720c */ /* 0x000fe20003f05270 */ /*04a0*/ BSSY B4, 0x5e0 ; /* 0x0000013000047945 */ /* 0x000fe20003800000 */ /*04b0*/ IMAD.MOV.U32 R21, RZ, RZ, R9 ; /* 0x000000ffff157224 */ /* 0x000fd600078e0009 */ /*04c0*/ @!P0 BRA 0x5d0 ; /* 0x0000010000008947 */ /* 0x000fea0003800000 */ /*04d0*/ IMAD R2, R20, 0x14, R19 ; /* 0x0000001414027824 */ /* 0x000fe400078e0213 */ /*04e0*/ IMAD.MOV.U32 R3, RZ, RZ, 0x4 ; /* 0x00000004ff037424 */ /* 0x000fe400078e00ff */ /*04f0*/ IMAD.IADD R2, R9, 0x1, R2 ; /* 0x0000000109027824 */ /* 0x000fc800078e0202 */ /*0500*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*0510*/ LDG.E R22, [R2.64] ; /* 0x0000000402167981 */ /* 0x000ea2000c1e1900 */ /*0520*/ ISETP.NE.AND P0, PT, R16, 0x1, PT ; /* 0x000000011000780c */ /* 0x000fe20003f05270 */ /*0530*/ IMAD.MOV.U32 R21, RZ, RZ, R13 ; /* 0x000000ffff157224 */ /* 0x000fe400078e000d */ /*0540*/ IMAD.IADD R11, R11, 0x1, R22 ; /* 0x000000010b0b7824 */ /* 0x004fd400078e0216 */ /*0550*/ @!P0 BRA 0x5d0 ; /* 0x0000007000008947 */ /* 0x000fea0003800000 */ /*0560*/ ISETP.NE.AND P0, PT, R16, 0x2, PT ; /* 0x000000021000780c */ /* 0x000fe20003f05270 */ /*0570*/ LDG.E R22, [R2.64+0x4] ; /* 0x0000040402167981 */ /* 0x000e98000c1e1900 */ /*0580*/ @P0 LDG.E R24, [R2.64+0x8] ; /* 0x0000080402180981 */ /* 0x000ee2000c1e1900 */ /*0590*/ IMAD.MOV.U32 R21, RZ, RZ, R12 ; /* 0x000000ffff157224 */ /* 0x000fe400078e000c */ /*05a0*/ @P0 IMAD.MOV.U32 R21, RZ, RZ, R23 ; /* 0x000000ffff150224 */ /* 0x000fe400078e0017 */ /*05b0*/ IMAD.IADD R11, R11, 0x1, R22 ; /* 0x000000010b0b7824 */ /* 0x004fc800078e0216 */ /*05c0*/ @P0 IMAD.IADD R11, R11, 0x1, R24 ; /* 0x000000010b0b0824 */ /* 0x008fe400078e0218 */ /*05d0*/ BSYNC B4 ; /* 0x0000000000047941 */ /* 0x000fea0003800000 */ /*05e0*/ @!P1 BRA 0xa90 ; /* 0x000004a000009947 */ /* 0x000fea0003800000 */ /*05f0*/ IMAD.IADD R2, R14, 0x1, -R21.reuse ; /* 0x000000010e027824 */ /* 0x100fe200078e0a15 */ /*0600*/ BSSY B4, 0x8a0 ; /* 0x0000029000047945 */ /* 0x000fe20003800000 */ /*0610*/ IMAD.IADD R3, R19, 0x1, R21 ; /* 0x0000000113037824 */ /* 0x000fe200078e0215 */ /*0620*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x80, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0f070 */ /*0630*/ ISETP.GT.AND P2, PT, R2, 0xc, PT ; /* 0x0000000c0200780c */ /* 0x000fe20003f44270 */ /*0640*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fe400078e00ff */ /*0650*/ IMAD R3, R20, 0x14, R3 ; /* 0x0000001414037824 */ /* 0x000fc800078e0203 */ /*0660*/ IMAD.WIDE R2, R3, R2, c[0x0][0x160] ; /* 0x0000580003027625 */ /* 0x000fcc00078e0202 */ /*0670*/ @!P2 BRA 0x890 ; /* 0x000002100000a947 */ /* 0x000fea0003800000 */ /*0680*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0690*/ IADD3 R22, R14, -0xc, RZ ; /* 0xfffffff40e167810 */ /* 0x000fe40007ffe0ff */ /*06a0*/ LDG.E R24, [R2.64] ; /* 0x0000000402187981 */ /* 0x000ea8000c1e1900 */ /*06b0*/ LDG.E R25, [R2.64+0x4] ; /* 0x0000040402197981 */ /* 0x000ea8000c1e1900 */ /*06c0*/ LDG.E R27, [R2.64+0x8] ; /* 0x00000804021b7981 */ /* 0x000ee8000c1e1900 */ /*06d0*/ LDG.E R26, [R2.64+0xc] ; /* 0x00000c04021a7981 */ /* 0x000ee2000c1e1900 */ /*06e0*/ IADD3 R24, R25, R24, R11 ; /* 0x0000001819187210 */ /* 0x004fc60007ffe00b */ /*06f0*/ LDG.E R11, [R2.64+0x10] ; /* 0x00001004020b7981 */ /* 0x000ea8000c1e1900 */ /*0700*/ LDG.E R25, [R2.64+0x14] ; /* 0x0000140402197981 */ /* 0x000ea2000c1e1900 */ /*0710*/ IADD3 R24, R26, R27, R24 ; /* 0x0000001b1a187210 */ /* 0x008fc60007ffe018 */ /*0720*/ LDG.E R26, [R2.64+0x18] ; /* 0x00001804021a7981 */ /* 0x000ee8000c1e1900 */ /*0730*/ LDG.E R27, [R2.64+0x1c] ; /* 0x00001c04021b7981 */ /* 0x000ee2000c1e1900 */ /*0740*/ IADD3 R11, R25, R11, R24 ; /* 0x0000000b190b7210 */ /* 0x004fc60007ffe018 */ /*0750*/ LDG.E R24, [R2.64+0x20] ; /* 0x0000200402187981 */ /* 0x000ea8000c1e1900 */ /*0760*/ LDG.E R25, [R2.64+0x24] ; /* 0x0000240402197981 */ /* 0x000ea2000c1e1900 */ /*0770*/ IADD3 R11, R27, R26, R11 ; /* 0x0000001a1b0b7210 */ /* 0x008fc60007ffe00b */ /*0780*/ LDG.E R26, [R2.64+0x28] ; /* 0x00002804021a7981 */ /* 0x000ee8000c1e1900 */ /*0790*/ LDG.E R27, [R2.64+0x2c] ; /* 0x00002c04021b7981 */ /* 0x000ee2000c1e1900 */ /*07a0*/ IADD3 R11, R25, R24, R11 ; /* 0x00000018190b7210 */ /* 0x004fc60007ffe00b */ /*07b0*/ LDG.E R25, [R2.64+0x30] ; /* 0x0000300402197981 */ /* 0x000ea8000c1e1900 */ /*07c0*/ LDG.E R24, [R2.64+0x38] ; /* 0x0000380402187981 */ /* 0x000f22000c1e1900 */ /*07d0*/ IADD3 R26, R27, R26, R11 ; /* 0x0000001a1b1a7210 */ /* 0x008fc60007ffe00b */ /*07e0*/ LDG.E R27, [R2.64+0x34] ; /* 0x00003404021b7981 */ /* 0x000ea8000c1e1900 */ /*07f0*/ LDG.E R11, [R2.64+0x3c] ; /* 0x00003c04020b7981 */ /* 0x000f22000c1e1900 */ /*0800*/ IADD3 R21, R21, 0x10, RZ ; /* 0x0000001015157810 */ /* 0x000fc80007ffe0ff */ /*0810*/ ISETP.GE.AND P2, PT, R21, R22, PT ; /* 0x000000161500720c */ /* 0x000fe40003f46270 */ /*0820*/ IADD3 R25, R27, R25, R26 ; /* 0x000000191b197210 */ /* 0x004fe40007ffe01a */ /*0830*/ IADD3 R26, P3, R2, 0x40, RZ ; /* 0x00000040021a7810 */ /* 0x000fca0007f7e0ff */ /*0840*/ IMAD.X R27, RZ, RZ, R3, P3 ; /* 0x000000ffff1b7224 */ /* 0x000fe200018e0603 */ /*0850*/ IADD3 R11, R11, R24, R25 ; /* 0x000000180b0b7210 */ /* 0x010fe20007ffe019 */ /*0860*/ IMAD.MOV.U32 R2, RZ, RZ, R26 ; /* 0x000000ffff027224 */ /* 0x000fe400078e001a */ /*0870*/ IMAD.MOV.U32 R3, RZ, RZ, R27 ; /* 0x000000ffff037224 */ /* 0x000fe200078e001b */ /*0880*/ @!P2 BRA 0x6a0 ; /* 0xfffffe100000a947 */ /* 0x000fea000383ffff */ /*0890*/ BSYNC B4 ; /* 0x0000000000047941 */ /* 0x000fea0003800000 */ /*08a0*/ IMAD.IADD R22, R14, 0x1, -R21 ; /* 0x000000010e167824 */ /* 0x000fe200078e0a15 */ /*08b0*/ BSSY B4, 0xa10 ; /* 0x0000015000047945 */ /* 0x000fe80003800000 */ /*08c0*/ ISETP.GT.AND P2, PT, R22, 0x4, PT ; /* 0x000000041600780c */ /* 0x000fda0003f44270 */ /*08d0*/ @!P2 BRA 0xa00 ; /* 0x000001200000a947 */ /* 0x000fea0003800000 */ /*08e0*/ LDG.E R22, [R2.64] ; /* 0x0000000402167981 */ /* 0x000ea8000c1e1900 */ /*08f0*/ LDG.E R24, [R2.64+0x4] ; /* 0x0000040402187981 */ /* 0x000ea8000c1e1900 */ /*0900*/ LDG.E R25, [R2.64+0x8] ; /* 0x0000080402197981 */ /* 0x000ee8000c1e1900 */ /*0910*/ LDG.E R26, [R2.64+0xc] ; /* 0x00000c04021a7981 */ /* 0x000ee2000c1e1900 */ /*0920*/ IADD3 R22, R24, R22, R11 ; /* 0x0000001618167210 */ /* 0x004fc60007ffe00b */ /*0930*/ LDG.E R11, [R2.64+0x10] ; /* 0x00001004020b7981 */ /* 0x000ea8000c1e1900 */ /*0940*/ LDG.E R24, [R2.64+0x14] ; /* 0x0000140402187981 */ /* 0x000ea2000c1e1900 */ /*0950*/ IADD3 R22, R26, R25, R22 ; /* 0x000000191a167210 */ /* 0x008fc60007ffe016 */ /*0960*/ LDG.E R26, [R2.64+0x18] ; /* 0x00001804021a7981 */ /* 0x000ee8000c1e1900 */ /*0970*/ LDG.E R25, [R2.64+0x1c] ; /* 0x00001c0402197981 */ /* 0x000ee2000c1e1900 */ /*0980*/ PLOP3.LUT P0, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f0e170 */ /*0990*/ IADD3 R21, R21, 0x8, RZ ; /* 0x0000000815157810 */ /* 0x000fe40007ffe0ff */ /*09a0*/ IADD3 R11, R24, R11, R22 ; /* 0x0000000b180b7210 */ /* 0x004fe40007ffe016 */ /*09b0*/ IADD3 R22, P2, R2, 0x20, RZ ; /* 0x0000002002167810 */ /* 0x000fca0007f5e0ff */ /*09c0*/ IMAD.X R27, RZ, RZ, R3, P2 ; /* 0x000000ffff1b7224 */ /* 0x000fe200010e0603 */ /*09d0*/ IADD3 R11, R25, R26, R11 ; /* 0x0000001a190b7210 */ /* 0x008fe20007ffe00b */ /*09e0*/ IMAD.MOV.U32 R2, RZ, RZ, R22 ; /* 0x000000ffff027224 */ /* 0x000fe400078e0016 */ /*09f0*/ IMAD.MOV.U32 R3, RZ, RZ, R27 ; /* 0x000000ffff037224 */ /* 0x000fe400078e001b */ /*0a00*/ BSYNC B4 ; /* 0x0000000000047941 */ /* 0x000fea0003800000 */ /*0a10*/ ISETP.LT.OR P0, PT, R21, R14, P0 ; /* 0x0000000e1500720c */ /* 0x000fda0000701670 */ /*0a20*/ @!P0 BRA 0xa90 ; /* 0x0000006000008947 */ /* 0x000fea0003800000 */ /*0a30*/ LDG.E R22, [R2.64] ; /* 0x0000000402167981 */ /* 0x000ea8000c1e1900 */ /*0a40*/ LDG.E R21, [R2.64+0x4] ; /* 0x0000040402157981 */ /* 0x000ea8000c1e1900 */ /*0a50*/ LDG.E R24, [R2.64+0x8] ; /* 0x0000080402187981 */ /* 0x000ee8000c1e1900 */ /*0a60*/ LDG.E R25, [R2.64+0xc] ; /* 0x00000c0402197981 */ /* 0x000ee2000c1e1900 */ /*0a70*/ IADD3 R11, R21, R22, R11 ; /* 0x00000016150b7210 */ /* 0x004fc80007ffe00b */ /*0a80*/ IADD3 R11, R25, R24, R11 ; /* 0x00000018190b7210 */ /* 0x008fe40007ffe00b */ /*0a90*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0aa0*/ IADD3 R20, R20, 0x1, RZ ; /* 0x0000000114147810 */ /* 0x000fc80007ffe0ff */ /*0ab0*/ ISETP.GE.AND P0, PT, R20, R17, PT ; /* 0x000000111400720c */ /* 0x000fda0003f06270 */ /*0ac0*/ @!P0 BRA 0x460 ; /* 0xfffff99000008947 */ /* 0x000fea000383ffff */ /*0ad0*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0ae0*/ IADD3 R18, R18, 0x1, RZ ; /* 0x0000000112127810 */ /* 0x000fc80007ffe0ff */ /*0af0*/ ISETP.GE.AND P0, PT, R18, R15, PT ; /* 0x0000000f1200720c */ /* 0x000fda0003f06270 */ /*0b00*/ @!P0 BRA 0x400 ; /* 0xfffff8f000008947 */ /* 0x000fea000383ffff */ /*0b10*/ BSYNC B2 ; /* 0x0000000000027941 */ /* 0x000fea0003800000 */ /*0b20*/ IADD3 R10, R10, 0x1, RZ ; /* 0x000000010a0a7810 */ /* 0x000fc80007ffe0ff */ /*0b30*/ ISETP.GE.AND P0, PT, R10, R5, PT ; /* 0x000000050a00720c */ /* 0x000fda0003f06270 */ /*0b40*/ @!P0 BRA 0x3c0 ; /* 0xfffff87000008947 */ /* 0x000fea000383ffff */ /*0b50*/ BSYNC B3 ; /* 0x0000000000037941 */ /* 0x000fea0003800000 */ /*0b60*/ IMAD.IADD R11, R11, 0x1, -R4 ; /* 0x000000010b0b7824 */ /* 0x020fca00078e0a04 */ /*0b70*/ LOP3.LUT R2, R11.reuse, 0xfffffffe, RZ, 0xc0, !PT ; /* 0xfffffffe0b027812 */ /* 0x040fe400078ec0ff */ /*0b80*/ ISETP.NE.AND P1, PT, R11, 0x3, PT ; /* 0x000000030b00780c */ /* 0x000fe40003f25270 */ /*0b90*/ ISETP.NE.AND P0, PT, R2, 0x2, PT ; /* 0x000000020200780c */ /* 0x000fe40003f05270 */ /*0ba0*/ ISETP.EQ.AND P1, PT, R4.reuse, RZ, !P1 ; /* 0x000000ff0400720c */ /* 0x040fe40004f22270 */ /*0bb0*/ ISETP.EQ.AND P0, PT, R4, 0x1, P0 ; /* 0x000000010400780c */ /* 0x000fe40000702270 */ /*0bc0*/ LEA R2, P2, R0, c[0x0][0x168], 0x2 ; /* 0x00005a0000027a11 */ /* 0x000fc400078410ff */ /*0bd0*/ SEL R4, R4, 0x1, !P1 ; /* 0x0000000104047807 */ /* 0x000fe40004800000 */ /*0be0*/ LEA.HI.X R3, R0, c[0x0][0x16c], R7, 0x2, P2 ; /* 0x00005b0000037a11 */ /* 0x000fe400010f1407 */ /*0bf0*/ SEL R5, R4, RZ, !P0 ; /* 0x000000ff04057207 */ /* 0x000fca0004000000 */ /*0c00*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0c10*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0c20*/ BRA 0xc20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0c30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ca0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cb0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ce0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0cf0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <iostream> constexpr int LEN_INITAL = 8; constexpr int STEPS = 6; constexpr int LEN_W = 1 + 2 * STEPS; constexpr int LEN_Z = LEN_W; constexpr int LEN_Y = LEN_INITAL + 2 * STEPS; constexpr int LEN_X = LEN_INITAL + 2 * STEPS; constexpr int LEN_TOTAL = LEN_W * LEN_Z * LEN_Y * LEN_X; constexpr int SIZE_W = LEN_Z * LEN_Y * LEN_X; constexpr int SIZE_Z = LEN_Y * LEN_X; constexpr int SIZE_Y = LEN_X; constexpr int NUM_THREADS = 512; constexpr int NUM_BLOCKS = LEN_TOTAL / NUM_THREADS; inline int coord_to_idx(const int w, const int z, const int y, const int x) { return w * SIZE_W + z * SIZE_Z + y * SIZE_Y + x; } inline void print_slice(const int* grid, const int w, const int z) { for (int row = 0; row < LEN_Y; ++row) { for (int col = 0; col < LEN_X; ++col) { std::cout << grid[coord_to_idx(w, z, row, col)] << " "; } std::cout << "\n"; } } __device__ int coord_to_idx_dev(const int w, const int z, const int y, const int x) { return w * SIZE_W + z * SIZE_Z + y * SIZE_Y + x; } __global__ void step(const int* grid, int* grid_next) { // Find out where we are. const int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < LEN_TOTAL) { int left = idx; int w = idx / SIZE_W; left = idx - w * SIZE_W; int z = left / SIZE_Z; left = left - z * SIZE_Z; int y = left / SIZE_Y; int x = left - y * SIZE_Y; // TODO: for loop here? const int active = grid[idx]; // Count active neighbors. int active_neighbors = 0; int min_nw = max(0, w - 1); int max_nw = min(LEN_W, w + 2); int min_nz = max(0, z - 1); int max_nz = min(LEN_Z, z + 2); int min_ny = max(0, y - 1); int max_ny = min(LEN_Y, y + 2); int min_nx = max(0, x - 1); int max_nx = min(LEN_X, x + 2); for (int nw = min_nw; nw < max_nw; ++nw) { for (int nz = min_nz; nz < max_nz; ++nz) { for (int ny = min_ny; ny < max_ny; ++ny) { for (int nx = min_nx; nx < max_nx; ++nx) { active_neighbors += grid[coord_to_idx_dev(nw, nz, ny, nx)]; } } } } active_neighbors -= active; // Rules int active_next = active; if (active == 1 && (active_neighbors < 2 || active_neighbors > 3)) { active_next = 0; } else if (active == 0 && active_neighbors == 3) { active_next = 1; } //active_next = idx; grid_next[idx] = active_next; } } int main() { // Initialize grid. int* grid; cudaMallocManaged(&grid, LEN_TOTAL * sizeof(int)); cudaMemset(grid, 0, LEN_TOTAL); int initial_grid[LEN_INITAL][LEN_INITAL] = { {1, 1, 0, 0, 1, 0, 1, 0}, {1, 1, 1, 0, 1, 0, 1, 1}, {0, 0, 1, 1, 1, 0, 0, 1}, {0, 1, 0, 0, 0, 0, 1, 1}, {0, 1, 0, 0, 1, 1, 1, 1}, {1, 1, 1, 1, 1, 0, 0, 0}, {1, 1, 1, 1, 1, 1, 1, 0}, {1, 0, 1, 1, 0, 1, 0, 1} }; for (int row = 0; row < LEN_INITAL; ++row) { for (int col = 0; col < LEN_INITAL; ++col) { grid[coord_to_idx(STEPS, STEPS, STEPS + row, STEPS + col)] = initial_grid[row][col]; } } //print_slice(grid, STEPS, STEPS); int* grid_next; cudaMallocManaged(&grid_next, LEN_TOTAL * sizeof(int)); for (int i = 0; i < STEPS; ++i) { //std::cout << "Step " << i << "\n"; step<<<NUM_BLOCKS, NUM_THREADS>>>(grid, grid_next); cudaDeviceSynchronize(); std::swap(grid, grid_next); //print_slice(grid, STEPS, STEPS); } // Count actives. int count = 0; for (int i = 0; i < LEN_TOTAL; ++i) { count += grid[i]; } std::cout << "Active: " << count << "\n"; cudaFree(grid); cudaFree(grid_next); return 0; }
.file "tmpxft_00178d59_00000000-6_17.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3676: .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 .LFE3676: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z16coord_to_idx_deviiii .type _Z16coord_to_idx_deviiii, @function _Z16coord_to_idx_deviiii: .LFB3671: .cfi_startproc endbr64 pushq %rax .cfi_def_cfa_offset 16 popq %rax .cfi_def_cfa_offset 8 subq $24, %rsp .cfi_def_cfa_offset 32 movl $1, 12(%rsp) movl 12(%rsp), %edi call exit@PLT .cfi_endproc .LFE3671: .size _Z16coord_to_idx_deviiii, .-_Z16coord_to_idx_deviiii .globl _Z26__device_stub__Z4stepPKiPiPKiPi .type _Z26__device_stub__Z4stepPKiPiPKiPi, @function _Z26__device_stub__Z4stepPKiPiPKiPi: .LFB3698: .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 .L9 .L5: movq 104(%rsp), %rax subq %fs:40, %rax jne .L10 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .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 _Z4stepPKiPi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L5 .L10: call __stack_chk_fail@PLT .cfi_endproc .LFE3698: .size _Z26__device_stub__Z4stepPKiPiPKiPi, .-_Z26__device_stub__Z4stepPKiPiPKiPi .globl _Z4stepPKiPi .type _Z4stepPKiPi, @function _Z4stepPKiPi: .LFB3699: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z4stepPKiPiPKiPi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3699: .size _Z4stepPKiPi, .-_Z4stepPKiPi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Active: " .LC1: .string "\n" .text .globl main .type main, @function main: .LFB3672: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $320, %rsp .cfi_def_cfa_offset 336 movq %fs:40, %rax movq %rax, 312(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $1, %edx movl $270400, %esi call cudaMallocManaged@PLT movl $67600, %edx movl $0, %esi movq 8(%rsp), %rdi call cudaMemset@PLT movl $1, 48(%rsp) movl $1, 52(%rsp) movl $0, 56(%rsp) movl $0, 60(%rsp) movl $1, 64(%rsp) movl $0, 68(%rsp) movl $1, 72(%rsp) movl $0, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $1, 88(%rsp) movl $0, 92(%rsp) movl $1, 96(%rsp) movl $0, 100(%rsp) movl $1, 104(%rsp) movl $1, 108(%rsp) movl $0, 112(%rsp) movl $0, 116(%rsp) movl $1, 120(%rsp) movl $1, 124(%rsp) movl $1, 128(%rsp) movl $0, 132(%rsp) movl $0, 136(%rsp) movl $1, 140(%rsp) movl $0, 144(%rsp) movl $1, 148(%rsp) movl $0, 152(%rsp) movl $0, 156(%rsp) movl $0, 160(%rsp) movl $0, 164(%rsp) movl $1, 168(%rsp) movl $1, 172(%rsp) movl $0, 176(%rsp) movl $1, 180(%rsp) movl $0, 184(%rsp) movl $0, 188(%rsp) movl $1, 192(%rsp) movl $1, 196(%rsp) movl $1, 200(%rsp) movl $1, 204(%rsp) movl $1, 208(%rsp) movl $1, 212(%rsp) movl $1, 216(%rsp) movl $1, 220(%rsp) movl $1, 224(%rsp) movl $0, 228(%rsp) movl $0, 232(%rsp) movl $0, 236(%rsp) movl $1, 240(%rsp) movl $1, 244(%rsp) movl $1, 248(%rsp) movl $1, 252(%rsp) movl $1, 256(%rsp) movl $1, 260(%rsp) movl $1, 264(%rsp) movl $0, 268(%rsp) movl $1, 272(%rsp) movl $0, 276(%rsp) movl $1, 280(%rsp) movl $1, 284(%rsp) movl $0, 288(%rsp) movl $1, 292(%rsp) movl $0, 296(%rsp) movl $1, 300(%rsp) leaq 48(%rsp), %rdi leaq -336(%rsp), %r8 movl $134936, %esi .L14: leaq -32(%rsi), %rax .L15: movl -134904(%rdi,%rax), %ecx movq 8(%rsp), %rdx movl %ecx, (%rdx,%rax) addq $4, %rax cmpq %rsi, %rax jne .L15 subq $48, %rdi addq $80, %rsi cmpq %r8, %rdi jne .L14 leaq 16(%rsp), %rdi movl $1, %edx movl $270400, %esi call cudaMallocManaged@PLT movl $6, %ebx jmp .L18 .L17: call cudaDeviceSynchronize@PLT movq 8(%rsp), %rax movq 16(%rsp), %rcx movq %rcx, 8(%rsp) movq %rax, 16(%rsp) subl $1, %ebx je .L25 .L18: movl $512, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $132, 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 jne .L17 movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z26__device_stub__Z4stepPKiPiPKiPi jmp .L17 .L25: movq %rcx, %rax addq $270400, %rcx movl $0, %edx .L19: movl %edx, %ebx addl (%rax), %ebx movl %ebx, %edx addq $4, %rax cmpq %rcx, %rax jne .L19 leaq .LC0(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movl %ebx, %esi call _ZNSolsEi@PLT movq %rax, %rdi leaq .LC1(%rip), %rsi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 312(%rsp), %rax subq %fs:40, %rax jne .L26 movl $0, %eax addq $320, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L26: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3672: .size main, .-main .section .rodata.str1.1 .LC2: .string "_Z4stepPKiPi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3701: .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 _Z4stepPKiPi(%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 .LFE3701: .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: