system
stringclasses
7 values
input
stringlengths
10
94.5k
output
stringlengths
14
55.5k
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include<stdio.h> #define NUM_THREADS_PER_BLOCK 256 __global__ void print_hello() { int idx = threadIdx.x; printf("Hello World! My threadId is %d\n", idx); } int main() { print_hello<<<1, NUM_THREADS_PER_BLOCK>>>(); cudaDeviceSynchronize(); return 0; }
.file "tmpxft_00185772_00000000-6_hello_world.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 _Z30__device_stub__Z11print_hellovv .type _Z30__device_stub__Z11print_hellovv, @function _Z30__device_stub__Z11print_hellovv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z11print_hellov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z30__device_stub__Z11print_hellovv, .-_Z30__device_stub__Z11print_hellovv .globl _Z11print_hellov .type _Z11print_hellov, @function _Z11print_hellov: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z11print_hellovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z11print_hellov, .-_Z11print_hellov .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $256, 20(%rsp) movl $1, 24(%rsp) movl $1, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z30__device_stub__Z11print_hellovv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z11print_hellov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z11print_hellov(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include<stdio.h> #define NUM_THREADS_PER_BLOCK 256 __global__ void print_hello() { int idx = threadIdx.x; printf("Hello World! My threadId is %d\n", idx); } int main() { print_hello<<<1, NUM_THREADS_PER_BLOCK>>>(); cudaDeviceSynchronize(); return 0; }
#include <hip/hip_runtime.h> #include<stdio.h> #define NUM_THREADS_PER_BLOCK 256 __global__ void print_hello() { int idx = threadIdx.x; printf("Hello World! My threadId is %d\n", idx); } int main() { print_hello<<<1, NUM_THREADS_PER_BLOCK>>>(); hipDeviceSynchronize(); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include<stdio.h> #define NUM_THREADS_PER_BLOCK 256 __global__ void print_hello() { int idx = threadIdx.x; printf("Hello World! My threadId is %d\n", idx); } int main() { print_hello<<<1, NUM_THREADS_PER_BLOCK>>>(); hipDeviceSynchronize(); return 0; }
.text .file "hello_world.hip" .globl _Z26__device_stub__print_hellov # -- Begin function _Z26__device_stub__print_hellov .p2align 4, 0x90 .type _Z26__device_stub__print_hellov,@function _Z26__device_stub__print_hellov: # @_Z26__device_stub__print_hellov .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 $_Z11print_hellov, %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 _Z26__device_stub__print_hellov, .Lfunc_end0-_Z26__device_stub__print_hellov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movabsq $4294967297, %rdi # imm = 0x100000001 leaq 255(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z11print_hellov, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11print_hellov, %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 _Z11print_hellov,@object # @_Z11print_hellov .section .rodata,"a",@progbits .globl _Z11print_hellov .p2align 3, 0x0 _Z11print_hellov: .quad _Z26__device_stub__print_hellov .size _Z11print_hellov, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11print_hellov" .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 _Z26__device_stub__print_hellov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11print_hellov .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_00185772_00000000-6_hello_world.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 _Z30__device_stub__Z11print_hellovv .type _Z30__device_stub__Z11print_hellovv, @function _Z30__device_stub__Z11print_hellovv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z11print_hellov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z30__device_stub__Z11print_hellovv, .-_Z30__device_stub__Z11print_hellovv .globl _Z11print_hellov .type _Z11print_hellov, @function _Z11print_hellov: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z11print_hellovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z11print_hellov, .-_Z11print_hellov .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $256, 20(%rsp) movl $1, 24(%rsp) movl $1, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z30__device_stub__Z11print_hellovv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z11print_hellov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z11print_hellov(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello_world.hip" .globl _Z26__device_stub__print_hellov # -- Begin function _Z26__device_stub__print_hellov .p2align 4, 0x90 .type _Z26__device_stub__print_hellov,@function _Z26__device_stub__print_hellov: # @_Z26__device_stub__print_hellov .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 $_Z11print_hellov, %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 _Z26__device_stub__print_hellov, .Lfunc_end0-_Z26__device_stub__print_hellov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movabsq $4294967297, %rdi # imm = 0x100000001 leaq 255(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z11print_hellov, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11print_hellov, %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 _Z11print_hellov,@object # @_Z11print_hellov .section .rodata,"a",@progbits .globl _Z11print_hellov .p2align 3, 0x0 _Z11print_hellov: .quad _Z26__device_stub__print_hellov .size _Z11print_hellov, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11print_hellov" .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 _Z26__device_stub__print_hellov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11print_hellov .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 <string> #include <fstream> #include <sstream> #include <list> #include <iterator> #include "file_handler.cuh" void prepareFile(std::string fileName, bool truncate) { std::ofstream resFile (fileName, (truncate ? std::ofstream::out | std::ofstream::trunc : std::ios_base::app)); resFile << "["; resFile.close(); } void finishFile(std::string fileName, bool removeLast) { std::ifstream resFile1 (fileName); std::stringstream buffer; buffer << resFile1.rdbuf(); std::string contents = buffer.str(); resFile1.close(); if(removeLast) contents.pop_back(); contents += "]\n"; std::ofstream resFile2 (fileName, std::ofstream::out | std::ofstream::trunc); resFile2 << contents; resFile2.close(); } void writeLogNormConstToFile(double logNormConstant) { // std::ofstream resFile (fileName); std::ofstream resFile (Z_FILE_NAME, std::ios_base::app); // If append to file is wanted if(resFile.is_open()) { resFile << logNormConstant << ","; resFile.close(); } else { printf("Could not open file %s\n", Z_FILE_NAME.c_str()); } } void writeESSToFile(std::list<double> essList) { // std::ofstream resFile (fileName); std::ofstream resFile (ESS_FILE_NAME, std::ios_base::app); // If append to file is wanted if(resFile.is_open()) { resFile << "["; std::list <double> :: iterator it; for(it = essList.begin(); it != essList.end(); ++it) resFile << *it << " "; resFile << "],\n"; resFile.close(); } else { printf("Could not open file %s\n", ESS_FILE_NAME.c_str()); } }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <string> #include <fstream> #include <sstream> #include <list> #include <iterator> #include "file_handler.cuh" void prepareFile(std::string fileName, bool truncate) { std::ofstream resFile (fileName, (truncate ? std::ofstream::out | std::ofstream::trunc : std::ios_base::app)); resFile << "["; resFile.close(); } void finishFile(std::string fileName, bool removeLast) { std::ifstream resFile1 (fileName); std::stringstream buffer; buffer << resFile1.rdbuf(); std::string contents = buffer.str(); resFile1.close(); if(removeLast) contents.pop_back(); contents += "]\n"; std::ofstream resFile2 (fileName, std::ofstream::out | std::ofstream::trunc); resFile2 << contents; resFile2.close(); } void writeLogNormConstToFile(double logNormConstant) { // std::ofstream resFile (fileName); std::ofstream resFile (Z_FILE_NAME, std::ios_base::app); // If append to file is wanted if(resFile.is_open()) { resFile << logNormConstant << ","; resFile.close(); } else { printf("Could not open file %s\n", Z_FILE_NAME.c_str()); } } void writeESSToFile(std::list<double> essList) { // std::ofstream resFile (fileName); std::ofstream resFile (ESS_FILE_NAME, std::ios_base::app); // If append to file is wanted if(resFile.is_open()) { resFile << "["; std::list <double> :: iterator it; for(it = essList.begin(); it != essList.end(); ++it) resFile << *it << " "; resFile << "],\n"; resFile.close(); } else { printf("Could not open file %s\n", ESS_FILE_NAME.c_str()); } }
#ifndef FILE_HANDLER_INCLUDED #define FILE_HANDLER_INCLUDED #include <string> #include <list> const std::string Z_FILE_NAME = "log_norm_const.txt"; const std::string ESS_FILE_NAME = "ess.txt"; /** * Optionally clears the file, and writes an opening bracket to it. * * @param fileName the name of the file to save to. * @param truncate if the file should be cleared first. */ void prepareFile(std::string fileName, bool truncate); /** * Optionally removes the last character of the file, and writes a closing bracket and new line to it. * * @param fileName the name of the file to save to. * @param removeLast if the last character of the file should be deleted before closing it. */ void finishFile(std::string fileName, bool removeLast); /** * Writes the approximated log normalization constant to file. The file is cleared before the * program starts and if multiple runs is performed, the values from each run is appended to this file. * * @param logNormConstant */ void writeLogNormConstToFile(double logNormConstant); /** * Writes the effective sample size (ESS) to file. The file is cleared before the program * starts. * * @param ess */ void writeESSToFile(std::list<double> essList); #endif
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#ifndef FILE_HANDLER_INCLUDED #define FILE_HANDLER_INCLUDED #include <string> #include <list> const std::string Z_FILE_NAME = "log_norm_const.txt"; const std::string ESS_FILE_NAME = "ess.txt"; /** * Optionally clears the file, and writes an opening bracket to it. * * @param fileName the name of the file to save to. * @param truncate if the file should be cleared first. */ void prepareFile(std::string fileName, bool truncate); /** * Optionally removes the last character of the file, and writes a closing bracket and new line to it. * * @param fileName the name of the file to save to. * @param removeLast if the last character of the file should be deleted before closing it. */ void finishFile(std::string fileName, bool removeLast); /** * Writes the approximated log normalization constant to file. The file is cleared before the * program starts and if multiple runs is performed, the values from each run is appended to this file. * * @param logNormConstant */ void writeLogNormConstToFile(double logNormConstant); /** * Writes the effective sample size (ESS) to file. The file is cleared before the program * starts. * * @param ess */ void writeESSToFile(std::list<double> essList); #endif
.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 FILE_HANDLER_INCLUDED #define FILE_HANDLER_INCLUDED #include <string> #include <list> const std::string Z_FILE_NAME = "log_norm_const.txt"; const std::string ESS_FILE_NAME = "ess.txt"; /** * Optionally clears the file, and writes an opening bracket to it. * * @param fileName the name of the file to save to. * @param truncate if the file should be cleared first. */ void prepareFile(std::string fileName, bool truncate); /** * Optionally removes the last character of the file, and writes a closing bracket and new line to it. * * @param fileName the name of the file to save to. * @param removeLast if the last character of the file should be deleted before closing it. */ void finishFile(std::string fileName, bool removeLast); /** * Writes the approximated log normalization constant to file. The file is cleared before the * program starts and if multiple runs is performed, the values from each run is appended to this file. * * @param logNormConstant */ void writeLogNormConstToFile(double logNormConstant); /** * Writes the effective sample size (ESS) to file. The file is cleared before the program * starts. * * @param ess */ void writeESSToFile(std::list<double> essList); #endif
.text .file "file_handler.hip" .section .text._ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev,"axG",@progbits,_ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev,comdat .weak _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev # -- Begin function _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev .p2align 4, 0x90 .type _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev,@function _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev: # @_ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev .cfi_startproc # %bb.0: movq %rdi, %rax movq (%rdi), %rdi addq $16, %rax cmpq %rax, %rdi jne _ZdlPv # TAILCALL # %bb.1: # %_ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEE10_M_disposeEv.exit retq .Lfunc_end0: .size _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev, .Lfunc_end0-_ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev .cfi_endproc # -- End function .section .text.startup,"ax",@progbits .p2align 4, 0x90 # -- Begin function _GLOBAL__sub_I_file_handler.hip .type _GLOBAL__sub_I_file_handler.hip,@function _GLOBAL__sub_I_file_handler.hip: # @_GLOBAL__sub_I_file_handler.hip .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movq $_ZL11Z_FILE_NAMEB5cxx11+16, _ZL11Z_FILE_NAMEB5cxx11(%rip) movl $19, %edi callq _Znwm movq %rax, _ZL11Z_FILE_NAMEB5cxx11(%rip) movq $18, _ZL11Z_FILE_NAMEB5cxx11+16(%rip) movups .L.str(%rip), %xmm0 movups %xmm0, (%rax) movw $29816, 16(%rax) # imm = 0x7478 movq $18, _ZL11Z_FILE_NAMEB5cxx11+8(%rip) movb $0, 18(%rax) movl $_ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEED2Ev, %edi movl $_ZL11Z_FILE_NAMEB5cxx11, %esi movl $__dso_handle, %edx popq %rax .cfi_def_cfa_offset 8 jmp __cxa_atexit # TAILCALL .Lfunc_end1: .size _GLOBAL__sub_I_file_handler.hip, .Lfunc_end1-_GLOBAL__sub_I_file_handler.hip .cfi_endproc # -- End function .type _ZL11Z_FILE_NAMEB5cxx11,@object # @_ZL11Z_FILE_NAMEB5cxx11 .local _ZL11Z_FILE_NAMEB5cxx11 .comm _ZL11Z_FILE_NAMEB5cxx11,32,8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "log_norm_const.txt" .size .L.str, 19 .hidden __dso_handle .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad _GLOBAL__sub_I_file_handler.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 __gxx_personality_v0 .addrsig_sym _GLOBAL__sub_I_file_handler.hip .addrsig_sym _ZL11Z_FILE_NAMEB5cxx11 .addrsig_sym __dso_handle .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 CUDA source code to CUDA host assembly.
#include <stdio.h> #include <sys/time.h> #include <time.h> #include <cuda.h> #include <curand.h> #include <curand_kernel.h> const int boardSize = 21; const int totalSize = boardSize * boardSize; struct BoardPoint{ int color; int groupID; int libertyNumber; bool isBlackLegal; bool isWhiteLegal; }; __global__ void randomInit(curandState *state, long randSeed){ int index = threadIdx.y*boardSize + threadIdx.x; //curandState state; // long seed = 123456; curand_init(randSeed, index, 0, &state[index]); // boardPoint[index].color = curand(&state); } __global__ void randomTest(BoardPoint *boardPoint, curandState *state){ int index = threadIdx.y*boardSize + threadIdx.x; // curandState state; // long seed = 123456; // curand_init(seed, index, 0, &state); boardPoint[index].color = (curand(&state[index])>>16)%361; } int main() { BoardPoint boardHost[totalSize]; BoardPoint *boardDevice; curandState *stateDevice; // DebugFlag debugFlagHost[totalSize]; // DebugFlag *debugFlagDevice; // const int valueSizeDevice = totalSize*sizeof(BoardPoint); // const int debugFlagSize = totalSize*sizeof(DebugFlag); // cudaMalloc( (void**)&boardDevice, valueSizeDevice ); cudaMalloc( (void**)&stateDevice, valueSizeDevice ); // cudaMalloc( (void**)&debugFlagDevice, debugFlagSize ); // // dim3 threadShape( boardSize, boardSize ); int numberOfBlock = 1; srand((unsigned int)time(NULL)); randomInit<<<numberOfBlock, threadShape>>>(stateDevice, rand()); struct timeval start_tv; gettimeofday(&start_tv,NULL); randomTest<<<numberOfBlock, threadShape>>>(boardDevice, stateDevice); // for (int i=0; i<19; i++){ // playBoard<<<numberOfBlock, threadShape>>>(boardDevice, globalFlag, i, i, 2); // } cudaDeviceSynchronize(); cudaMemcpy( boardHost, boardDevice, valueSizeDevice, cudaMemcpyDeviceToHost ); // cudaMemcpy( debugFlagHost, debugFlagDevice, debugFlagSize, cudaMemcpyDeviceToHost ); // // cudaFree( boardDevice ); // cudaFree( debugFlagDevice ); // cudaDeviceSynchronize(); // struct timeval end_tv; gettimeofday(&end_tv,NULL); for (int i=boardSize-1; i>=0; i--){ for (int j=0; j<boardSize; j++){ int index = i*boardSize + j; printf("%d| ",boardHost[index].color); } printf("\n"); } if(end_tv.tv_usec >= start_tv.tv_usec){ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec, end_tv.tv_usec - start_tv.tv_usec); }else{ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec - 1, 1000000 - start_tv.tv_usec + end_tv.tv_usec); } return EXIT_SUCCESS; }
.file "tmpxft_0003d616_00000000-6_random_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2274: .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 .LFE2274: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl .type _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl, @function _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl: .LFB2296: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10randomInitP17curandStateXORWOWl(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2296: .size _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl, .-_Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl .globl _Z10randomInitP17curandStateXORWOWl .type _Z10randomInitP17curandStateXORWOWl, @function _Z10randomInitP17curandStateXORWOWl: .LFB2297: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2297: .size _Z10randomInitP17curandStateXORWOWl, .-_Z10randomInitP17curandStateXORWOWl .globl _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW .type _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW, @function _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW: .LFB2298: .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 .L15 .L11: movq 104(%rsp), %rax subq %fs:40, %rax jne .L16 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z10randomTestP10BoardPointP17curandStateXORWOW(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2298: .size _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW, .-_Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW .globl _Z10randomTestP10BoardPointP17curandStateXORWOW .type _Z10randomTestP10BoardPointP17curandStateXORWOW, @function _Z10randomTestP10BoardPointP17curandStateXORWOW: .LFB2299: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2299: .size _Z10randomTestP10BoardPointP17curandStateXORWOW, .-_Z10randomTestP10BoardPointP17curandStateXORWOW .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d| " .LC1: .string "\n" .LC2: .string "time %lu:%lu\n" .text .globl main .type main, @function main: .LFB2271: .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 $4096, %rsp .cfi_def_cfa_offset 4144 orq $0, (%rsp) subq $3040, %rsp .cfi_def_cfa_offset 7184 movq %fs:40, %rax movq %rax, 7128(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $7056, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $7056, %esi call cudaMalloc@PLT movl $21, 20(%rsp) movl $21, 24(%rsp) movl $1, 28(%rsp) movl $0, %edi call time@PLT movl %eax, %edi call srand@PLT movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L30 .L20: leaq 32(%rsp), %rdi movl $0, %esi call gettimeofday@PLT movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L31 .L21: call cudaDeviceSynchronize@PLT leaq 64(%rsp), %r13 movl $2, %ecx movl $7056, %edx movq (%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT movq (%rsp), %rdi call cudaFree@PLT call cudaDeviceSynchronize@PLT leaq 48(%rsp), %rdi movl $0, %esi call gettimeofday@PLT leaq 7120(%rsp), %rbp leaq .LC0(%rip), %r12 leaq .LC1(%rip), %r14 .L22: leaq -336(%rbp), %rbx .L23: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $16, %rbx cmpq %rbp, %rbx jne .L23 movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT subq $336, %rbp cmpq %r13, %rbp jne .L22 movq 56(%rsp), %rax movq 40(%rsp), %rdx cmpq %rdx, %rax jl .L25 subq %rdx, %rax movq %rax, %rcx movq 48(%rsp), %rdx subq 32(%rsp), %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L26: movq 7128(%rsp), %rax subq %fs:40, %rax jne .L32 movl $0, %eax addq $7136, %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 .L30: .cfi_restore_state call rand@PLT movslq %eax, %rsi movq 8(%rsp), %rdi call _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl jmp .L20 .L31: movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW jmp .L21 .L25: subq %rdx, %rax leaq 1000000(%rax), %rcx movq 48(%rsp), %rdx subq 32(%rsp), %rdx subq $1, %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L26 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2271: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "_Z10randomTestP10BoardPointP17curandStateXORWOW" .align 8 .LC4: .string "_Z10randomInitP17curandStateXORWOWl" .section .rodata.str1.1 .LC5: .string "precalc_xorwow_matrix" .LC6: .string "precalc_xorwow_offset_matrix" .LC7: .string "mrg32k3aM1" .LC8: .string "mrg32k3aM2" .LC9: .string "mrg32k3aM1SubSeq" .LC10: .string "mrg32k3aM2SubSeq" .LC11: .string "mrg32k3aM1Seq" .LC12: .string "mrg32k3aM2Seq" .LC13: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2301: .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z10randomTestP10BoardPointP17curandStateXORWOW(%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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z10randomInitP17curandStateXORWOWl(%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 .LC5(%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 .LC6(%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 .LC7(%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 .LC8(%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 .LC9(%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 .LC10(%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 .LC11(%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 .LC12(%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 .LC13(%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 .LFE2301: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <sys/time.h> #include <time.h> #include <cuda.h> #include <curand.h> #include <curand_kernel.h> const int boardSize = 21; const int totalSize = boardSize * boardSize; struct BoardPoint{ int color; int groupID; int libertyNumber; bool isBlackLegal; bool isWhiteLegal; }; __global__ void randomInit(curandState *state, long randSeed){ int index = threadIdx.y*boardSize + threadIdx.x; //curandState state; // long seed = 123456; curand_init(randSeed, index, 0, &state[index]); // boardPoint[index].color = curand(&state); } __global__ void randomTest(BoardPoint *boardPoint, curandState *state){ int index = threadIdx.y*boardSize + threadIdx.x; // curandState state; // long seed = 123456; // curand_init(seed, index, 0, &state); boardPoint[index].color = (curand(&state[index])>>16)%361; } int main() { BoardPoint boardHost[totalSize]; BoardPoint *boardDevice; curandState *stateDevice; // DebugFlag debugFlagHost[totalSize]; // DebugFlag *debugFlagDevice; // const int valueSizeDevice = totalSize*sizeof(BoardPoint); // const int debugFlagSize = totalSize*sizeof(DebugFlag); // cudaMalloc( (void**)&boardDevice, valueSizeDevice ); cudaMalloc( (void**)&stateDevice, valueSizeDevice ); // cudaMalloc( (void**)&debugFlagDevice, debugFlagSize ); // // dim3 threadShape( boardSize, boardSize ); int numberOfBlock = 1; srand((unsigned int)time(NULL)); randomInit<<<numberOfBlock, threadShape>>>(stateDevice, rand()); struct timeval start_tv; gettimeofday(&start_tv,NULL); randomTest<<<numberOfBlock, threadShape>>>(boardDevice, stateDevice); // for (int i=0; i<19; i++){ // playBoard<<<numberOfBlock, threadShape>>>(boardDevice, globalFlag, i, i, 2); // } cudaDeviceSynchronize(); cudaMemcpy( boardHost, boardDevice, valueSizeDevice, cudaMemcpyDeviceToHost ); // cudaMemcpy( debugFlagHost, debugFlagDevice, debugFlagSize, cudaMemcpyDeviceToHost ); // // cudaFree( boardDevice ); // cudaFree( debugFlagDevice ); // cudaDeviceSynchronize(); // struct timeval end_tv; gettimeofday(&end_tv,NULL); for (int i=boardSize-1; i>=0; i--){ for (int j=0; j<boardSize; j++){ int index = i*boardSize + j; printf("%d| ",boardHost[index].color); } printf("\n"); } if(end_tv.tv_usec >= start_tv.tv_usec){ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec, end_tv.tv_usec - start_tv.tv_usec); }else{ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec - 1, 1000000 - start_tv.tv_usec + end_tv.tv_usec); } return EXIT_SUCCESS; }
#include <stdio.h> #include <sys/time.h> #include <time.h> #include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> const int boardSize = 21; const int totalSize = boardSize * boardSize; struct BoardPoint{ int color; int groupID; int libertyNumber; bool isBlackLegal; bool isWhiteLegal; }; __global__ void randomInit(hiprandState *state, long randSeed){ int index = threadIdx.y*boardSize + threadIdx.x; //curandState state; // long seed = 123456; hiprand_init(randSeed, index, 0, &state[index]); // boardPoint[index].color = curand(&state); } __global__ void randomTest(BoardPoint *boardPoint, hiprandState *state){ int index = threadIdx.y*boardSize + threadIdx.x; // curandState state; // long seed = 123456; // curand_init(seed, index, 0, &state); boardPoint[index].color = (hiprand(&state[index])>>16)%361; } int main() { BoardPoint boardHost[totalSize]; BoardPoint *boardDevice; hiprandState *stateDevice; // DebugFlag debugFlagHost[totalSize]; // DebugFlag *debugFlagDevice; // const int valueSizeDevice = totalSize*sizeof(BoardPoint); // const int debugFlagSize = totalSize*sizeof(DebugFlag); // hipMalloc( (void**)&boardDevice, valueSizeDevice ); hipMalloc( (void**)&stateDevice, valueSizeDevice ); // cudaMalloc( (void**)&debugFlagDevice, debugFlagSize ); // // dim3 threadShape( boardSize, boardSize ); int numberOfBlock = 1; srand((unsigned int)time(NULL)); randomInit<<<numberOfBlock, threadShape>>>(stateDevice, rand()); struct timeval start_tv; gettimeofday(&start_tv,NULL); randomTest<<<numberOfBlock, threadShape>>>(boardDevice, stateDevice); // for (int i=0; i<19; i++){ // playBoard<<<numberOfBlock, threadShape>>>(boardDevice, globalFlag, i, i, 2); // } hipDeviceSynchronize(); hipMemcpy( boardHost, boardDevice, valueSizeDevice, hipMemcpyDeviceToHost ); // cudaMemcpy( debugFlagHost, debugFlagDevice, debugFlagSize, cudaMemcpyDeviceToHost ); // // hipFree( boardDevice ); // cudaFree( debugFlagDevice ); // hipDeviceSynchronize(); // struct timeval end_tv; gettimeofday(&end_tv,NULL); for (int i=boardSize-1; i>=0; i--){ for (int j=0; j<boardSize; j++){ int index = i*boardSize + j; printf("%d| ",boardHost[index].color); } printf("\n"); } if(end_tv.tv_usec >= start_tv.tv_usec){ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec, end_tv.tv_usec - start_tv.tv_usec); }else{ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec - 1, 1000000 - start_tv.tv_usec + end_tv.tv_usec); } return EXIT_SUCCESS; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <stdio.h> #include <sys/time.h> #include <time.h> #include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #include <hiprand/hiprand_kernel.h> const int boardSize = 21; const int totalSize = boardSize * boardSize; struct BoardPoint{ int color; int groupID; int libertyNumber; bool isBlackLegal; bool isWhiteLegal; }; __global__ void randomInit(hiprandState *state, long randSeed){ int index = threadIdx.y*boardSize + threadIdx.x; //curandState state; // long seed = 123456; hiprand_init(randSeed, index, 0, &state[index]); // boardPoint[index].color = curand(&state); } __global__ void randomTest(BoardPoint *boardPoint, hiprandState *state){ int index = threadIdx.y*boardSize + threadIdx.x; // curandState state; // long seed = 123456; // curand_init(seed, index, 0, &state); boardPoint[index].color = (hiprand(&state[index])>>16)%361; } int main() { BoardPoint boardHost[totalSize]; BoardPoint *boardDevice; hiprandState *stateDevice; // DebugFlag debugFlagHost[totalSize]; // DebugFlag *debugFlagDevice; // const int valueSizeDevice = totalSize*sizeof(BoardPoint); // const int debugFlagSize = totalSize*sizeof(DebugFlag); // hipMalloc( (void**)&boardDevice, valueSizeDevice ); hipMalloc( (void**)&stateDevice, valueSizeDevice ); // cudaMalloc( (void**)&debugFlagDevice, debugFlagSize ); // // dim3 threadShape( boardSize, boardSize ); int numberOfBlock = 1; srand((unsigned int)time(NULL)); randomInit<<<numberOfBlock, threadShape>>>(stateDevice, rand()); struct timeval start_tv; gettimeofday(&start_tv,NULL); randomTest<<<numberOfBlock, threadShape>>>(boardDevice, stateDevice); // for (int i=0; i<19; i++){ // playBoard<<<numberOfBlock, threadShape>>>(boardDevice, globalFlag, i, i, 2); // } hipDeviceSynchronize(); hipMemcpy( boardHost, boardDevice, valueSizeDevice, hipMemcpyDeviceToHost ); // cudaMemcpy( debugFlagHost, debugFlagDevice, debugFlagSize, cudaMemcpyDeviceToHost ); // // hipFree( boardDevice ); // cudaFree( debugFlagDevice ); // hipDeviceSynchronize(); // struct timeval end_tv; gettimeofday(&end_tv,NULL); for (int i=boardSize-1; i>=0; i--){ for (int j=0; j<boardSize; j++){ int index = i*boardSize + j; printf("%d| ",boardHost[index].color); } printf("\n"); } if(end_tv.tv_usec >= start_tv.tv_usec){ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec, end_tv.tv_usec - start_tv.tv_usec); }else{ printf("time %lu:%lu\n",end_tv.tv_sec - start_tv.tv_sec - 1, 1000000 - start_tv.tv_usec + end_tv.tv_usec); } return EXIT_SUCCESS; }
.text .file "random_test.hip" .globl _Z25__device_stub__randomInitP12hiprandStatel # -- Begin function _Z25__device_stub__randomInitP12hiprandStatel .p2align 4, 0x90 .type _Z25__device_stub__randomInitP12hiprandStatel,@function _Z25__device_stub__randomInitP12hiprandStatel: # @_Z25__device_stub__randomInitP12hiprandStatel .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10randomInitP12hiprandStatel, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z25__device_stub__randomInitP12hiprandStatel, .Lfunc_end0-_Z25__device_stub__randomInitP12hiprandStatel .cfi_endproc # -- End function .globl _Z25__device_stub__randomTestP10BoardPointP12hiprandState # -- Begin function _Z25__device_stub__randomTestP10BoardPointP12hiprandState .p2align 4, 0x90 .type _Z25__device_stub__randomTestP10BoardPointP12hiprandState,@function _Z25__device_stub__randomTestP10BoardPointP12hiprandState: # @_Z25__device_stub__randomTestP10BoardPointP12hiprandState .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10randomTestP10BoardPointP12hiprandState, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end1: .size _Z25__device_stub__randomTestP10BoardPointP12hiprandState, .Lfunc_end1-_Z25__device_stub__randomTestP10BoardPointP12hiprandState .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 $7152, %rsp # imm = 0x1BF0 .cfi_def_cfa_offset 7184 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movabsq $90194313237, %rbx # imm = 0x1500000015 movabsq $4294967297, %r14 # imm = 0x100000001 leaq 16(%rsp), %rdi movl $7056, %esi # imm = 0x1B90 callq hipMalloc leaq 80(%rsp), %rdi movl $7056, %esi # imm = 0x1B90 callq hipMalloc xorl %edi, %edi callq time movl %eax, %edi callq srand movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movq 80(%rsp), %r15 callq rand cltq movq %r15, 48(%rsp) movq %rax, 40(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rax movq %rax, 104(%rsp) leaq 64(%rsp), %rdi movq %rsp, %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq (%rsp), %rcx movl 8(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z10randomInitP12hiprandStatel, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_2: leaq 64(%rsp), %rdi xorl %esi, %esi callq gettimeofday movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 16(%rsp), %rax movq 80(%rsp), %rcx movq %rax, 40(%rsp) movq %rcx, 32(%rsp) leaq 40(%rsp), %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rax movq %rax, 104(%rsp) movq %rsp, %rdi leaq 48(%rsp), %rsi leaq 24(%rsp), %rdx leaq 88(%rsp), %rcx callq __hipPopCallConfiguration movq (%rsp), %rsi movl 8(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z10randomTestP10BoardPointP12hiprandState, %edi pushq 88(%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_4: callq hipDeviceSynchronize movq 16(%rsp), %rsi leaq 96(%rsp), %rdi movl $7056, %edx # imm = 0x1B90 movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree callq hipDeviceSynchronize movq %rsp, %rdi xorl %esi, %esi callq gettimeofday movl $20, %ebx leaq 6816(%rsp), %r14 .p2align 4, 0x90 .LBB2_5: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB2_6 Depth 2 xorl %r15d, %r15d .p2align 4, 0x90 .LBB2_6: # Parent Loop BB2_5 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r14,%r15), %esi movl $.L.str, %edi xorl %eax, %eax callq printf addq $16, %r15 cmpq $336, %r15 # imm = 0x150 jne .LBB2_6 # %bb.7: # in Loop: Header=BB2_5 Depth=1 movl $10, %edi callq putchar@PLT addq $-336, %r14 # imm = 0xFEB0 subq $1, %rbx jae .LBB2_5 # %bb.8: movq 8(%rsp), %rax movq 64(%rsp), %rsi movq 72(%rsp), %rcx leaq 1000000(%rax), %rdx movq %rsi, %rdi negq %rdi cmpq %rcx, %rax notq %rsi cmovgeq %rax, %rdx cmovgeq %rdi, %rsi addq (%rsp), %rsi subq %rcx, %rdx movl $.L.str.2, %edi xorl %eax, %eax callq printf xorl %eax, %eax addq $7152, %rsp # imm = 0x1BF0 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: 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 $_Z10randomInitP12hiprandStatel, %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 $_Z10randomTestP10BoardPointP12hiprandState, %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 _Z10randomInitP12hiprandStatel,@object # @_Z10randomInitP12hiprandStatel .section .rodata,"a",@progbits .globl _Z10randomInitP12hiprandStatel .p2align 3, 0x0 _Z10randomInitP12hiprandStatel: .quad _Z25__device_stub__randomInitP12hiprandStatel .size _Z10randomInitP12hiprandStatel, 8 .type _Z10randomTestP10BoardPointP12hiprandState,@object # @_Z10randomTestP10BoardPointP12hiprandState .globl _Z10randomTestP10BoardPointP12hiprandState .p2align 3, 0x0 _Z10randomTestP10BoardPointP12hiprandState: .quad _Z25__device_stub__randomTestP10BoardPointP12hiprandState .size _Z10randomTestP10BoardPointP12hiprandState, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d| " .size .L.str, 5 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "time %lu:%lu\n" .size .L.str.2, 14 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10randomInitP12hiprandStatel" .size .L__unnamed_1, 31 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z10randomTestP10BoardPointP12hiprandState" .size .L__unnamed_2, 43 .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__randomInitP12hiprandStatel .addrsig_sym _Z25__device_stub__randomTestP10BoardPointP12hiprandState .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10randomInitP12hiprandStatel .addrsig_sym _Z10randomTestP10BoardPointP12hiprandState .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_0003d616_00000000-6_random_test.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2274: .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 .LFE2274: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl .type _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl, @function _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl: .LFB2296: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10randomInitP17curandStateXORWOWl(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2296: .size _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl, .-_Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl .globl _Z10randomInitP17curandStateXORWOWl .type _Z10randomInitP17curandStateXORWOWl, @function _Z10randomInitP17curandStateXORWOWl: .LFB2297: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2297: .size _Z10randomInitP17curandStateXORWOWl, .-_Z10randomInitP17curandStateXORWOWl .globl _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW .type _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW, @function _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW: .LFB2298: .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 .L15 .L11: movq 104(%rsp), %rax subq %fs:40, %rax jne .L16 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .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 _Z10randomTestP10BoardPointP17curandStateXORWOW(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2298: .size _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW, .-_Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW .globl _Z10randomTestP10BoardPointP17curandStateXORWOW .type _Z10randomTestP10BoardPointP17curandStateXORWOW, @function _Z10randomTestP10BoardPointP17curandStateXORWOW: .LFB2299: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2299: .size _Z10randomTestP10BoardPointP17curandStateXORWOW, .-_Z10randomTestP10BoardPointP17curandStateXORWOW .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d| " .LC1: .string "\n" .LC2: .string "time %lu:%lu\n" .text .globl main .type main, @function main: .LFB2271: .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 $4096, %rsp .cfi_def_cfa_offset 4144 orq $0, (%rsp) subq $3040, %rsp .cfi_def_cfa_offset 7184 movq %fs:40, %rax movq %rax, 7128(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $7056, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $7056, %esi call cudaMalloc@PLT movl $21, 20(%rsp) movl $21, 24(%rsp) movl $1, 28(%rsp) movl $0, %edi call time@PLT movl %eax, %edi call srand@PLT movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L30 .L20: leaq 32(%rsp), %rdi movl $0, %esi call gettimeofday@PLT movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L31 .L21: call cudaDeviceSynchronize@PLT leaq 64(%rsp), %r13 movl $2, %ecx movl $7056, %edx movq (%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT movq (%rsp), %rdi call cudaFree@PLT call cudaDeviceSynchronize@PLT leaq 48(%rsp), %rdi movl $0, %esi call gettimeofday@PLT leaq 7120(%rsp), %rbp leaq .LC0(%rip), %r12 leaq .LC1(%rip), %r14 .L22: leaq -336(%rbp), %rbx .L23: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $16, %rbx cmpq %rbp, %rbx jne .L23 movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT subq $336, %rbp cmpq %r13, %rbp jne .L22 movq 56(%rsp), %rax movq 40(%rsp), %rdx cmpq %rdx, %rax jl .L25 subq %rdx, %rax movq %rax, %rcx movq 48(%rsp), %rdx subq 32(%rsp), %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L26: movq 7128(%rsp), %rax subq %fs:40, %rax jne .L32 movl $0, %eax addq $7136, %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 .L30: .cfi_restore_state call rand@PLT movslq %eax, %rsi movq 8(%rsp), %rdi call _Z49__device_stub__Z10randomInitP17curandStateXORWOWlP17curandStateXORWOWl jmp .L20 .L31: movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z61__device_stub__Z10randomTestP10BoardPointP17curandStateXORWOWP10BoardPointP17curandStateXORWOW jmp .L21 .L25: subq %rdx, %rax leaq 1000000(%rax), %rcx movq 48(%rsp), %rdx subq 32(%rsp), %rdx subq $1, %rdx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L26 .L32: call __stack_chk_fail@PLT .cfi_endproc .LFE2271: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "_Z10randomTestP10BoardPointP17curandStateXORWOW" .align 8 .LC4: .string "_Z10randomInitP17curandStateXORWOWl" .section .rodata.str1.1 .LC5: .string "precalc_xorwow_matrix" .LC6: .string "precalc_xorwow_offset_matrix" .LC7: .string "mrg32k3aM1" .LC8: .string "mrg32k3aM2" .LC9: .string "mrg32k3aM1SubSeq" .LC10: .string "mrg32k3aM2SubSeq" .LC11: .string "mrg32k3aM1Seq" .LC12: .string "mrg32k3aM2Seq" .LC13: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2301: .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z10randomTestP10BoardPointP17curandStateXORWOW(%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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z10randomInitP17curandStateXORWOWl(%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 .LC5(%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 .LC6(%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 .LC7(%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 .LC8(%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 .LC9(%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 .LC10(%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 .LC11(%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 .LC12(%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 .LC13(%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 .LFE2301: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "random_test.hip" .globl _Z25__device_stub__randomInitP12hiprandStatel # -- Begin function _Z25__device_stub__randomInitP12hiprandStatel .p2align 4, 0x90 .type _Z25__device_stub__randomInitP12hiprandStatel,@function _Z25__device_stub__randomInitP12hiprandStatel: # @_Z25__device_stub__randomInitP12hiprandStatel .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10randomInitP12hiprandStatel, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z25__device_stub__randomInitP12hiprandStatel, .Lfunc_end0-_Z25__device_stub__randomInitP12hiprandStatel .cfi_endproc # -- End function .globl _Z25__device_stub__randomTestP10BoardPointP12hiprandState # -- Begin function _Z25__device_stub__randomTestP10BoardPointP12hiprandState .p2align 4, 0x90 .type _Z25__device_stub__randomTestP10BoardPointP12hiprandState,@function _Z25__device_stub__randomTestP10BoardPointP12hiprandState: # @_Z25__device_stub__randomTestP10BoardPointP12hiprandState .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10randomTestP10BoardPointP12hiprandState, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end1: .size _Z25__device_stub__randomTestP10BoardPointP12hiprandState, .Lfunc_end1-_Z25__device_stub__randomTestP10BoardPointP12hiprandState .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 $7152, %rsp # imm = 0x1BF0 .cfi_def_cfa_offset 7184 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movabsq $90194313237, %rbx # imm = 0x1500000015 movabsq $4294967297, %r14 # imm = 0x100000001 leaq 16(%rsp), %rdi movl $7056, %esi # imm = 0x1B90 callq hipMalloc leaq 80(%rsp), %rdi movl $7056, %esi # imm = 0x1B90 callq hipMalloc xorl %edi, %edi callq time movl %eax, %edi callq srand movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_2 # %bb.1: movq 80(%rsp), %r15 callq rand cltq movq %r15, 48(%rsp) movq %rax, 40(%rsp) leaq 48(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rax movq %rax, 104(%rsp) leaq 64(%rsp), %rdi movq %rsp, %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq (%rsp), %rcx movl 8(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z10randomInitP12hiprandStatel, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_2: leaq 64(%rsp), %rdi xorl %esi, %esi callq gettimeofday movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 16(%rsp), %rax movq 80(%rsp), %rcx movq %rax, 40(%rsp) movq %rcx, 32(%rsp) leaq 40(%rsp), %rax movq %rax, 96(%rsp) leaq 32(%rsp), %rax movq %rax, 104(%rsp) movq %rsp, %rdi leaq 48(%rsp), %rsi leaq 24(%rsp), %rdx leaq 88(%rsp), %rcx callq __hipPopCallConfiguration movq (%rsp), %rsi movl 8(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z10randomTestP10BoardPointP12hiprandState, %edi pushq 88(%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_4: callq hipDeviceSynchronize movq 16(%rsp), %rsi leaq 96(%rsp), %rdi movl $7056, %edx # imm = 0x1B90 movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree callq hipDeviceSynchronize movq %rsp, %rdi xorl %esi, %esi callq gettimeofday movl $20, %ebx leaq 6816(%rsp), %r14 .p2align 4, 0x90 .LBB2_5: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB2_6 Depth 2 xorl %r15d, %r15d .p2align 4, 0x90 .LBB2_6: # Parent Loop BB2_5 Depth=1 # => This Inner Loop Header: Depth=2 movl (%r14,%r15), %esi movl $.L.str, %edi xorl %eax, %eax callq printf addq $16, %r15 cmpq $336, %r15 # imm = 0x150 jne .LBB2_6 # %bb.7: # in Loop: Header=BB2_5 Depth=1 movl $10, %edi callq putchar@PLT addq $-336, %r14 # imm = 0xFEB0 subq $1, %rbx jae .LBB2_5 # %bb.8: movq 8(%rsp), %rax movq 64(%rsp), %rsi movq 72(%rsp), %rcx leaq 1000000(%rax), %rdx movq %rsi, %rdi negq %rdi cmpq %rcx, %rax notq %rsi cmovgeq %rax, %rdx cmovgeq %rdi, %rsi addq (%rsp), %rsi subq %rcx, %rdx movl $.L.str.2, %edi xorl %eax, %eax callq printf xorl %eax, %eax addq $7152, %rsp # imm = 0x1BF0 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: 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 $_Z10randomInitP12hiprandStatel, %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 $_Z10randomTestP10BoardPointP12hiprandState, %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 _Z10randomInitP12hiprandStatel,@object # @_Z10randomInitP12hiprandStatel .section .rodata,"a",@progbits .globl _Z10randomInitP12hiprandStatel .p2align 3, 0x0 _Z10randomInitP12hiprandStatel: .quad _Z25__device_stub__randomInitP12hiprandStatel .size _Z10randomInitP12hiprandStatel, 8 .type _Z10randomTestP10BoardPointP12hiprandState,@object # @_Z10randomTestP10BoardPointP12hiprandState .globl _Z10randomTestP10BoardPointP12hiprandState .p2align 3, 0x0 _Z10randomTestP10BoardPointP12hiprandState: .quad _Z25__device_stub__randomTestP10BoardPointP12hiprandState .size _Z10randomTestP10BoardPointP12hiprandState, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "%d| " .size .L.str, 5 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "time %lu:%lu\n" .size .L.str.2, 14 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10randomInitP12hiprandStatel" .size .L__unnamed_1, 31 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z10randomTestP10BoardPointP12hiprandState" .size .L__unnamed_2, 43 .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__randomInitP12hiprandStatel .addrsig_sym _Z25__device_stub__randomTestP10BoardPointP12hiprandState .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10randomInitP12hiprandStatel .addrsig_sym _Z10randomTestP10BoardPointP12hiprandState .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 <cmath> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
code for sm_80 Function : _Z11my_copysignPd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_TID.X ; /* 0x0000000000087919 */ /* 0x000e220000002100 */ /*0020*/ BSSY B0, 0xd0 ; /* 0x000000a000007945 */ /* 0x000fe20003800000 */ /*0030*/ I2F.F64 R2, R8 ; /* 0x0000000800027312 */ /* 0x001e240000201c00 */ /*0040*/ LOP3.LUT R0, R3, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000003007812 */ /* 0x001fc800078ec0ff */ /*0050*/ LEA.HI R5, R0, 0xfffffc0c, RZ, 0xc ; /* 0xfffffc0c00057811 */ /* 0x000fc800078f60ff */ /*0060*/ SHF.L.U32 R0, R2.reuse, R5.reuse, RZ ; /* 0x0000000502007219 */ /* 0x0c0fe400000006ff */ /*0070*/ SHF.L.U64.HI R5, R2, R5, R3 ; /* 0x0000000502057219 */ /* 0x000fe40000010203 */ /*0080*/ ISETP.NE.U32.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe40003f05070 */ /*0090*/ MOV R0, 0xc0 ; /* 0x000000c000007802 */ /* 0x000fe40000000f00 */ /*00a0*/ ISETP.NE.AND.EX P0, PT, R5, -0x80000000, PT, P0 ; /* 0x800000000500780c */ /* 0x000fd00003f05300 */ /*00b0*/ CALL.REL.NOINC 0x2f0 ; /* 0x0000023000007944 */ /* 0x000fea0003c00000 */ /*00c0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*00d0*/ FRND.F64.TRUNC R4, R2 ; /* 0x0000000200047313 */ /* 0x000e62000030d800 */ /*00e0*/ DADD R6, R2, -1 ; /* 0xbff0000002067429 */ /* 0x000ea20000000000 */ /*00f0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0100*/ BSSY B0, 0x260 ; /* 0x0000015000007945 */ /* 0x000ff00003800000 */ /*0110*/ LOP3.LUT R6, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000007067812 */ /* 0x004fc800078ec0ff */ /*0120*/ ISETP.NE.AND P2, PT, R6, 0x7ff00000, PT ; /* 0x7ff000000600780c */ /* 0x000fe20003f45270 */ /*0130*/ DSETP.NEU.AND P1, PT, R4, R2, PT ; /* 0x000000020400722a */ /* 0x0022a40003f2d000 */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, R15 ; /* 0x000000ffff057224 */ /* 0x002fe400078e000f */ /*0150*/ IMAD.MOV.U32 R4, RZ, RZ, R14 ; /* 0x000000ffff047224 */ /* 0x000fc600078e000e */ /*0160*/ @!P0 LOP3.LUT R7, R5, 0x80000000, RZ, 0x3c, !PT ; /* 0x8000000005078812 */ /* 0x000fca00078e3cff */ /*0170*/ @!P0 IMAD.MOV.U32 R5, RZ, RZ, R7 ; /* 0x000000ffff058224 */ /* 0x000fe400078e0007 */ /*0180*/ @P1 IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff041424 */ /* 0x004fe400078e00ff */ /*0190*/ @P1 IMAD.MOV.U32 R5, RZ, RZ, -0x80000 ; /* 0xfff80000ff051424 */ /* 0x000fe200078e00ff */ /*01a0*/ @P2 BRA 0x250 ; /* 0x000000a000002947 */ /* 0x000fea0003800000 */ /*01b0*/ DSETP.GTU.AND P0, PT, |R2|, +INF , PT ; /* 0x7ff000000200742a */ /* 0x000e5c0003f0c200 */ /*01c0*/ @P0 BRA 0x240 ; /* 0x0000007000000947 */ /* 0x002fea0003800000 */ /*01d0*/ ISETP.NE.AND P0, PT, R2, RZ, PT ; /* 0x000000ff0200720c */ /* 0x000fe40003f05270 */ /*01e0*/ LOP3.LUT R2, R3, 0x7fffffff, RZ, 0xc0, !PT ; /* 0x7fffffff03027812 */ /* 0x000fc800078ec0ff */ /*01f0*/ ISETP.EQ.AND P0, PT, R2, 0x7ff00000, !P0 ; /* 0x7ff000000200780c */ /* 0x000fda0004702270 */ /*0200*/ @!P0 BRA 0x250 ; /* 0x0000004000008947 */ /* 0x000fea0003800000 */ /*0210*/ IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff047424 */ /* 0x000fe400078e00ff */ /*0220*/ IMAD.MOV.U32 R5, RZ, RZ, 0x3ff00000 ; /* 0x3ff00000ff057424 */ /* 0x000fe200078e00ff */ /*0230*/ BRA 0x250 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*0240*/ DADD R4, R2, -1 ; /* 0xbff0000002047429 */ /* 0x00028c0000000000 */ /*0250*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0260*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x160] ; /* 0x00005800ff027624 */ /* 0x002fe400078e00ff */ /*0270*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff037624 */ /* 0x000fca00078e00ff */ /*0280*/ LDG.E.64 R6, [R2.64] ; /* 0x0000000402067981 */ /* 0x000ee2000c1e1b00 */ /*0290*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fc80003f05270 */ /*02a0*/ FSEL R4, R4, RZ, P0 ; /* 0x000000ff04047208 */ /* 0x004fe40000000000 */ /*02b0*/ FSEL R5, R5, 1.875, P0 ; /* 0x3ff0000005057808 */ /* 0x000fcc0000000000 */ /*02c0*/ DMUL R6, R6, R4 ; /* 0x0000000406067228 */ /* 0x008e4e0000000000 */ /*02d0*/ STG.E.64 [R2.64], R6 ; /* 0x0000000602007986 */ /* 0x002fe2000c101b04 */ /*02e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02f0*/ MUFU.RCP64H R11, 2 ; /* 0x40000000000b7908 */ /* 0x000e220000001800 */ /*0300*/ IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff047424 */ /* 0x000fe400078e00ff */ /*0310*/ IMAD.MOV.U32 R5, RZ, RZ, 0x3ff00000 ; /* 0x3ff00000ff057424 */ /* 0x000fe400078e00ff */ /*0320*/ IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e00ff */ /*0330*/ IMAD.MOV.U32 R6, RZ, RZ, 0x7d2cafe2 ; /* 0x7d2cafe2ff067424 */ /* 0x000fe400078e00ff */ /*0340*/ IMAD.MOV.U32 R7, RZ, RZ, 0x3eb0f5ff ; /* 0x3eb0f5ffff077424 */ /* 0x000fc400078e00ff */ /*0350*/ IMAD.SHL.U32 R9, R3, 0x2, RZ ; /* 0x0000000203097824 */ /* 0x000fe200078e00ff */ /*0360*/ DFMA R4, R10, -2, R4 ; /* 0xc00000000a04782b */ /* 0x001e080000000004 */ /*0370*/ ISETP.GT.U32.AND P1, PT, R9, -0x2000001, PT ; /* 0xfdffffff0900780c */ /* 0x000fe40003f24070 */ /*0380*/ DFMA R4, R4, R4, R4 ; /* 0x000000040404722b */ /* 0x001e0c0000000004 */ /*0390*/ DFMA R10, R10, R4, R10 ; /* 0x000000040a0a722b */ /* 0x001e0c000000000a */ /*03a0*/ DMUL R4, RZ, R10 ; /* 0x0000000aff047228 */ /* 0x001e0c0000000000 */ /*03b0*/ DFMA R4, RZ, R10, R4 ; /* 0x0000000aff04722b */ /* 0x001e0c0000000004 */ /*03c0*/ DMUL R12, R4, R4 ; /* 0x00000004040c7228 */ /* 0x001e080000000000 */ /*03d0*/ DADD R14, RZ, -R4 ; /* 0x00000000ff0e7229 */ /* 0x000e480000000804 */ /*03e0*/ DFMA R6, R12, R6, c[0x2][0x0] ; /* 0x008000000c06762b */ /* 0x001e080000000006 */ /*03f0*/ DADD R16, R14, R14 ; /* 0x000000000e107229 */ /* 0x002fc8000000000e */ /*0400*/ DFMA R6, R12, R6, c[0x2][0x8] ; /* 0x008002000c06762b */ /* 0x001e080000000006 */ /*0410*/ DMUL R14, R4, R4 ; /* 0x00000004040e7228 */ /* 0x000fc80000000000 */ /*0420*/ DFMA R6, R12, R6, c[0x2][0x10] ; /* 0x008004000c06762b */ /* 0x001e080000000006 */ /*0430*/ DFMA R20, RZ, -R4, R16 ; /* 0x80000004ff14722b */ /* 0x000e480000000010 */ /*0440*/ DFMA R6, R12, R6, c[0x2][0x18] ; /* 0x008006000c06762b */ /* 0x001e080000000006 */ /*0450*/ DMUL R10, R10, R20 ; /* 0x000000140a0a7228 */ /* 0x002fc80000000000 */ /*0460*/ DFMA R6, R12, R6, c[0x2][0x20] ; /* 0x008008000c06762b */ /* 0x001e080000000006 */ /*0470*/ DFMA R16, R4, R4, -R14 ; /* 0x000000040410722b */ /* 0x000fc8000000080e */ /*0480*/ DFMA R18, R12, R6, c[0x2][0x28] ; /* 0x00800a000c12762b */ /* 0x001e0c0000000006 */ /*0490*/ DFMA R6, R12, R18, c[0x2][0x30] ; /* 0x00800c000c06762b */ /* 0x001e0c0000000012 */ /*04a0*/ DADD R22, -R6, c[0x2][0x30] ; /* 0x00800c0006167629 */ /* 0x001e0c0000000100 */ /*04b0*/ DFMA R22, R12, R18, R22 ; /* 0x000000120c16722b */ /* 0x001e080000000016 */ /*04c0*/ DMUL R12, R4, R14 ; /* 0x0000000e040c7228 */ /* 0x000e480000000000 */ /*04d0*/ DADD R18, RZ, R22 ; /* 0x00000000ff127229 */ /* 0x001e080000000016 */ /*04e0*/ DFMA R20, R4, R14, -R12 ; /* 0x0000000e0414722b */ /* 0x002e48000000080c */ /*04f0*/ DADD R18, R18, c[0x2][0x38] ; /* 0x00800e0012127629 */ /* 0x001e080000000000 */ /*0500*/ DFMA R22, R10, R14, R20 ; /* 0x0000000e0a16722b */ /* 0x0023e40000000014 */ /*0510*/ IADD3 R21, R11, 0x100000, RZ ; /* 0x001000000b157810 */ /* 0x002fe20007ffe0ff */ /*0520*/ IMAD.MOV.U32 R20, RZ, RZ, R10 ; /* 0x000000ffff147224 */ /* 0x000fe200078e000a */ /*0530*/ DADD R14, R6, R18 ; /* 0x00000000060e7229 */ /* 0x001e0a0000000012 */ /*0540*/ DFMA R20, R4, R20, R16 ; /* 0x000000140414722b */ /* 0x000e480000000010 */ /*0550*/ DMUL R16, R14, R12 ; /* 0x0000000c0e107228 */ /* 0x001e080000000000 */ /*0560*/ DFMA R20, R4, R20, R22 ; /* 0x000000140414722b */ /* 0x002fc80000000016 */ /*0570*/ DADD R6, R6, -R14 ; /* 0x0000000006067229 */ /* 0x000e48000000080e */ /*0580*/ DFMA R22, R14, R12, -R16 ; /* 0x0000000c0e16722b */ /* 0x001e080000000810 */ /*0590*/ DADD R6, R18, R6 ; /* 0x0000000012067229 */ /* 0x0023e40000000006 */ /*05a0*/ IMAD.MOV.U32 R18, RZ, RZ, 0x69ce2bdf ; /* 0x69ce2bdfff127424 */ /* 0x002fe400078e00ff */ /*05b0*/ DFMA R20, R14, R20, R22 ; /* 0x000000140e14722b */ /* 0x001e220000000016 */ /*05c0*/ IMAD.MOV.U32 R19, RZ, RZ, 0x3e5ade15 ; /* 0x3e5ade15ff137424 */ /* 0x000fca00078e00ff */ /*05d0*/ DFMA R20, R6, R12, R20 ; /* 0x0000000c0614722b */ /* 0x001e0c0000000014 */ /*05e0*/ DADD R12, R16, R20 ; /* 0x00000000100c7229 */ /* 0x001e0c0000000014 */ /*05f0*/ DADD R6, R4, R12 ; /* 0x0000000004067229 */ /* 0x001e08000000000c */ /*0600*/ DADD R16, R16, -R12 ; /* 0x0000000010107229 */ /* 0x000e48000000080c */ /*0610*/ DADD R4, R4, -R6 ; /* 0x0000000004047229 */ /* 0x001e080000000806 */ /*0620*/ DADD R16, R20, R16 ; /* 0x0000000014107229 */ /* 0x002fc80000000010 */ /*0630*/ DADD R4, R12, R4 ; /* 0x000000000c047229 */ /* 0x001e0c0000000004 */ /*0640*/ DADD R4, R16, R4 ; /* 0x0000000010047229 */ /* 0x001e0c0000000004 */ /*0650*/ DADD R10, R10, R4 ; /* 0x000000000a0a7229 */ /* 0x001e0c0000000004 */ /*0660*/ DADD R12, R6, R10 ; /* 0x00000000060c7229 */ /* 0x001e0c000000000a */ /*0670*/ DFMA R4, RZ, c[0x2][0x40], R12 ; /* 0x00801000ff047a2b */ /* 0x001e08000000000c */ /*0680*/ DADD R6, R6, -R12 ; /* 0x0000000006067229 */ /* 0x000e48000000080c */ /*0690*/ DFMA R14, -RZ, c[0x2][0x40], R4 ; /* 0x00801000ff0e7a2b */ /* 0x001e080000000104 */ /*06a0*/ DADD R6, R10, R6 ; /* 0x000000000a067229 */ /* 0x002fc80000000006 */ /*06b0*/ DADD R14, -R12, R14 ; /* 0x000000000c0e7229 */ /* 0x001e0c000000010e */ /*06c0*/ DADD R6, R6, -R14 ; /* 0x0000000006067229 */ /* 0x001064000000080e */ /*06d0*/ LOP3.LUT R15, R3, 0xff0fffff, RZ, 0xc0, !PT ; /* 0xff0fffff030f7812 */ /* 0x001fe200078ec0ff */ /*06e0*/ IMAD.MOV.U32 R14, RZ, RZ, R2 ; /* 0x000000ffff0e7224 */ /* 0x000fc600078e0002 */ /*06f0*/ DFMA R6, RZ, c[0x2][0x48], R6 ; /* 0x00801200ff067a2b */ /* 0x002e220000000006 */ /*0700*/ SEL R15, R15, R3, P1 ; /* 0x000000030f0f7207 */ /* 0x000fca0000800000 */ /*0710*/ DADD R10, R4, R6 ; /* 0x00000000040a7229 */ /* 0x001e0c0000000006 */ /*0720*/ DADD R12, R4, -R10 ; /* 0x00000000040c7229 */ /* 0x001e08000000080a */ /*0730*/ DMUL R4, R10, R14 ; /* 0x0000000e0a047228 */ /* 0x000e480000000000 */ /*0740*/ DADD R6, R6, R12 ; /* 0x0000000006067229 */ /* 0x0011e4000000000c */ /*0750*/ IMAD.MOV.U32 R12, RZ, RZ, 0x652b82fe ; /* 0x652b82feff0c7424 */ /* 0x001fe400078e00ff */ /*0760*/ DFMA R10, R10, R14, -R4 ; /* 0x0000000e0a0a722b */ /* 0x002e220000000804 */ /*0770*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3ff71547 ; /* 0x3ff71547ff0d7424 */ /* 0x000fca00078e00ff */ /*0780*/ DFMA R6, R6, R14, R10 ; /* 0x0000000e0606722b */ /* 0x001e0c000000000a */ /*0790*/ DADD R10, R4, R6 ; /* 0x00000000040a7229 */ /* 0x001e0c0000000006 */ /*07a0*/ DFMA R12, R10, R12, 6.75539944105574400000e+15 ; /* 0x433800000a0c742b */ /* 0x001e08000000000c */ /*07b0*/ FSETP.GEU.AND P1, PT, |R11|, 4.1917929649353027344, PT ; /* 0x4086232b0b00780b */ /* 0x000fe40003f2e200 */ /*07c0*/ DADD R14, R12, -6.75539944105574400000e+15 ; /* 0xc33800000c0e7429 */ /* 0x001e0c0000000000 */ /*07d0*/ DFMA R16, R14, c[0x2][0x50], R10 ; /* 0x008014000e107a2b */ /* 0x001e0c000000000a */ /*07e0*/ DFMA R14, R14, c[0x2][0x58], R16 ; /* 0x008016000e0e7a2b */ /* 0x001e0c0000000010 */ /*07f0*/ DFMA R16, R14, R18, c[0x2][0x60] ; /* 0x008018000e10762b */ /* 0x001e0c0000000012 */ /*0800*/ DFMA R16, R14, R16, c[0x2][0x68] ; /* 0x00801a000e10762b */ /* 0x001e0c0000000010 */ /*0810*/ DFMA R16, R14, R16, c[0x2][0x70] ; /* 0x00801c000e10762b */ /* 0x001e0c0000000010 */ /*0820*/ DFMA R16, R14, R16, c[0x2][0x78] ; /* 0x00801e000e10762b */ /* 0x001e0c0000000010 */ /*0830*/ DFMA R16, R14, R16, c[0x2][0x80] ; /* 0x008020000e10762b */ /* 0x001e0c0000000010 */ /*0840*/ DFMA R16, R14, R16, c[0x2][0x88] ; /* 0x008022000e10762b */ /* 0x001e0c0000000010 */ /*0850*/ DFMA R16, R14, R16, c[0x2][0x90] ; /* 0x008024000e10762b */ /* 0x001e0c0000000010 */ /*0860*/ DFMA R16, R14, R16, c[0x2][0x98] ; /* 0x008026000e10762b */ /* 0x001e0c0000000010 */ /*0870*/ DFMA R16, R14, R16, c[0x2][0xa0] ; /* 0x008028000e10762b */ /* 0x001e0c0000000010 */ /*0880*/ DFMA R16, R14, R16, 1 ; /* 0x3ff000000e10742b */ /* 0x001e0c0000000010 */ /*0890*/ DFMA R16, R14, R16, 1 ; /* 0x3ff000000e10742b */ /* 0x001e140000000010 */ /*08a0*/ IMAD R15, R12, 0x100000, R17 ; /* 0x001000000c0f7824 */ /* 0x001fe400078e0211 */ /*08b0*/ IMAD.MOV.U32 R14, RZ, RZ, R16 ; /* 0x000000ffff0e7224 */ /* 0x000fe200078e0010 */ /*08c0*/ @!P1 BRA 0x990 ; /* 0x000000c000009947 */ /* 0x000fea0003800000 */ /*08d0*/ FSETP.GEU.AND P2, PT, |R11|, 4.2275390625, PT ; /* 0x408748000b00780b */ /* 0x000fe20003f4e200 */ /*08e0*/ DADD R14, R10, +INF ; /* 0x7ff000000a0e7429 */ /* 0x000fc80000000000 */ /*08f0*/ DSETP.GEU.AND P1, PT, R10, RZ, PT ; /* 0x000000ff0a00722a */ /* 0x000e0c0003f2e000 */ /*0900*/ FSEL R14, R14, RZ, P1 ; /* 0x000000ff0e0e7208 */ /* 0x001fe40000800000 */ /*0910*/ @!P2 LEA.HI R9, R12, R12, RZ, 0x1 ; /* 0x0000000c0c09a211 */ /* 0x000fe400078f08ff */ /*0920*/ FSEL R15, R15, RZ, P1 ; /* 0x000000ff0f0f7208 */ /* 0x000fe40000800000 */ /*0930*/ @!P2 SHF.R.S32.HI R9, RZ, 0x1, R9 ; /* 0x00000001ff09a819 */ /* 0x000fca0000011409 */ /*0940*/ @!P2 IMAD.IADD R12, R12, 0x1, -R9 ; /* 0x000000010c0ca824 */ /* 0x000fe400078e0a09 */ /*0950*/ @!P2 IMAD R17, R9, 0x100000, R17 ; /* 0x001000000911a824 */ /* 0x000fc600078e0211 */ /*0960*/ @!P2 LEA R13, R12, 0x3ff00000, 0x14 ; /* 0x3ff000000c0da811 */ /* 0x000fe200078ea0ff */ /*0970*/ @!P2 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0ca224 */ /* 0x000fcc00078e00ff */ /*0980*/ @!P2 DMUL R14, R16, R12 ; /* 0x0000000c100ea228 */ /* 0x0000540000000000 */ /*0990*/ LOP3.LUT R9, R15, 0x7fffffff, RZ, 0xc0, !PT ; /* 0x7fffffff0f097812 */ /* 0x002fe200078ec0ff */ /*09a0*/ DADD R4, R4, -R10 ; /* 0x0000000004047229 */ /* 0x000e46000000080a */ /*09b0*/ ISETP.NE.AND P1, PT, R9, 0x7ff00000, PT ; /* 0x7ff000000900780c */ /* 0x000fc60003f25270 */ /*09c0*/ DADD R4, R6, R4 ; /* 0x0000000006047229 */ /* 0x002e620000000004 */ /*09d0*/ ISETP.EQ.AND P1, PT, R14, RZ, !P1 ; /* 0x000000ff0e00720c */ /* 0x000fda0004f22270 */ /*09e0*/ @!P1 DFMA R14, R4, R14, R14 ; /* 0x0000000e040e922b */ /* 0x0022a4000000000e */ /*09f0*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x002fe400078e0000 */ /*0a00*/ IMAD.MOV.U32 R5, RZ, RZ, 0x0 ; /* 0x00000000ff057424 */ /* 0x000fc800078e00ff */ /*0a10*/ RET.REL.NODEC R4 0x0 ; /* 0xfffff5e004007950 */ /* 0x004fea0003c3ffff */ /*0a20*/ BRA 0xa20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cmath> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
.file "tmpxft_0013d6ca_00000000-6_pow.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 _Z31__device_stub__Z11my_copysignPdPd .type _Z31__device_stub__Z11my_copysignPdPd, @function _Z31__device_stub__Z11my_copysignPdPd: .LFB2051: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .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 _Z11my_copysignPd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z31__device_stub__Z11my_copysignPdPd, .-_Z31__device_stub__Z11my_copysignPdPd .globl _Z11my_copysignPd .type _Z11my_copysignPd, @function _Z11my_copysignPd: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z11my_copysignPdPd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z11my_copysignPd, .-_Z11my_copysignPd .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z11my_copysignPd" .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 _Z11my_copysignPd(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <cmath> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
#include <hip/hip_runtime.h> #include <cmath> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <cmath> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z11my_copysignPd .globl _Z11my_copysignPd .p2align 8 .type _Z11my_copysignPd,@function _Z11my_copysignPd: v_cvt_f64_i32_e32 v[1:2], v0 s_mov_b32 s3, 0x3ff71547 s_mov_b32 s2, 0x652b82fe s_mov_b32 s5, 0x3e5ade15 s_mov_b32 s4, 0x6a5dcb37 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[1:2], 0 v_fma_f64 v[5:6], v[1:2], 0, -v[3:4] v_cmp_class_f64_e64 vcc_lo, v[3:4], 0x204 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_fma_f64 v[5:6], v[1:2], 0, v[5:6] v_mul_f64 v[1:2], v[1:2], 0.5 v_add_f64 v[7:8], v[3:4], v[5:6] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_dual_cndmask_b32 v10, v8, v4 :: v_dual_cndmask_b32 v9, v7, v3 v_add_f64 v[3:4], v[7:8], -v[3:4] v_mul_f64 v[11:12], v[9:10], s[2:3] s_mov_b32 s3, 0xbfe62e42 s_mov_b32 s2, 0xfefa39ef v_cmp_neq_f64_e64 vcc_lo, 0x7ff00000, |v[9:10]| s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_f64 v[3:4], v[5:6], -v[3:4] v_rndne_f64_e32 v[11:12], v[11:12] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_dual_cndmask_b32 v4, 0, v4 :: v_dual_cndmask_b32 v3, 0, v3 v_fma_f64 v[13:14], v[11:12], s[2:3], v[9:10] s_mov_b32 s3, 0xbc7abc9e s_mov_b32 s2, 0x3b39803f v_cvt_i32_f64_e32 v17, v[11:12] s_delay_alu instid0(VALU_DEP_2) v_fma_f64 v[13:14], v[11:12], s[2:3], v[13:14] s_mov_b32 s3, 0x3e928af3 s_mov_b32 s2, 0xfca7ab0c s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], s[4:5], s[2:3] s_mov_b32 s3, 0x3ec71dee s_mov_b32 s2, 0x623fde64 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3efa0199 s_mov_b32 s2, 0x7c89e6b0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f2a01a0 s_mov_b32 s2, 0x14761f6e s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f56c16c s_mov_b32 s2, 0x1852b7b0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f811111 s_mov_b32 s2, 0x11122322 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fa55555 s_mov_b32 s2, 0x555502a1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fc55555 s_mov_b32 s2, 0x55555511 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fe00000 s_mov_b32 s2, 11 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] v_cmp_nlt_f64_e64 s2, 0x40900000, v[9:10] v_cmp_ngt_f64_e64 s3, 0xc090cc00, v[9:10] s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_fma_f64 v[15:16], v[13:14], v[15:16], 1.0 s_and_b32 vcc_lo, s3, s2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[11:12], v[13:14], v[15:16], 1.0 v_ldexp_f64 v[7:8], v[11:12], v17 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e64 v5, 0x7ff00000, v8, s2 v_cndmask_b32_e64 v6, 0, v5, s3 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v5, 0, v7, vcc_lo v_trunc_f64_e32 v[7:8], v[1:2] s_load_b64 s[2:3], s[0:1], 0x0 v_cmp_eq_u32_e64 s0, 0, v0 v_fma_f64 v[3:4], v[5:6], v[3:4], v[5:6] v_cmp_class_f64_e64 vcc_lo, v[5:6], 0x204 v_mov_b32_e32 v9, 0x3ff00000 s_waitcnt lgkmcnt(0) s_load_b64 s[4:5], s[2:3], 0x0 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v0, v3, v5 :: v_dual_cndmask_b32 v3, v4, v6 v_cmp_neq_f64_e32 vcc_lo, v[7:8], v[1:2] v_cndmask_b32_e64 v9, 0xbff00000, v9, s0 v_dual_mov_b32 v2, 0 :: v_dual_cndmask_b32 v1, 0x3ff00000, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_bfi_b32 v1, 0x7fffffff, v3, v1 s_waitcnt lgkmcnt(0) v_mul_f64 v[0:1], v[0:1], s[4:5] global_store_b64 v2, v[0:1], s[2:3] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11my_copysignPd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 8 .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 18 .amdhsa_next_free_sgpr 6 .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 _Z11my_copysignPd, .Lfunc_end0-_Z11my_copysignPd .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 8 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11my_copysignPd .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z11my_copysignPd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 18 .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> __global__ void my_copysign(double* v) { int i = threadIdx.x; *v = std::pow(-1, i) * (*v); }
.text .file "pow.hip" .globl _Z26__device_stub__my_copysignPd # -- Begin function _Z26__device_stub__my_copysignPd .p2align 4, 0x90 .type _Z26__device_stub__my_copysignPd,@function _Z26__device_stub__my_copysignPd: # @_Z26__device_stub__my_copysignPd .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movq %rdi, 64(%rsp) leaq 64(%rsp), %rax movq %rax, (%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d movq %rsp, %r9 movl $_Z11my_copysignPd, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size _Z26__device_stub__my_copysignPd, .Lfunc_end0-_Z26__device_stub__my_copysignPd .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 $_Z11my_copysignPd, %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 _Z11my_copysignPd,@object # @_Z11my_copysignPd .section .rodata,"a",@progbits .globl _Z11my_copysignPd .p2align 3, 0x0 _Z11my_copysignPd: .quad _Z26__device_stub__my_copysignPd .size _Z11my_copysignPd, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11my_copysignPd" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .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__my_copysignPd .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11my_copysignPd .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 : _Z11my_copysignPd .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_TID.X ; /* 0x0000000000087919 */ /* 0x000e220000002100 */ /*0020*/ BSSY B0, 0xd0 ; /* 0x000000a000007945 */ /* 0x000fe20003800000 */ /*0030*/ I2F.F64 R2, R8 ; /* 0x0000000800027312 */ /* 0x001e240000201c00 */ /*0040*/ LOP3.LUT R0, R3, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000003007812 */ /* 0x001fc800078ec0ff */ /*0050*/ LEA.HI R5, R0, 0xfffffc0c, RZ, 0xc ; /* 0xfffffc0c00057811 */ /* 0x000fc800078f60ff */ /*0060*/ SHF.L.U32 R0, R2.reuse, R5.reuse, RZ ; /* 0x0000000502007219 */ /* 0x0c0fe400000006ff */ /*0070*/ SHF.L.U64.HI R5, R2, R5, R3 ; /* 0x0000000502057219 */ /* 0x000fe40000010203 */ /*0080*/ ISETP.NE.U32.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe40003f05070 */ /*0090*/ MOV R0, 0xc0 ; /* 0x000000c000007802 */ /* 0x000fe40000000f00 */ /*00a0*/ ISETP.NE.AND.EX P0, PT, R5, -0x80000000, PT, P0 ; /* 0x800000000500780c */ /* 0x000fd00003f05300 */ /*00b0*/ CALL.REL.NOINC 0x2f0 ; /* 0x0000023000007944 */ /* 0x000fea0003c00000 */ /*00c0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*00d0*/ FRND.F64.TRUNC R4, R2 ; /* 0x0000000200047313 */ /* 0x000e62000030d800 */ /*00e0*/ DADD R6, R2, -1 ; /* 0xbff0000002067429 */ /* 0x000ea20000000000 */ /*00f0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0100*/ BSSY B0, 0x260 ; /* 0x0000015000007945 */ /* 0x000ff00003800000 */ /*0110*/ LOP3.LUT R6, R7, 0x7ff00000, RZ, 0xc0, !PT ; /* 0x7ff0000007067812 */ /* 0x004fc800078ec0ff */ /*0120*/ ISETP.NE.AND P2, PT, R6, 0x7ff00000, PT ; /* 0x7ff000000600780c */ /* 0x000fe20003f45270 */ /*0130*/ DSETP.NEU.AND P1, PT, R4, R2, PT ; /* 0x000000020400722a */ /* 0x0022a40003f2d000 */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, R15 ; /* 0x000000ffff057224 */ /* 0x002fe400078e000f */ /*0150*/ IMAD.MOV.U32 R4, RZ, RZ, R14 ; /* 0x000000ffff047224 */ /* 0x000fc600078e000e */ /*0160*/ @!P0 LOP3.LUT R7, R5, 0x80000000, RZ, 0x3c, !PT ; /* 0x8000000005078812 */ /* 0x000fca00078e3cff */ /*0170*/ @!P0 IMAD.MOV.U32 R5, RZ, RZ, R7 ; /* 0x000000ffff058224 */ /* 0x000fe400078e0007 */ /*0180*/ @P1 IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff041424 */ /* 0x004fe400078e00ff */ /*0190*/ @P1 IMAD.MOV.U32 R5, RZ, RZ, -0x80000 ; /* 0xfff80000ff051424 */ /* 0x000fe200078e00ff */ /*01a0*/ @P2 BRA 0x250 ; /* 0x000000a000002947 */ /* 0x000fea0003800000 */ /*01b0*/ DSETP.GTU.AND P0, PT, |R2|, +INF , PT ; /* 0x7ff000000200742a */ /* 0x000e5c0003f0c200 */ /*01c0*/ @P0 BRA 0x240 ; /* 0x0000007000000947 */ /* 0x002fea0003800000 */ /*01d0*/ ISETP.NE.AND P0, PT, R2, RZ, PT ; /* 0x000000ff0200720c */ /* 0x000fe40003f05270 */ /*01e0*/ LOP3.LUT R2, R3, 0x7fffffff, RZ, 0xc0, !PT ; /* 0x7fffffff03027812 */ /* 0x000fc800078ec0ff */ /*01f0*/ ISETP.EQ.AND P0, PT, R2, 0x7ff00000, !P0 ; /* 0x7ff000000200780c */ /* 0x000fda0004702270 */ /*0200*/ @!P0 BRA 0x250 ; /* 0x0000004000008947 */ /* 0x000fea0003800000 */ /*0210*/ IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff047424 */ /* 0x000fe400078e00ff */ /*0220*/ IMAD.MOV.U32 R5, RZ, RZ, 0x3ff00000 ; /* 0x3ff00000ff057424 */ /* 0x000fe200078e00ff */ /*0230*/ BRA 0x250 ; /* 0x0000001000007947 */ /* 0x000fea0003800000 */ /*0240*/ DADD R4, R2, -1 ; /* 0xbff0000002047429 */ /* 0x00028c0000000000 */ /*0250*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0260*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x160] ; /* 0x00005800ff027624 */ /* 0x002fe400078e00ff */ /*0270*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x164] ; /* 0x00005900ff037624 */ /* 0x000fca00078e00ff */ /*0280*/ LDG.E.64 R6, [R2.64] ; /* 0x0000000402067981 */ /* 0x000ee2000c1e1b00 */ /*0290*/ ISETP.NE.AND P0, PT, R8, RZ, PT ; /* 0x000000ff0800720c */ /* 0x000fc80003f05270 */ /*02a0*/ FSEL R4, R4, RZ, P0 ; /* 0x000000ff04047208 */ /* 0x004fe40000000000 */ /*02b0*/ FSEL R5, R5, 1.875, P0 ; /* 0x3ff0000005057808 */ /* 0x000fcc0000000000 */ /*02c0*/ DMUL R6, R6, R4 ; /* 0x0000000406067228 */ /* 0x008e4e0000000000 */ /*02d0*/ STG.E.64 [R2.64], R6 ; /* 0x0000000602007986 */ /* 0x002fe2000c101b04 */ /*02e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02f0*/ MUFU.RCP64H R11, 2 ; /* 0x40000000000b7908 */ /* 0x000e220000001800 */ /*0300*/ IMAD.MOV.U32 R4, RZ, RZ, 0x0 ; /* 0x00000000ff047424 */ /* 0x000fe400078e00ff */ /*0310*/ IMAD.MOV.U32 R5, RZ, RZ, 0x3ff00000 ; /* 0x3ff00000ff057424 */ /* 0x000fe400078e00ff */ /*0320*/ IMAD.MOV.U32 R10, RZ, RZ, RZ ; /* 0x000000ffff0a7224 */ /* 0x000fe400078e00ff */ /*0330*/ IMAD.MOV.U32 R6, RZ, RZ, 0x7d2cafe2 ; /* 0x7d2cafe2ff067424 */ /* 0x000fe400078e00ff */ /*0340*/ IMAD.MOV.U32 R7, RZ, RZ, 0x3eb0f5ff ; /* 0x3eb0f5ffff077424 */ /* 0x000fc400078e00ff */ /*0350*/ IMAD.SHL.U32 R9, R3, 0x2, RZ ; /* 0x0000000203097824 */ /* 0x000fe200078e00ff */ /*0360*/ DFMA R4, R10, -2, R4 ; /* 0xc00000000a04782b */ /* 0x001e080000000004 */ /*0370*/ ISETP.GT.U32.AND P1, PT, R9, -0x2000001, PT ; /* 0xfdffffff0900780c */ /* 0x000fe40003f24070 */ /*0380*/ DFMA R4, R4, R4, R4 ; /* 0x000000040404722b */ /* 0x001e0c0000000004 */ /*0390*/ DFMA R10, R10, R4, R10 ; /* 0x000000040a0a722b */ /* 0x001e0c000000000a */ /*03a0*/ DMUL R4, RZ, R10 ; /* 0x0000000aff047228 */ /* 0x001e0c0000000000 */ /*03b0*/ DFMA R4, RZ, R10, R4 ; /* 0x0000000aff04722b */ /* 0x001e0c0000000004 */ /*03c0*/ DMUL R12, R4, R4 ; /* 0x00000004040c7228 */ /* 0x001e080000000000 */ /*03d0*/ DADD R14, RZ, -R4 ; /* 0x00000000ff0e7229 */ /* 0x000e480000000804 */ /*03e0*/ DFMA R6, R12, R6, c[0x2][0x0] ; /* 0x008000000c06762b */ /* 0x001e080000000006 */ /*03f0*/ DADD R16, R14, R14 ; /* 0x000000000e107229 */ /* 0x002fc8000000000e */ /*0400*/ DFMA R6, R12, R6, c[0x2][0x8] ; /* 0x008002000c06762b */ /* 0x001e080000000006 */ /*0410*/ DMUL R14, R4, R4 ; /* 0x00000004040e7228 */ /* 0x000fc80000000000 */ /*0420*/ DFMA R6, R12, R6, c[0x2][0x10] ; /* 0x008004000c06762b */ /* 0x001e080000000006 */ /*0430*/ DFMA R20, RZ, -R4, R16 ; /* 0x80000004ff14722b */ /* 0x000e480000000010 */ /*0440*/ DFMA R6, R12, R6, c[0x2][0x18] ; /* 0x008006000c06762b */ /* 0x001e080000000006 */ /*0450*/ DMUL R10, R10, R20 ; /* 0x000000140a0a7228 */ /* 0x002fc80000000000 */ /*0460*/ DFMA R6, R12, R6, c[0x2][0x20] ; /* 0x008008000c06762b */ /* 0x001e080000000006 */ /*0470*/ DFMA R16, R4, R4, -R14 ; /* 0x000000040410722b */ /* 0x000fc8000000080e */ /*0480*/ DFMA R18, R12, R6, c[0x2][0x28] ; /* 0x00800a000c12762b */ /* 0x001e0c0000000006 */ /*0490*/ DFMA R6, R12, R18, c[0x2][0x30] ; /* 0x00800c000c06762b */ /* 0x001e0c0000000012 */ /*04a0*/ DADD R22, -R6, c[0x2][0x30] ; /* 0x00800c0006167629 */ /* 0x001e0c0000000100 */ /*04b0*/ DFMA R22, R12, R18, R22 ; /* 0x000000120c16722b */ /* 0x001e080000000016 */ /*04c0*/ DMUL R12, R4, R14 ; /* 0x0000000e040c7228 */ /* 0x000e480000000000 */ /*04d0*/ DADD R18, RZ, R22 ; /* 0x00000000ff127229 */ /* 0x001e080000000016 */ /*04e0*/ DFMA R20, R4, R14, -R12 ; /* 0x0000000e0414722b */ /* 0x002e48000000080c */ /*04f0*/ DADD R18, R18, c[0x2][0x38] ; /* 0x00800e0012127629 */ /* 0x001e080000000000 */ /*0500*/ DFMA R22, R10, R14, R20 ; /* 0x0000000e0a16722b */ /* 0x0023e40000000014 */ /*0510*/ IADD3 R21, R11, 0x100000, RZ ; /* 0x001000000b157810 */ /* 0x002fe20007ffe0ff */ /*0520*/ IMAD.MOV.U32 R20, RZ, RZ, R10 ; /* 0x000000ffff147224 */ /* 0x000fe200078e000a */ /*0530*/ DADD R14, R6, R18 ; /* 0x00000000060e7229 */ /* 0x001e0a0000000012 */ /*0540*/ DFMA R20, R4, R20, R16 ; /* 0x000000140414722b */ /* 0x000e480000000010 */ /*0550*/ DMUL R16, R14, R12 ; /* 0x0000000c0e107228 */ /* 0x001e080000000000 */ /*0560*/ DFMA R20, R4, R20, R22 ; /* 0x000000140414722b */ /* 0x002fc80000000016 */ /*0570*/ DADD R6, R6, -R14 ; /* 0x0000000006067229 */ /* 0x000e48000000080e */ /*0580*/ DFMA R22, R14, R12, -R16 ; /* 0x0000000c0e16722b */ /* 0x001e080000000810 */ /*0590*/ DADD R6, R18, R6 ; /* 0x0000000012067229 */ /* 0x0023e40000000006 */ /*05a0*/ IMAD.MOV.U32 R18, RZ, RZ, 0x69ce2bdf ; /* 0x69ce2bdfff127424 */ /* 0x002fe400078e00ff */ /*05b0*/ DFMA R20, R14, R20, R22 ; /* 0x000000140e14722b */ /* 0x001e220000000016 */ /*05c0*/ IMAD.MOV.U32 R19, RZ, RZ, 0x3e5ade15 ; /* 0x3e5ade15ff137424 */ /* 0x000fca00078e00ff */ /*05d0*/ DFMA R20, R6, R12, R20 ; /* 0x0000000c0614722b */ /* 0x001e0c0000000014 */ /*05e0*/ DADD R12, R16, R20 ; /* 0x00000000100c7229 */ /* 0x001e0c0000000014 */ /*05f0*/ DADD R6, R4, R12 ; /* 0x0000000004067229 */ /* 0x001e08000000000c */ /*0600*/ DADD R16, R16, -R12 ; /* 0x0000000010107229 */ /* 0x000e48000000080c */ /*0610*/ DADD R4, R4, -R6 ; /* 0x0000000004047229 */ /* 0x001e080000000806 */ /*0620*/ DADD R16, R20, R16 ; /* 0x0000000014107229 */ /* 0x002fc80000000010 */ /*0630*/ DADD R4, R12, R4 ; /* 0x000000000c047229 */ /* 0x001e0c0000000004 */ /*0640*/ DADD R4, R16, R4 ; /* 0x0000000010047229 */ /* 0x001e0c0000000004 */ /*0650*/ DADD R10, R10, R4 ; /* 0x000000000a0a7229 */ /* 0x001e0c0000000004 */ /*0660*/ DADD R12, R6, R10 ; /* 0x00000000060c7229 */ /* 0x001e0c000000000a */ /*0670*/ DFMA R4, RZ, c[0x2][0x40], R12 ; /* 0x00801000ff047a2b */ /* 0x001e08000000000c */ /*0680*/ DADD R6, R6, -R12 ; /* 0x0000000006067229 */ /* 0x000e48000000080c */ /*0690*/ DFMA R14, -RZ, c[0x2][0x40], R4 ; /* 0x00801000ff0e7a2b */ /* 0x001e080000000104 */ /*06a0*/ DADD R6, R10, R6 ; /* 0x000000000a067229 */ /* 0x002fc80000000006 */ /*06b0*/ DADD R14, -R12, R14 ; /* 0x000000000c0e7229 */ /* 0x001e0c000000010e */ /*06c0*/ DADD R6, R6, -R14 ; /* 0x0000000006067229 */ /* 0x001064000000080e */ /*06d0*/ LOP3.LUT R15, R3, 0xff0fffff, RZ, 0xc0, !PT ; /* 0xff0fffff030f7812 */ /* 0x001fe200078ec0ff */ /*06e0*/ IMAD.MOV.U32 R14, RZ, RZ, R2 ; /* 0x000000ffff0e7224 */ /* 0x000fc600078e0002 */ /*06f0*/ DFMA R6, RZ, c[0x2][0x48], R6 ; /* 0x00801200ff067a2b */ /* 0x002e220000000006 */ /*0700*/ SEL R15, R15, R3, P1 ; /* 0x000000030f0f7207 */ /* 0x000fca0000800000 */ /*0710*/ DADD R10, R4, R6 ; /* 0x00000000040a7229 */ /* 0x001e0c0000000006 */ /*0720*/ DADD R12, R4, -R10 ; /* 0x00000000040c7229 */ /* 0x001e08000000080a */ /*0730*/ DMUL R4, R10, R14 ; /* 0x0000000e0a047228 */ /* 0x000e480000000000 */ /*0740*/ DADD R6, R6, R12 ; /* 0x0000000006067229 */ /* 0x0011e4000000000c */ /*0750*/ IMAD.MOV.U32 R12, RZ, RZ, 0x652b82fe ; /* 0x652b82feff0c7424 */ /* 0x001fe400078e00ff */ /*0760*/ DFMA R10, R10, R14, -R4 ; /* 0x0000000e0a0a722b */ /* 0x002e220000000804 */ /*0770*/ IMAD.MOV.U32 R13, RZ, RZ, 0x3ff71547 ; /* 0x3ff71547ff0d7424 */ /* 0x000fca00078e00ff */ /*0780*/ DFMA R6, R6, R14, R10 ; /* 0x0000000e0606722b */ /* 0x001e0c000000000a */ /*0790*/ DADD R10, R4, R6 ; /* 0x00000000040a7229 */ /* 0x001e0c0000000006 */ /*07a0*/ DFMA R12, R10, R12, 6.75539944105574400000e+15 ; /* 0x433800000a0c742b */ /* 0x001e08000000000c */ /*07b0*/ FSETP.GEU.AND P1, PT, |R11|, 4.1917929649353027344, PT ; /* 0x4086232b0b00780b */ /* 0x000fe40003f2e200 */ /*07c0*/ DADD R14, R12, -6.75539944105574400000e+15 ; /* 0xc33800000c0e7429 */ /* 0x001e0c0000000000 */ /*07d0*/ DFMA R16, R14, c[0x2][0x50], R10 ; /* 0x008014000e107a2b */ /* 0x001e0c000000000a */ /*07e0*/ DFMA R14, R14, c[0x2][0x58], R16 ; /* 0x008016000e0e7a2b */ /* 0x001e0c0000000010 */ /*07f0*/ DFMA R16, R14, R18, c[0x2][0x60] ; /* 0x008018000e10762b */ /* 0x001e0c0000000012 */ /*0800*/ DFMA R16, R14, R16, c[0x2][0x68] ; /* 0x00801a000e10762b */ /* 0x001e0c0000000010 */ /*0810*/ DFMA R16, R14, R16, c[0x2][0x70] ; /* 0x00801c000e10762b */ /* 0x001e0c0000000010 */ /*0820*/ DFMA R16, R14, R16, c[0x2][0x78] ; /* 0x00801e000e10762b */ /* 0x001e0c0000000010 */ /*0830*/ DFMA R16, R14, R16, c[0x2][0x80] ; /* 0x008020000e10762b */ /* 0x001e0c0000000010 */ /*0840*/ DFMA R16, R14, R16, c[0x2][0x88] ; /* 0x008022000e10762b */ /* 0x001e0c0000000010 */ /*0850*/ DFMA R16, R14, R16, c[0x2][0x90] ; /* 0x008024000e10762b */ /* 0x001e0c0000000010 */ /*0860*/ DFMA R16, R14, R16, c[0x2][0x98] ; /* 0x008026000e10762b */ /* 0x001e0c0000000010 */ /*0870*/ DFMA R16, R14, R16, c[0x2][0xa0] ; /* 0x008028000e10762b */ /* 0x001e0c0000000010 */ /*0880*/ DFMA R16, R14, R16, 1 ; /* 0x3ff000000e10742b */ /* 0x001e0c0000000010 */ /*0890*/ DFMA R16, R14, R16, 1 ; /* 0x3ff000000e10742b */ /* 0x001e140000000010 */ /*08a0*/ IMAD R15, R12, 0x100000, R17 ; /* 0x001000000c0f7824 */ /* 0x001fe400078e0211 */ /*08b0*/ IMAD.MOV.U32 R14, RZ, RZ, R16 ; /* 0x000000ffff0e7224 */ /* 0x000fe200078e0010 */ /*08c0*/ @!P1 BRA 0x990 ; /* 0x000000c000009947 */ /* 0x000fea0003800000 */ /*08d0*/ FSETP.GEU.AND P2, PT, |R11|, 4.2275390625, PT ; /* 0x408748000b00780b */ /* 0x000fe20003f4e200 */ /*08e0*/ DADD R14, R10, +INF ; /* 0x7ff000000a0e7429 */ /* 0x000fc80000000000 */ /*08f0*/ DSETP.GEU.AND P1, PT, R10, RZ, PT ; /* 0x000000ff0a00722a */ /* 0x000e0c0003f2e000 */ /*0900*/ FSEL R14, R14, RZ, P1 ; /* 0x000000ff0e0e7208 */ /* 0x001fe40000800000 */ /*0910*/ @!P2 LEA.HI R9, R12, R12, RZ, 0x1 ; /* 0x0000000c0c09a211 */ /* 0x000fe400078f08ff */ /*0920*/ FSEL R15, R15, RZ, P1 ; /* 0x000000ff0f0f7208 */ /* 0x000fe40000800000 */ /*0930*/ @!P2 SHF.R.S32.HI R9, RZ, 0x1, R9 ; /* 0x00000001ff09a819 */ /* 0x000fca0000011409 */ /*0940*/ @!P2 IMAD.IADD R12, R12, 0x1, -R9 ; /* 0x000000010c0ca824 */ /* 0x000fe400078e0a09 */ /*0950*/ @!P2 IMAD R17, R9, 0x100000, R17 ; /* 0x001000000911a824 */ /* 0x000fc600078e0211 */ /*0960*/ @!P2 LEA R13, R12, 0x3ff00000, 0x14 ; /* 0x3ff000000c0da811 */ /* 0x000fe200078ea0ff */ /*0970*/ @!P2 IMAD.MOV.U32 R12, RZ, RZ, RZ ; /* 0x000000ffff0ca224 */ /* 0x000fcc00078e00ff */ /*0980*/ @!P2 DMUL R14, R16, R12 ; /* 0x0000000c100ea228 */ /* 0x0000540000000000 */ /*0990*/ LOP3.LUT R9, R15, 0x7fffffff, RZ, 0xc0, !PT ; /* 0x7fffffff0f097812 */ /* 0x002fe200078ec0ff */ /*09a0*/ DADD R4, R4, -R10 ; /* 0x0000000004047229 */ /* 0x000e46000000080a */ /*09b0*/ ISETP.NE.AND P1, PT, R9, 0x7ff00000, PT ; /* 0x7ff000000900780c */ /* 0x000fc60003f25270 */ /*09c0*/ DADD R4, R6, R4 ; /* 0x0000000006047229 */ /* 0x002e620000000004 */ /*09d0*/ ISETP.EQ.AND P1, PT, R14, RZ, !P1 ; /* 0x000000ff0e00720c */ /* 0x000fda0004f22270 */ /*09e0*/ @!P1 DFMA R14, R4, R14, R14 ; /* 0x0000000e040e922b */ /* 0x0022a4000000000e */ /*09f0*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x002fe400078e0000 */ /*0a00*/ IMAD.MOV.U32 R5, RZ, RZ, 0x0 ; /* 0x00000000ff057424 */ /* 0x000fc800078e00ff */ /*0a10*/ RET.REL.NODEC R4 0x0 ; /* 0xfffff5e004007950 */ /* 0x004fea0003c3ffff */ /*0a20*/ BRA 0xa20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z11my_copysignPd .globl _Z11my_copysignPd .p2align 8 .type _Z11my_copysignPd,@function _Z11my_copysignPd: v_cvt_f64_i32_e32 v[1:2], v0 s_mov_b32 s3, 0x3ff71547 s_mov_b32 s2, 0x652b82fe s_mov_b32 s5, 0x3e5ade15 s_mov_b32 s4, 0x6a5dcb37 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[3:4], v[1:2], 0 v_fma_f64 v[5:6], v[1:2], 0, -v[3:4] v_cmp_class_f64_e64 vcc_lo, v[3:4], 0x204 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_fma_f64 v[5:6], v[1:2], 0, v[5:6] v_mul_f64 v[1:2], v[1:2], 0.5 v_add_f64 v[7:8], v[3:4], v[5:6] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_dual_cndmask_b32 v10, v8, v4 :: v_dual_cndmask_b32 v9, v7, v3 v_add_f64 v[3:4], v[7:8], -v[3:4] v_mul_f64 v[11:12], v[9:10], s[2:3] s_mov_b32 s3, 0xbfe62e42 s_mov_b32 s2, 0xfefa39ef v_cmp_neq_f64_e64 vcc_lo, 0x7ff00000, |v[9:10]| s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_f64 v[3:4], v[5:6], -v[3:4] v_rndne_f64_e32 v[11:12], v[11:12] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_dual_cndmask_b32 v4, 0, v4 :: v_dual_cndmask_b32 v3, 0, v3 v_fma_f64 v[13:14], v[11:12], s[2:3], v[9:10] s_mov_b32 s3, 0xbc7abc9e s_mov_b32 s2, 0x3b39803f v_cvt_i32_f64_e32 v17, v[11:12] s_delay_alu instid0(VALU_DEP_2) v_fma_f64 v[13:14], v[11:12], s[2:3], v[13:14] s_mov_b32 s3, 0x3e928af3 s_mov_b32 s2, 0xfca7ab0c s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], s[4:5], s[2:3] s_mov_b32 s3, 0x3ec71dee s_mov_b32 s2, 0x623fde64 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3efa0199 s_mov_b32 s2, 0x7c89e6b0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f2a01a0 s_mov_b32 s2, 0x14761f6e s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f56c16c s_mov_b32 s2, 0x1852b7b0 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3f811111 s_mov_b32 s2, 0x11122322 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fa55555 s_mov_b32 s2, 0x555502a1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fc55555 s_mov_b32 s2, 0x55555511 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] s_mov_b32 s3, 0x3fe00000 s_mov_b32 s2, 11 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fma_f64 v[15:16], v[13:14], v[15:16], s[2:3] v_cmp_nlt_f64_e64 s2, 0x40900000, v[9:10] v_cmp_ngt_f64_e64 s3, 0xc090cc00, v[9:10] s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_fma_f64 v[15:16], v[13:14], v[15:16], 1.0 s_and_b32 vcc_lo, s3, s2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[11:12], v[13:14], v[15:16], 1.0 v_ldexp_f64 v[7:8], v[11:12], v17 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e64 v5, 0x7ff00000, v8, s2 v_cndmask_b32_e64 v6, 0, v5, s3 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_3) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v5, 0, v7, vcc_lo v_trunc_f64_e32 v[7:8], v[1:2] s_load_b64 s[2:3], s[0:1], 0x0 v_cmp_eq_u32_e64 s0, 0, v0 v_fma_f64 v[3:4], v[5:6], v[3:4], v[5:6] v_cmp_class_f64_e64 vcc_lo, v[5:6], 0x204 v_mov_b32_e32 v9, 0x3ff00000 s_waitcnt lgkmcnt(0) s_load_b64 s[4:5], s[2:3], 0x0 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v0, v3, v5 :: v_dual_cndmask_b32 v3, v4, v6 v_cmp_neq_f64_e32 vcc_lo, v[7:8], v[1:2] v_cndmask_b32_e64 v9, 0xbff00000, v9, s0 v_dual_mov_b32 v2, 0 :: v_dual_cndmask_b32 v1, 0x3ff00000, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_bfi_b32 v1, 0x7fffffff, v3, v1 s_waitcnt lgkmcnt(0) v_mul_f64 v[0:1], v[0:1], s[4:5] global_store_b64 v2, v[0:1], s[2:3] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11my_copysignPd .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 8 .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 18 .amdhsa_next_free_sgpr 6 .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 _Z11my_copysignPd, .Lfunc_end0-_Z11my_copysignPd .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 8 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11my_copysignPd .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z11my_copysignPd.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 18 .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_0013d6ca_00000000-6_pow.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 _Z31__device_stub__Z11my_copysignPdPd .type _Z31__device_stub__Z11my_copysignPdPd, @function _Z31__device_stub__Z11my_copysignPdPd: .LFB2051: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .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 _Z11my_copysignPd(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z31__device_stub__Z11my_copysignPdPd, .-_Z31__device_stub__Z11my_copysignPdPd .globl _Z11my_copysignPd .type _Z11my_copysignPd, @function _Z11my_copysignPd: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z11my_copysignPdPd addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z11my_copysignPd, .-_Z11my_copysignPd .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z11my_copysignPd" .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 _Z11my_copysignPd(%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 "pow.hip" .globl _Z26__device_stub__my_copysignPd # -- Begin function _Z26__device_stub__my_copysignPd .p2align 4, 0x90 .type _Z26__device_stub__my_copysignPd,@function _Z26__device_stub__my_copysignPd: # @_Z26__device_stub__my_copysignPd .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movq %rdi, 64(%rsp) leaq 64(%rsp), %rax movq %rax, (%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d movq %rsp, %r9 movl $_Z11my_copysignPd, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size _Z26__device_stub__my_copysignPd, .Lfunc_end0-_Z26__device_stub__my_copysignPd .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 $_Z11my_copysignPd, %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 _Z11my_copysignPd,@object # @_Z11my_copysignPd .section .rodata,"a",@progbits .globl _Z11my_copysignPd .p2align 3, 0x0 _Z11my_copysignPd: .quad _Z26__device_stub__my_copysignPd .size _Z11my_copysignPd, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z11my_copysignPd" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .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__my_copysignPd .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11my_copysignPd .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
code for sm_80 Function : _Z14KernelTilesMulPiS_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 R9, SR_CTAID.X ; /* 0x0000000000097919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff067624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe20000000a00 */ /*0040*/ HFMA2.MMA R21, -RZ, RZ, 0, 0 ; /* 0x00000000ff157435 */ /* 0x000fe200000001ff */ /*0050*/ S2R R12, SR_TID.X ; /* 0x00000000000c7919 */ /* 0x000e240000002100 */ /*0060*/ ISETP.GE.AND P1, PT, R6, 0x1, PT ; /* 0x000000010600780c */ /* 0x000fe40003f26270 */ /*0070*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e680000002600 */ /*0080*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0090*/ LEA R0, R9, R12, 0x4 ; /* 0x0000000c09007211 */ /* 0x001fc800078e20ff */ /*00a0*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fe40003f06270 */ /*00b0*/ LEA R3, R3, R2, 0x4 ; /* 0x0000000203037211 */ /* 0x002fc800078e20ff */ /*00c0*/ ISETP.GE.OR P0, PT, R3, c[0x0][0x178], P0 ; /* 0x00005e0003007a0c */ /* 0x000fe20000706670 */ /*00d0*/ @!P1 BRA 0x5b0 ; /* 0x000004d000009947 */ /* 0x000fd80003800000 */ /*00e0*/ IADD3 R6, R6, 0xf, RZ ; /* 0x0000000f06067810 */ /* 0x000fe20007ffe0ff */ /*00f0*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe200078e00ff */ /*0100*/ SHF.L.U32 R15, R2.reuse, 0x6, RZ ; /* 0x00000006020f7819 */ /* 0x040fe200000006ff */ /*0110*/ IMAD R4, R3, c[0x0][0x17c], R12 ; /* 0x00005f0003047a24 */ /* 0x000fe200078e020c */ /*0120*/ SHF.R.S32.HI R7, RZ, 0x1f, R6 ; /* 0x0000001fff077819 */ /* 0x000fe20000011406 */ /*0130*/ IMAD R16, R2, c[0x0][0x180], R12 ; /* 0x0000600002107a24 */ /* 0x000fe200078e020c */ /*0140*/ MOV R21, RZ ; /* 0x000000ff00157202 */ /* 0x000fe20000000f00 */ /*0150*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0160*/ LEA.HI R7, R7, R6, RZ, 0x4 ; /* 0x0000000607077211 */ /* 0x000fe200078f20ff */ /*0170*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0180*/ LEA R16, R9, R16, 0x4 ; /* 0x0000001009107211 */ /* 0x000fe200078e20ff */ /*0190*/ IMAD.MOV.U32 R14, RZ, RZ, R4 ; /* 0x000000ffff0e7224 */ /* 0x000fe200078e0004 */ /*01a0*/ MOV R13, R5 ; /* 0x00000005000d7202 */ /* 0x000fe20000000f00 */ /*01b0*/ IMAD R19, R12.reuse, 0x4, R15 ; /* 0x000000040c137824 */ /* 0x040fe200078e020f */ /*01c0*/ LEA R18, R12, 0x400, 0x2 ; /* 0x000004000c127811 */ /* 0x000fc400078e10ff */ /*01d0*/ SHF.R.S32.HI R17, RZ, 0x4, R7 ; /* 0x00000004ff117819 */ /* 0x000fe40000011407 */ /*01e0*/ ISETP.GE.AND P1, PT, R2, c[0x0][0x17c], PT ; /* 0x00005f0002007a0c */ /* 0x000fe20003f26270 */ /*01f0*/ HFMA2.MMA R22, -RZ, RZ, 0, 0 ; /* 0x00000000ff167435 */ /* 0x000fe200000001ff */ /*0200*/ ISETP.GE.AND P2, PT, R12, c[0x0][0x17c], PT ; /* 0x00005f000c007a0c */ /* 0x000fe20003f46270 */ /*0210*/ IMAD.MOV.U32 R28, RZ, RZ, RZ ; /* 0x000000ffff1c7224 */ /* 0x000fe200078e00ff */ /*0220*/ ISETP.GE.OR P1, PT, R0, c[0x0][0x180], P1 ; /* 0x0000600000007a0c */ /* 0x000fe40000f26670 */ /*0230*/ ISETP.GE.OR P2, PT, R3, c[0x0][0x178], P2 ; /* 0x00005e0003007a0c */ /* 0x000fd60001746670 */ /*0240*/ @!P1 MOV R25, 0x4 ; /* 0x0000000400199802 */ /* 0x000fe40000000f00 */ /*0250*/ @!P2 MOV R4, R14 ; /* 0x0000000e0004a202 */ /* 0x000fe40000000f00 */ /*0260*/ @!P2 MOV R5, R13 ; /* 0x0000000d0005a202 */ /* 0x000fe20000000f00 */ /*0270*/ @!P1 IMAD.WIDE R24, R16, R25, c[0x0][0x168] ; /* 0x00005a0010189625 */ /* 0x000fc800078e0219 */ /*0280*/ @!P2 LDG.E R22, [R4.64] ; /* 0x000000060416a981 */ /* 0x000ea8000c1e1900 */ /*0290*/ @!P1 LDG.E R28, [R24.64] ; /* 0x00000006181c9981 */ /* 0x000ee2000c1e1900 */ /*02a0*/ UIADD3 UR4, UR4, 0x1, URZ ; /* 0x0000000104047890 */ /* 0x000fe2000fffe03f */ /*02b0*/ IADD3 R14, P2, R14, 0x40, RZ ; /* 0x000000400e0e7810 */ /* 0x000fe40007f5e0ff */ /*02c0*/ IADD3 R2, R2, 0x10, RZ ; /* 0x0000001002027810 */ /* 0x000fe40007ffe0ff */ /*02d0*/ IADD3 R12, R12, 0x10, RZ ; /* 0x000000100c0c7810 */ /* 0x000fc40007ffe0ff */ /*02e0*/ ISETP.LE.AND P1, PT, R17, UR4, PT ; /* 0x0000000411007c0c */ /* 0x000fe4000bf23270 */ /*02f0*/ IADD3.X R13, RZ, R13, RZ, P2, !PT ; /* 0x0000000dff0d7210 */ /* 0x000fe200017fe4ff */ /*0300*/ STS [R19], R22 ; /* 0x0000001613007388 */ /* 0x004fe80000000800 */ /*0310*/ STS [R19+0x400], R28 ; /* 0x0004001c13007388 */ /* 0x008fe80000000800 */ /*0320*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0330*/ LDS R29, [R18] ; /* 0x00000000121d7984 */ /* 0x000fe80000000800 */ /*0340*/ LDS.128 R8, [R15] ; /* 0x000000000f087984 */ /* 0x000e280000000c00 */ /*0350*/ LDS R24, [R18+0x40] ; /* 0x0000400012187984 */ /* 0x000e680000000800 */ /*0360*/ LDS R27, [R18+0x80] ; /* 0x00008000121b7984 */ /* 0x000ea80000000800 */ /*0370*/ LDS R26, [R18+0xc0] ; /* 0x0000c000121a7984 */ /* 0x000ee80000000800 */ /*0380*/ LDS R23, [R18+0x100] ; /* 0x0001000012177984 */ /* 0x000fe80000000800 */ /*0390*/ LDS.128 R4, [R15+0x10] ; /* 0x000010000f047984 */ /* 0x000f280000000c00 */ /*03a0*/ LDS R20, [R18+0x140] ; /* 0x0001400012147984 */ /* 0x000f680000000800 */ /*03b0*/ LDS R25, [R18+0x180] ; /* 0x0001800012197984 */ /* 0x000f680000000800 */ /*03c0*/ LDS R22, [R18+0x1c0] ; /* 0x0001c00012167984 */ /* 0x000f620000000800 */ /*03d0*/ IMAD R8, R29, R8, R21 ; /* 0x000000081d087224 */ /* 0x001fc600078e0215 */ /*03e0*/ LDS R21, [R18+0x200] ; /* 0x0002000012157984 */ /* 0x000fe20000000800 */ /*03f0*/ IMAD R8, R24, R9, R8 ; /* 0x0000000918087224 */ /* 0x002fc600078e0208 */ /*0400*/ LDS R24, [R18+0x240] ; /* 0x0002400012187984 */ /* 0x000fe20000000800 */ /*0410*/ IMAD R8, R27, R10, R8 ; /* 0x0000000a1b087224 */ /* 0x004fc800078e0208 */ /*0420*/ IMAD R26, R26, R11, R8 ; /* 0x0000000b1a1a7224 */ /* 0x008fe400078e0208 */ /*0430*/ LDS.128 R8, [R15+0x20] ; /* 0x000020000f087984 */ /* 0x000e240000000c00 */ /*0440*/ IMAD R4, R23, R4, R26 ; /* 0x0000000417047224 */ /* 0x010fe400078e021a */ /*0450*/ LDS R23, [R18+0x280] ; /* 0x0002800012177984 */ /* 0x000e640000000800 */ /*0460*/ IMAD R4, R20, R5, R4 ; /* 0x0000000514047224 */ /* 0x020fe400078e0204 */ /*0470*/ LDS R20, [R18+0x2c0] ; /* 0x0002c00012147984 */ /* 0x000ea40000000800 */ /*0480*/ IMAD R4, R25, R6, R4 ; /* 0x0000000619047224 */ /* 0x000fc400078e0204 */ /*0490*/ LDS R25, [R18+0x300] ; /* 0x0003000012197984 */ /* 0x000fe40000000800 */ /*04a0*/ IMAD R26, R22, R7, R4 ; /* 0x00000007161a7224 */ /* 0x000fe400078e0204 */ /*04b0*/ LDS.128 R4, [R15+0x30] ; /* 0x000030000f047984 */ /* 0x000ee80000000c00 */ /*04c0*/ LDS R22, [R18+0x340] ; /* 0x0003400012167984 */ /* 0x000f220000000800 */ /*04d0*/ IMAD R26, R21, R8, R26 ; /* 0x00000008151a7224 */ /* 0x001fc600078e021a */ /*04e0*/ LDS R21, [R18+0x380] ; /* 0x0003800012157984 */ /* 0x000e220000000800 */ /*04f0*/ IMAD R9, R24, R9, R26 ; /* 0x0000000918097224 */ /* 0x000fc600078e021a */ /*0500*/ LDS R8, [R18+0x3c0] ; /* 0x0003c00012087984 */ /* 0x000f620000000800 */ /*0510*/ IMAD R9, R23, R10, R9 ; /* 0x0000000a17097224 */ /* 0x002fc800078e0209 */ /*0520*/ IMAD R9, R20, R11, R9 ; /* 0x0000000b14097224 */ /* 0x004fc800078e0209 */ /*0530*/ IMAD R4, R25, R4, R9 ; /* 0x0000000419047224 */ /* 0x008fc800078e0209 */ /*0540*/ IMAD R4, R22, R5, R4 ; /* 0x0000000516047224 */ /* 0x010fe400078e0204 */ /*0550*/ IMAD.MOV.U32 R5, RZ, RZ, 0x10 ; /* 0x00000010ff057424 */ /* 0x000fc800078e00ff */ /*0560*/ IMAD R16, R5, c[0x0][0x180], R16 ; /* 0x0000600005107a24 */ /* 0x000fe400078e0210 */ /*0570*/ IMAD R21, R21, R6, R4 ; /* 0x0000000615157224 */ /* 0x001fc800078e0204 */ /*0580*/ IMAD R21, R8, R7, R21 ; /* 0x0000000708157224 */ /* 0x020fe200078e0215 */ /*0590*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*05a0*/ @!P1 BRA 0x1e0 ; /* 0xfffffc3000009947 */ /* 0x000fea000383ffff */ /*05b0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*05c0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fe200000001ff */ /*05d0*/ IMAD R3, R3, c[0x0][0x180], R0 ; /* 0x0000600003037a24 */ /* 0x000fd200078e0200 */ /*05e0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x170] ; /* 0x00005c0003027625 */ /* 0x000fca00078e0202 */ /*05f0*/ STG.E [R2.64], R21 ; /* 0x0000001502007986 */ /* 0x000fe2000c101906 */ /*0600*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0610*/ BRA 0x610; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0620*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0630*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0640*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0650*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0660*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0670*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0680*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0690*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
.file "tmpxft_000b4cbc_00000000-6_KernelTilesMul.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__Z14KernelTilesMulPiS_S_iiiPiS_S_iii .type _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii, @function _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii: .LFB2051: .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 .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14KernelTilesMulPiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii, .-_Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii .globl _Z14KernelTilesMulPiS_S_iii .type _Z14KernelTilesMulPiS_S_iii, @function _Z14KernelTilesMulPiS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z14KernelTilesMulPiS_S_iii, .-_Z14KernelTilesMulPiS_S_iii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14KernelTilesMulPiS_S_iii" .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 _Z14KernelTilesMulPiS_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 .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14KernelTilesMulPiS_S_iii .globl _Z14KernelTilesMulPiS_S_iii .p2align 8 .type _Z14KernelTilesMulPiS_S_iii,@function _Z14KernelTilesMulPiS_S_iii: s_clause 0x1 s_load_b64 s[8:9], s[0:1], 0x18 s_load_b32 s3, s[0:1], 0x20 v_bfe_u32 v6, v0, 10, 10 v_and_b32_e32 v1, 0x3ff, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_lshl_add_u32 v5, s15, 4, v6 v_lshl_add_u32 v0, s14, 4, v1 s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cmp_gt_i32_e32 vcc_lo, s8, v5 v_cmp_gt_i32_e64 s2, s3, v0 s_cmp_lt_i32 s9, 1 s_cbranch_scc1 .LBB0_13 s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v2, 2, v1 v_lshlrev_b32_e32 v7, 6, v6 s_add_i32 s10, s9, 15 s_mov_b32 s11, 0 v_mov_b32_e32 v10, 0 v_add_nc_u32_e32 v8, 0x400, v2 v_mad_u64_u32 v[3:4], null, v5, s9, v[1:2] v_dual_mov_b32 v2, 0 :: v_dual_add_nc_u32 v9, v7, v2 s_delay_alu instid0(VALU_DEP_3) v_add_nc_u32_e32 v4, v8, v7 s_lshr_b32 s10, s10, 4 s_xor_b32 s12, vcc_lo, -1 s_max_i32 s10, s10, 1 s_xor_b32 s2, s2, -1 .LBB0_2: s_lshl_b32 s13, s11, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v11, s13, v1 v_cmp_le_i32_e32 vcc_lo, s9, v11 s_or_b32 s14, vcc_lo, s12 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_saveexec_b32 s15, s14 s_xor_b32 s14, exec_lo, s15 s_cbranch_execz .LBB0_4 ds_store_b32 v9, v10 .LBB0_4: s_and_not1_saveexec_b32 s14, s14 s_cbranch_execz .LBB0_6 v_add_nc_u32_e32 v11, s13, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v12, 31, v11 v_lshlrev_b64 v[11:12], 2, v[11:12] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v11, vcc_lo, s4, v11 v_add_co_ci_u32_e32 v12, vcc_lo, s5, v12, vcc_lo global_load_b32 v11, v[11:12], off s_waitcnt vmcnt(0) ds_store_b32 v9, v11 .LBB0_6: s_or_b32 exec_lo, exec_lo, s14 v_add_nc_u32_e32 v11, s13, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_cmp_le_i32_e32 vcc_lo, s9, v11 s_or_b32 s13, vcc_lo, s2 s_and_saveexec_b32 s14, s13 s_delay_alu instid0(SALU_CYCLE_1) s_xor_b32 s13, exec_lo, s14 s_cbranch_execz .LBB0_8 ds_store_b32 v4, v10 .LBB0_8: s_and_not1_saveexec_b32 s13, s13 s_cbranch_execz .LBB0_10 v_mad_u64_u32 v[12:13], null, v11, s3, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v13, 31, v12 v_lshlrev_b64 v[11:12], 2, v[12:13] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v11, vcc_lo, s6, v11 v_add_co_ci_u32_e32 v12, vcc_lo, s7, v12, vcc_lo global_load_b32 v11, v[11:12], off s_waitcnt vmcnt(0) ds_store_b32 v4, v11 .LBB0_10: s_or_b32 exec_lo, exec_lo, s13 v_mov_b32_e32 v11, v8 s_mov_b32 s13, 0 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_11: v_add_nc_u32_e32 v12, s13, v7 s_add_i32 s13, s13, 4 ds_load_b32 v14, v11 ds_load_b32 v15, v12 s_cmp_eq_u32 s13, 64 s_waitcnt lgkmcnt(0) v_mad_u64_u32 v[12:13], null, v14, v15, v[2:3] s_delay_alu instid0(VALU_DEP_1) v_dual_mov_b32 v2, v12 :: v_dual_add_nc_u32 v11, 64, v11 s_cbranch_scc0 .LBB0_11 s_add_i32 s11, s11, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s11, s10 s_barrier buffer_gl0_inv s_cbranch_scc0 .LBB0_2 s_branch .LBB0_14 .LBB0_13: v_mov_b32_e32 v2, 0 .LBB0_14: v_cmp_gt_i32_e32 vcc_lo, s8, v5 v_cmp_gt_i32_e64 s2, s3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, vcc_lo, s2 s_and_saveexec_b32 s4, s2 s_cbranch_execz .LBB0_16 s_load_b64 s[0:1], s[0:1], 0x10 v_mad_u64_u32 v[3:4], null, v5, s3, 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_16: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14KernelTilesMulPiS_S_iii .amdhsa_group_segment_fixed_size 2048 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 36 .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 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z14KernelTilesMulPiS_S_iii, .Lfunc_end0-_Z14KernelTilesMulPiS_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 .group_segment_fixed_size: 2048 .kernarg_segment_align: 8 .kernarg_segment_size: 36 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14KernelTilesMulPiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z14KernelTilesMulPiS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void KernelTilesMul(int *Mat1,int *Mat2,int *Mat3,int rowM1,int colM1,int colM2){ __shared__ int Mds[TILE_WIDTH][TILE_WIDTH]; __shared__ int Nds[TILE_WIDTH][TILE_WIDTH]; int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int row = by * TILE_WIDTH + ty; int col = bx * TILE_WIDTH + tx; int Pvalue = 0; for(int k = 0; k < (colM1+TILE_WIDTH-1)/(TILE_WIDTH); ++k){ if(k*TILE_WIDTH + tx < colM1 && row < rowM1){ Mds[ty][tx] = Mat1[row*colM1 + k*TILE_WIDTH + tx]; }else{ Mds[ty][tx] = 0; } if(k*TILE_WIDTH + ty < colM1 && col < colM2){ Nds[ty][tx] = Mat2[(k*TILE_WIDTH + ty) * colM2 + col]; }else{ Nds[ty][tx] =0; } __syncthreads(); for(int k = 0; k < TILE_WIDTH; ++k){ Pvalue += Mds[ty][k] * Nds[k][tx]; } __syncthreads(); } if (row < rowM1 && col < colM2){ Mat3[row*colM2+col] = Pvalue; } }
.text .file "KernelTilesMul.hip" .globl _Z29__device_stub__KernelTilesMulPiS_S_iii # -- Begin function _Z29__device_stub__KernelTilesMulPiS_S_iii .p2align 4, 0x90 .type _Z29__device_stub__KernelTilesMulPiS_S_iii,@function _Z29__device_stub__KernelTilesMulPiS_S_iii: # @_Z29__device_stub__KernelTilesMulPiS_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 $_Z14KernelTilesMulPiS_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_end0: .size _Z29__device_stub__KernelTilesMulPiS_S_iii, .Lfunc_end0-_Z29__device_stub__KernelTilesMulPiS_S_iii .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 $_Z14KernelTilesMulPiS_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_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 _Z14KernelTilesMulPiS_S_iii,@object # @_Z14KernelTilesMulPiS_S_iii .section .rodata,"a",@progbits .globl _Z14KernelTilesMulPiS_S_iii .p2align 3, 0x0 _Z14KernelTilesMulPiS_S_iii: .quad _Z29__device_stub__KernelTilesMulPiS_S_iii .size _Z14KernelTilesMulPiS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14KernelTilesMulPiS_S_iii" .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__KernelTilesMulPiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14KernelTilesMulPiS_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 : _Z14KernelTilesMulPiS_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 R9, SR_CTAID.X ; /* 0x0000000000097919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R6, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff067624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe20000000a00 */ /*0040*/ HFMA2.MMA R21, -RZ, RZ, 0, 0 ; /* 0x00000000ff157435 */ /* 0x000fe200000001ff */ /*0050*/ S2R R12, SR_TID.X ; /* 0x00000000000c7919 */ /* 0x000e240000002100 */ /*0060*/ ISETP.GE.AND P1, PT, R6, 0x1, PT ; /* 0x000000010600780c */ /* 0x000fe40003f26270 */ /*0070*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e680000002600 */ /*0080*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0090*/ LEA R0, R9, R12, 0x4 ; /* 0x0000000c09007211 */ /* 0x001fc800078e20ff */ /*00a0*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x180], PT ; /* 0x0000600000007a0c */ /* 0x000fe40003f06270 */ /*00b0*/ LEA R3, R3, R2, 0x4 ; /* 0x0000000203037211 */ /* 0x002fc800078e20ff */ /*00c0*/ ISETP.GE.OR P0, PT, R3, c[0x0][0x178], P0 ; /* 0x00005e0003007a0c */ /* 0x000fe20000706670 */ /*00d0*/ @!P1 BRA 0x5b0 ; /* 0x000004d000009947 */ /* 0x000fd80003800000 */ /*00e0*/ IADD3 R6, R6, 0xf, RZ ; /* 0x0000000f06067810 */ /* 0x000fe20007ffe0ff */ /*00f0*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe200078e00ff */ /*0100*/ SHF.L.U32 R15, R2.reuse, 0x6, RZ ; /* 0x00000006020f7819 */ /* 0x040fe200000006ff */ /*0110*/ IMAD R4, R3, c[0x0][0x17c], R12 ; /* 0x00005f0003047a24 */ /* 0x000fe200078e020c */ /*0120*/ SHF.R.S32.HI R7, RZ, 0x1f, R6 ; /* 0x0000001fff077819 */ /* 0x000fe20000011406 */ /*0130*/ IMAD R16, R2, c[0x0][0x180], R12 ; /* 0x0000600002107a24 */ /* 0x000fe200078e020c */ /*0140*/ MOV R21, RZ ; /* 0x000000ff00157202 */ /* 0x000fe20000000f00 */ /*0150*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe200078e0205 */ /*0160*/ LEA.HI R7, R7, R6, RZ, 0x4 ; /* 0x0000000607077211 */ /* 0x000fe200078f20ff */ /*0170*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0180*/ LEA R16, R9, R16, 0x4 ; /* 0x0000001009107211 */ /* 0x000fe200078e20ff */ /*0190*/ IMAD.MOV.U32 R14, RZ, RZ, R4 ; /* 0x000000ffff0e7224 */ /* 0x000fe200078e0004 */ /*01a0*/ MOV R13, R5 ; /* 0x00000005000d7202 */ /* 0x000fe20000000f00 */ /*01b0*/ IMAD R19, R12.reuse, 0x4, R15 ; /* 0x000000040c137824 */ /* 0x040fe200078e020f */ /*01c0*/ LEA R18, R12, 0x400, 0x2 ; /* 0x000004000c127811 */ /* 0x000fc400078e10ff */ /*01d0*/ SHF.R.S32.HI R17, RZ, 0x4, R7 ; /* 0x00000004ff117819 */ /* 0x000fe40000011407 */ /*01e0*/ ISETP.GE.AND P1, PT, R2, c[0x0][0x17c], PT ; /* 0x00005f0002007a0c */ /* 0x000fe20003f26270 */ /*01f0*/ HFMA2.MMA R22, -RZ, RZ, 0, 0 ; /* 0x00000000ff167435 */ /* 0x000fe200000001ff */ /*0200*/ ISETP.GE.AND P2, PT, R12, c[0x0][0x17c], PT ; /* 0x00005f000c007a0c */ /* 0x000fe20003f46270 */ /*0210*/ IMAD.MOV.U32 R28, RZ, RZ, RZ ; /* 0x000000ffff1c7224 */ /* 0x000fe200078e00ff */ /*0220*/ ISETP.GE.OR P1, PT, R0, c[0x0][0x180], P1 ; /* 0x0000600000007a0c */ /* 0x000fe40000f26670 */ /*0230*/ ISETP.GE.OR P2, PT, R3, c[0x0][0x178], P2 ; /* 0x00005e0003007a0c */ /* 0x000fd60001746670 */ /*0240*/ @!P1 MOV R25, 0x4 ; /* 0x0000000400199802 */ /* 0x000fe40000000f00 */ /*0250*/ @!P2 MOV R4, R14 ; /* 0x0000000e0004a202 */ /* 0x000fe40000000f00 */ /*0260*/ @!P2 MOV R5, R13 ; /* 0x0000000d0005a202 */ /* 0x000fe20000000f00 */ /*0270*/ @!P1 IMAD.WIDE R24, R16, R25, c[0x0][0x168] ; /* 0x00005a0010189625 */ /* 0x000fc800078e0219 */ /*0280*/ @!P2 LDG.E R22, [R4.64] ; /* 0x000000060416a981 */ /* 0x000ea8000c1e1900 */ /*0290*/ @!P1 LDG.E R28, [R24.64] ; /* 0x00000006181c9981 */ /* 0x000ee2000c1e1900 */ /*02a0*/ UIADD3 UR4, UR4, 0x1, URZ ; /* 0x0000000104047890 */ /* 0x000fe2000fffe03f */ /*02b0*/ IADD3 R14, P2, R14, 0x40, RZ ; /* 0x000000400e0e7810 */ /* 0x000fe40007f5e0ff */ /*02c0*/ IADD3 R2, R2, 0x10, RZ ; /* 0x0000001002027810 */ /* 0x000fe40007ffe0ff */ /*02d0*/ IADD3 R12, R12, 0x10, RZ ; /* 0x000000100c0c7810 */ /* 0x000fc40007ffe0ff */ /*02e0*/ ISETP.LE.AND P1, PT, R17, UR4, PT ; /* 0x0000000411007c0c */ /* 0x000fe4000bf23270 */ /*02f0*/ IADD3.X R13, RZ, R13, RZ, P2, !PT ; /* 0x0000000dff0d7210 */ /* 0x000fe200017fe4ff */ /*0300*/ STS [R19], R22 ; /* 0x0000001613007388 */ /* 0x004fe80000000800 */ /*0310*/ STS [R19+0x400], R28 ; /* 0x0004001c13007388 */ /* 0x008fe80000000800 */ /*0320*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0330*/ LDS R29, [R18] ; /* 0x00000000121d7984 */ /* 0x000fe80000000800 */ /*0340*/ LDS.128 R8, [R15] ; /* 0x000000000f087984 */ /* 0x000e280000000c00 */ /*0350*/ LDS R24, [R18+0x40] ; /* 0x0000400012187984 */ /* 0x000e680000000800 */ /*0360*/ LDS R27, [R18+0x80] ; /* 0x00008000121b7984 */ /* 0x000ea80000000800 */ /*0370*/ LDS R26, [R18+0xc0] ; /* 0x0000c000121a7984 */ /* 0x000ee80000000800 */ /*0380*/ LDS R23, [R18+0x100] ; /* 0x0001000012177984 */ /* 0x000fe80000000800 */ /*0390*/ LDS.128 R4, [R15+0x10] ; /* 0x000010000f047984 */ /* 0x000f280000000c00 */ /*03a0*/ LDS R20, [R18+0x140] ; /* 0x0001400012147984 */ /* 0x000f680000000800 */ /*03b0*/ LDS R25, [R18+0x180] ; /* 0x0001800012197984 */ /* 0x000f680000000800 */ /*03c0*/ LDS R22, [R18+0x1c0] ; /* 0x0001c00012167984 */ /* 0x000f620000000800 */ /*03d0*/ IMAD R8, R29, R8, R21 ; /* 0x000000081d087224 */ /* 0x001fc600078e0215 */ /*03e0*/ LDS R21, [R18+0x200] ; /* 0x0002000012157984 */ /* 0x000fe20000000800 */ /*03f0*/ IMAD R8, R24, R9, R8 ; /* 0x0000000918087224 */ /* 0x002fc600078e0208 */ /*0400*/ LDS R24, [R18+0x240] ; /* 0x0002400012187984 */ /* 0x000fe20000000800 */ /*0410*/ IMAD R8, R27, R10, R8 ; /* 0x0000000a1b087224 */ /* 0x004fc800078e0208 */ /*0420*/ IMAD R26, R26, R11, R8 ; /* 0x0000000b1a1a7224 */ /* 0x008fe400078e0208 */ /*0430*/ LDS.128 R8, [R15+0x20] ; /* 0x000020000f087984 */ /* 0x000e240000000c00 */ /*0440*/ IMAD R4, R23, R4, R26 ; /* 0x0000000417047224 */ /* 0x010fe400078e021a */ /*0450*/ LDS R23, [R18+0x280] ; /* 0x0002800012177984 */ /* 0x000e640000000800 */ /*0460*/ IMAD R4, R20, R5, R4 ; /* 0x0000000514047224 */ /* 0x020fe400078e0204 */ /*0470*/ LDS R20, [R18+0x2c0] ; /* 0x0002c00012147984 */ /* 0x000ea40000000800 */ /*0480*/ IMAD R4, R25, R6, R4 ; /* 0x0000000619047224 */ /* 0x000fc400078e0204 */ /*0490*/ LDS R25, [R18+0x300] ; /* 0x0003000012197984 */ /* 0x000fe40000000800 */ /*04a0*/ IMAD R26, R22, R7, R4 ; /* 0x00000007161a7224 */ /* 0x000fe400078e0204 */ /*04b0*/ LDS.128 R4, [R15+0x30] ; /* 0x000030000f047984 */ /* 0x000ee80000000c00 */ /*04c0*/ LDS R22, [R18+0x340] ; /* 0x0003400012167984 */ /* 0x000f220000000800 */ /*04d0*/ IMAD R26, R21, R8, R26 ; /* 0x00000008151a7224 */ /* 0x001fc600078e021a */ /*04e0*/ LDS R21, [R18+0x380] ; /* 0x0003800012157984 */ /* 0x000e220000000800 */ /*04f0*/ IMAD R9, R24, R9, R26 ; /* 0x0000000918097224 */ /* 0x000fc600078e021a */ /*0500*/ LDS R8, [R18+0x3c0] ; /* 0x0003c00012087984 */ /* 0x000f620000000800 */ /*0510*/ IMAD R9, R23, R10, R9 ; /* 0x0000000a17097224 */ /* 0x002fc800078e0209 */ /*0520*/ IMAD R9, R20, R11, R9 ; /* 0x0000000b14097224 */ /* 0x004fc800078e0209 */ /*0530*/ IMAD R4, R25, R4, R9 ; /* 0x0000000419047224 */ /* 0x008fc800078e0209 */ /*0540*/ IMAD R4, R22, R5, R4 ; /* 0x0000000516047224 */ /* 0x010fe400078e0204 */ /*0550*/ IMAD.MOV.U32 R5, RZ, RZ, 0x10 ; /* 0x00000010ff057424 */ /* 0x000fc800078e00ff */ /*0560*/ IMAD R16, R5, c[0x0][0x180], R16 ; /* 0x0000600005107a24 */ /* 0x000fe400078e0210 */ /*0570*/ IMAD R21, R21, R6, R4 ; /* 0x0000000615157224 */ /* 0x001fc800078e0204 */ /*0580*/ IMAD R21, R8, R7, R21 ; /* 0x0000000708157224 */ /* 0x020fe200078e0215 */ /*0590*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*05a0*/ @!P1 BRA 0x1e0 ; /* 0xfffffc3000009947 */ /* 0x000fea000383ffff */ /*05b0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*05c0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fe200000001ff */ /*05d0*/ IMAD R3, R3, c[0x0][0x180], R0 ; /* 0x0000600003037a24 */ /* 0x000fd200078e0200 */ /*05e0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x170] ; /* 0x00005c0003027625 */ /* 0x000fca00078e0202 */ /*05f0*/ STG.E [R2.64], R21 ; /* 0x0000001502007986 */ /* 0x000fe2000c101906 */ /*0600*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0610*/ BRA 0x610; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0620*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0630*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0640*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0650*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0660*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0670*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0680*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0690*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14KernelTilesMulPiS_S_iii .globl _Z14KernelTilesMulPiS_S_iii .p2align 8 .type _Z14KernelTilesMulPiS_S_iii,@function _Z14KernelTilesMulPiS_S_iii: s_clause 0x1 s_load_b64 s[8:9], s[0:1], 0x18 s_load_b32 s3, s[0:1], 0x20 v_bfe_u32 v6, v0, 10, 10 v_and_b32_e32 v1, 0x3ff, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_lshl_add_u32 v5, s15, 4, v6 v_lshl_add_u32 v0, s14, 4, v1 s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cmp_gt_i32_e32 vcc_lo, s8, v5 v_cmp_gt_i32_e64 s2, s3, v0 s_cmp_lt_i32 s9, 1 s_cbranch_scc1 .LBB0_13 s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v2, 2, v1 v_lshlrev_b32_e32 v7, 6, v6 s_add_i32 s10, s9, 15 s_mov_b32 s11, 0 v_mov_b32_e32 v10, 0 v_add_nc_u32_e32 v8, 0x400, v2 v_mad_u64_u32 v[3:4], null, v5, s9, v[1:2] v_dual_mov_b32 v2, 0 :: v_dual_add_nc_u32 v9, v7, v2 s_delay_alu instid0(VALU_DEP_3) v_add_nc_u32_e32 v4, v8, v7 s_lshr_b32 s10, s10, 4 s_xor_b32 s12, vcc_lo, -1 s_max_i32 s10, s10, 1 s_xor_b32 s2, s2, -1 .LBB0_2: s_lshl_b32 s13, s11, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v11, s13, v1 v_cmp_le_i32_e32 vcc_lo, s9, v11 s_or_b32 s14, vcc_lo, s12 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_saveexec_b32 s15, s14 s_xor_b32 s14, exec_lo, s15 s_cbranch_execz .LBB0_4 ds_store_b32 v9, v10 .LBB0_4: s_and_not1_saveexec_b32 s14, s14 s_cbranch_execz .LBB0_6 v_add_nc_u32_e32 v11, s13, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v12, 31, v11 v_lshlrev_b64 v[11:12], 2, v[11:12] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v11, vcc_lo, s4, v11 v_add_co_ci_u32_e32 v12, vcc_lo, s5, v12, vcc_lo global_load_b32 v11, v[11:12], off s_waitcnt vmcnt(0) ds_store_b32 v9, v11 .LBB0_6: s_or_b32 exec_lo, exec_lo, s14 v_add_nc_u32_e32 v11, s13, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_cmp_le_i32_e32 vcc_lo, s9, v11 s_or_b32 s13, vcc_lo, s2 s_and_saveexec_b32 s14, s13 s_delay_alu instid0(SALU_CYCLE_1) s_xor_b32 s13, exec_lo, s14 s_cbranch_execz .LBB0_8 ds_store_b32 v4, v10 .LBB0_8: s_and_not1_saveexec_b32 s13, s13 s_cbranch_execz .LBB0_10 v_mad_u64_u32 v[12:13], null, v11, s3, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v13, 31, v12 v_lshlrev_b64 v[11:12], 2, v[12:13] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v11, vcc_lo, s6, v11 v_add_co_ci_u32_e32 v12, vcc_lo, s7, v12, vcc_lo global_load_b32 v11, v[11:12], off s_waitcnt vmcnt(0) ds_store_b32 v4, v11 .LBB0_10: s_or_b32 exec_lo, exec_lo, s13 v_mov_b32_e32 v11, v8 s_mov_b32 s13, 0 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_11: v_add_nc_u32_e32 v12, s13, v7 s_add_i32 s13, s13, 4 ds_load_b32 v14, v11 ds_load_b32 v15, v12 s_cmp_eq_u32 s13, 64 s_waitcnt lgkmcnt(0) v_mad_u64_u32 v[12:13], null, v14, v15, v[2:3] s_delay_alu instid0(VALU_DEP_1) v_dual_mov_b32 v2, v12 :: v_dual_add_nc_u32 v11, 64, v11 s_cbranch_scc0 .LBB0_11 s_add_i32 s11, s11, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s11, s10 s_barrier buffer_gl0_inv s_cbranch_scc0 .LBB0_2 s_branch .LBB0_14 .LBB0_13: v_mov_b32_e32 v2, 0 .LBB0_14: v_cmp_gt_i32_e32 vcc_lo, s8, v5 v_cmp_gt_i32_e64 s2, s3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, vcc_lo, s2 s_and_saveexec_b32 s4, s2 s_cbranch_execz .LBB0_16 s_load_b64 s[0:1], s[0:1], 0x10 v_mad_u64_u32 v[3:4], null, v5, s3, 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_16: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14KernelTilesMulPiS_S_iii .amdhsa_group_segment_fixed_size 2048 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 36 .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 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z14KernelTilesMulPiS_S_iii, .Lfunc_end0-_Z14KernelTilesMulPiS_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 .group_segment_fixed_size: 2048 .kernarg_segment_align: 8 .kernarg_segment_size: 36 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14KernelTilesMulPiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z14KernelTilesMulPiS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .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_000b4cbc_00000000-6_KernelTilesMul.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__Z14KernelTilesMulPiS_S_iiiPiS_S_iii .type _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii, @function _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii: .LFB2051: .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 .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14KernelTilesMulPiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii, .-_Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii .globl _Z14KernelTilesMulPiS_S_iii .type _Z14KernelTilesMulPiS_S_iii, @function _Z14KernelTilesMulPiS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z41__device_stub__Z14KernelTilesMulPiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z14KernelTilesMulPiS_S_iii, .-_Z14KernelTilesMulPiS_S_iii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14KernelTilesMulPiS_S_iii" .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 _Z14KernelTilesMulPiS_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 .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 "KernelTilesMul.hip" .globl _Z29__device_stub__KernelTilesMulPiS_S_iii # -- Begin function _Z29__device_stub__KernelTilesMulPiS_S_iii .p2align 4, 0x90 .type _Z29__device_stub__KernelTilesMulPiS_S_iii,@function _Z29__device_stub__KernelTilesMulPiS_S_iii: # @_Z29__device_stub__KernelTilesMulPiS_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 $_Z14KernelTilesMulPiS_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_end0: .size _Z29__device_stub__KernelTilesMulPiS_S_iii, .Lfunc_end0-_Z29__device_stub__KernelTilesMulPiS_S_iii .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 $_Z14KernelTilesMulPiS_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_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 _Z14KernelTilesMulPiS_S_iii,@object # @_Z14KernelTilesMulPiS_S_iii .section .rodata,"a",@progbits .globl _Z14KernelTilesMulPiS_S_iii .p2align 3, 0x0 _Z14KernelTilesMulPiS_S_iii: .quad _Z29__device_stub__KernelTilesMulPiS_S_iii .size _Z14KernelTilesMulPiS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14KernelTilesMulPiS_S_iii" .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__KernelTilesMulPiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14KernelTilesMulPiS_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 device assembly.
extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
code for sm_80 Function : add .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 R12, SR_CTAID.X ; /* 0x00000000000c7919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R12, R12, c[0x0][0x0], R3 ; /* 0x000000000c0c7a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R12, c[0x0][0x160], PT ; /* 0x000058000c007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R13, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0d7435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R4, R12, R13, c[0x0][0x178] ; /* 0x00005e000c047625 */ /* 0x000fc800078e020d */ /*0090*/ IMAD.WIDE R2, R12.reuse, R13.reuse, c[0x0][0x168] ; /* 0x00005a000c027625 */ /* 0x0c0fe400078e020d */ /*00a0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ IMAD.WIDE R6, R12, R13, c[0x0][0x188] ; /* 0x000062000c067625 */ /* 0x000fc800078e020d */ /*00d0*/ IMAD.WIDE R10, R12, R13, c[0x0][0x180] ; /* 0x000060000c0a7625 */ /* 0x000fc800078e020d */ /*00e0*/ IMAD.WIDE R8, R12, R13, c[0x0][0x170] ; /* 0x00005c000c087625 */ /* 0x000fc800078e020d */ /*00f0*/ FADD R15, R4, R3 ; /* 0x00000003040f7221 */ /* 0x004fca0000000000 */ /*0100*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x000fe8000c101904 */ /*0110*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R9, [R8.64] ; /* 0x0000000408097981 */ /* 0x000ea2000c1e1900 */ /*0130*/ IMAD.WIDE R12, R12, R13, c[0x0][0x190] ; /* 0x000064000c0c7625 */ /* 0x000fc800078e020d */ /*0140*/ FADD R17, R10, R9 ; /* 0x000000090a117221 */ /* 0x004fca0000000000 */ /*0150*/ STG.E [R12.64], R17 ; /* 0x000000110c007986 */ /* 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.
extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
.file "tmpxft_00104bc7_00000000-6_JCudaVectorAddKernel.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 _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ .type _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_, @function _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movl %edi, 60(%rsp) movq %rsi, 48(%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 60(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) 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 add(%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 _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_, .-_Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ .globl add .type add, @function add: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size add, .-add .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "add" .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 add(%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.
extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
#include <hip/hip_runtime.h> extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected add .globl add .p2align 8 .type add,@function add: s_clause 0x1 s_load_b32 s2, s[0:1], 0x44 s_load_b32 s3, s[0:1], 0x0 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_2 s_load_b256 s[4:11], s[0:1], 0x8 v_ashrrev_i32_e32 v2, 31, v1 s_load_b128 s[0:3], s[0:1], 0x28 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) 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, s8, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s9, v1, vcc_lo global_load_b32 v6, v[2:3], off global_load_b32 v7, v[4:5], off v_add_co_u32 v2, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s1, 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 s_waitcnt vmcnt(0) v_add_f32_e32 v8, v6, v7 v_add_co_u32 v6, vcc_lo, s10, v0 v_add_co_ci_u32_e32 v7, vcc_lo, s11, v1, vcc_lo global_store_b32 v[2:3], v8, off global_load_b32 v2, v[4:5], off global_load_b32 v3, v[6:7], off v_add_co_u32 v0, vcc_lo, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo s_waitcnt vmcnt(0) v_add_f32_e32 v2, v2, v3 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 add .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 312 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 9 .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 add, .Lfunc_end0-add .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .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 - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 48 .size: 8 .value_kind: global_buffer - .offset: 56 .size: 4 .value_kind: hidden_block_count_x - .offset: 60 .size: 4 .value_kind: hidden_block_count_y - .offset: 64 .size: 4 .value_kind: hidden_block_count_z - .offset: 68 .size: 2 .value_kind: hidden_group_size_x - .offset: 70 .size: 2 .value_kind: hidden_group_size_y - .offset: 72 .size: 2 .value_kind: hidden_group_size_z - .offset: 74 .size: 2 .value_kind: hidden_remainder_x - .offset: 76 .size: 2 .value_kind: hidden_remainder_y - .offset: 78 .size: 2 .value_kind: hidden_remainder_z - .offset: 96 .size: 8 .value_kind: hidden_global_offset_x - .offset: 104 .size: 8 .value_kind: hidden_global_offset_y - .offset: 112 .size: 8 .value_kind: hidden_global_offset_z - .offset: 120 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 312 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: add .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: add.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 9 .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> extern "C" __global__ void add(int n, float *ax, float *ay, float *bx, float *by, float *sumx, float *sumy) { int i = blockIdx.x * blockDim.x + threadIdx.x; if (i<n) { sumx[i] = ax[i] + bx[i]; sumy[i] = ay[i] + by[i]; } }
.text .file "JCudaVectorAddKernel.hip" .globl __device_stub__add # -- Begin function __device_stub__add .p2align 4, 0x90 .type __device_stub__add,@function __device_stub__add: # @__device_stub__add .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 4(%rsp) movq %rsi, 88(%rsp) movq %rdx, 80(%rsp) movq %rcx, 72(%rsp) movq %r8, 64(%rsp) movq %r9, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 96(%rsp) leaq 88(%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 $add, %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 __device_stub__add, .Lfunc_end0-__device_stub__add .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 $add, %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 add,@object # @add .section .rodata,"a",@progbits .globl add .p2align 3, 0x0 add: .quad __device_stub__add .size add, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "add" .size .L__unnamed_1, 4 .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__add .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym add .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 : add .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 R12, SR_CTAID.X ; /* 0x00000000000c7919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R12, R12, c[0x0][0x0], R3 ; /* 0x000000000c0c7a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R12, c[0x0][0x160], PT ; /* 0x000058000c007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R13, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0d7435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R4, R12, R13, c[0x0][0x178] ; /* 0x00005e000c047625 */ /* 0x000fc800078e020d */ /*0090*/ IMAD.WIDE R2, R12.reuse, R13.reuse, c[0x0][0x168] ; /* 0x00005a000c027625 */ /* 0x0c0fe400078e020d */ /*00a0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ IMAD.WIDE R6, R12, R13, c[0x0][0x188] ; /* 0x000062000c067625 */ /* 0x000fc800078e020d */ /*00d0*/ IMAD.WIDE R10, R12, R13, c[0x0][0x180] ; /* 0x000060000c0a7625 */ /* 0x000fc800078e020d */ /*00e0*/ IMAD.WIDE R8, R12, R13, c[0x0][0x170] ; /* 0x00005c000c087625 */ /* 0x000fc800078e020d */ /*00f0*/ FADD R15, R4, R3 ; /* 0x00000003040f7221 */ /* 0x004fca0000000000 */ /*0100*/ STG.E [R6.64], R15 ; /* 0x0000000f06007986 */ /* 0x000fe8000c101904 */ /*0110*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ea8000c1e1900 */ /*0120*/ LDG.E R9, [R8.64] ; /* 0x0000000408097981 */ /* 0x000ea2000c1e1900 */ /*0130*/ IMAD.WIDE R12, R12, R13, c[0x0][0x190] ; /* 0x000064000c0c7625 */ /* 0x000fc800078e020d */ /*0140*/ FADD R17, R10, R9 ; /* 0x000000090a117221 */ /* 0x004fca0000000000 */ /*0150*/ STG.E [R12.64], R17 ; /* 0x000000110c007986 */ /* 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 add .globl add .p2align 8 .type add,@function add: s_clause 0x1 s_load_b32 s2, s[0:1], 0x44 s_load_b32 s3, s[0:1], 0x0 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_2 s_load_b256 s[4:11], s[0:1], 0x8 v_ashrrev_i32_e32 v2, 31, v1 s_load_b128 s[0:3], s[0:1], 0x28 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) 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, s8, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s9, v1, vcc_lo global_load_b32 v6, v[2:3], off global_load_b32 v7, v[4:5], off v_add_co_u32 v2, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v3, vcc_lo, s1, 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 s_waitcnt vmcnt(0) v_add_f32_e32 v8, v6, v7 v_add_co_u32 v6, vcc_lo, s10, v0 v_add_co_ci_u32_e32 v7, vcc_lo, s11, v1, vcc_lo global_store_b32 v[2:3], v8, off global_load_b32 v2, v[4:5], off global_load_b32 v3, v[6:7], off v_add_co_u32 v0, vcc_lo, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo s_waitcnt vmcnt(0) v_add_f32_e32 v2, v2, v3 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 add .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 312 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 9 .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 add, .Lfunc_end0-add .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .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 - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 48 .size: 8 .value_kind: global_buffer - .offset: 56 .size: 4 .value_kind: hidden_block_count_x - .offset: 60 .size: 4 .value_kind: hidden_block_count_y - .offset: 64 .size: 4 .value_kind: hidden_block_count_z - .offset: 68 .size: 2 .value_kind: hidden_group_size_x - .offset: 70 .size: 2 .value_kind: hidden_group_size_y - .offset: 72 .size: 2 .value_kind: hidden_group_size_z - .offset: 74 .size: 2 .value_kind: hidden_remainder_x - .offset: 76 .size: 2 .value_kind: hidden_remainder_y - .offset: 78 .size: 2 .value_kind: hidden_remainder_z - .offset: 96 .size: 8 .value_kind: hidden_global_offset_x - .offset: 104 .size: 8 .value_kind: hidden_global_offset_y - .offset: 112 .size: 8 .value_kind: hidden_global_offset_z - .offset: 120 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 312 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: add .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: add.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 9 .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_00104bc7_00000000-6_JCudaVectorAddKernel.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 _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ .type _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_, @function _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movl %edi, 60(%rsp) movq %rsi, 48(%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 60(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) 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 add(%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 _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_, .-_Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ .globl add .type add, @function add: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z33__device_stub__Z3addiPfS_S_S_S_S_iPfS_S_S_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size add, .-add .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "add" .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 add(%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 "JCudaVectorAddKernel.hip" .globl __device_stub__add # -- Begin function __device_stub__add .p2align 4, 0x90 .type __device_stub__add,@function __device_stub__add: # @__device_stub__add .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 4(%rsp) movq %rsi, 88(%rsp) movq %rdx, 80(%rsp) movq %rcx, 72(%rsp) movq %r8, 64(%rsp) movq %r9, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 96(%rsp) leaq 88(%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 $add, %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 __device_stub__add, .Lfunc_end0-__device_stub__add .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 $add, %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 add,@object # @add .section .rodata,"a",@progbits .globl add .p2align 3, 0x0 add: .quad __device_stub__add .size add, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "add" .size .L__unnamed_1, 4 .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__add .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym add .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
code for sm_80 Function : _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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.Y ; /* 0x0000000000007919 */ /* 0x000e280000002600 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e280000002500 */ /*0030*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e620000002100 */ /*0040*/ IMAD R0, R0, c[0x0][0xc], R3 ; /* 0x0000030000007a24 */ /* 0x001fc800078e0203 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R5 ; /* 0x0000000000007a24 */ /* 0x002fca00078e0205 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x16c], PT ; /* 0x00005b0000007a0c */ /* 0x000fda0003f06270 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ MOV R2, c[0x0][0x170] ; /* 0x00005c0000027a02 */ /* 0x000fe20000000f00 */ /*0090*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00a0*/ MOV R3, c[0x0][0x174] ; /* 0x00005d0000037a02 */ /* 0x000fca0000000f00 */ /*00b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fe200000001ff */ /*00d0*/ F2I.TRUNC.NTZ R5, R2 ; /* 0x0000000200057305 */ /* 0x004e24000020f100 */ /*00e0*/ IMAD R5, R5, c[0x0][0x16c], R0 ; /* 0x00005b0005057a24 */ /* 0x001fce00078e0200 */ /*00f0*/ IMAD.WIDE R4, R5, R6, c[0x0][0x160] ; /* 0x0000580005047625 */ /* 0x000fcc00078e0206 */ /*0100*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*0110*/ IMAD.WIDE R6, R0, R6, c[0x0][0x180] ; /* 0x0000600000067625 */ /* 0x000fca00078e0206 */ /*0120*/ STG.E [R6.64], R5 ; /* 0x0000000506007986 */ /* 0x004fe2000c101904 */ /*0130*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0140*/ BRA 0x140; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
.file "tmpxft_0001f87a_00000000-6_Matrix_getRow_FloatPointer_naive.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 _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii .type _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii, @function _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movq %rcx, 24(%rsp) movl %r8d, 20(%rsp) movl %r9d, 16(%rsp) movq 208(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 8(%rsp), %rax movq %rax, 160(%rsp) leaq 216(%rsp), %rax movq %rax, 168(%rsp) leaq 224(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii(%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 _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii, .-_Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .type _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, @function _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 pushq 40(%rsp) .cfi_def_cfa_offset 48 call _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii addq $40, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .-_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii" .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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 8 .type _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@function _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: s_clause 0x2 s_load_b32 s3, s[0:1], 0x30 s_load_b32 s4, s[0:1], 0x3c s_load_b32 s2, s[0:1], 0xc s_waitcnt lgkmcnt(0) s_mul_i32 s3, s3, s15 s_and_b32 s4, s4, 0xffff s_add_i32 s3, s3, s14 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s3, s4, v[0:1] s_mov_b32 s3, exec_lo v_cmpx_gt_i32_e64 s2, v1 s_cbranch_execz .LBB0_2 s_clause 0x2 s_load_b64 s[4:5], s[0:1], 0x10 s_load_b64 s[6:7], s[0:1], 0x20 s_load_b64 s[0:1], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_load_b32 s3, s[4:5], 0x0 s_waitcnt lgkmcnt(0) v_cvt_i32_f32_e32 v0, s3 s_delay_alu instid0(VALU_DEP_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[2:3], 2, v[2:3] v_add_co_u32 v2, vcc_lo, s0, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v3, vcc_lo, s1, v3, vcc_lo global_load_b32 v3, v[2:3], off v_ashrrev_i32_e32 v2, 31, v1 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 v0, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v1, vcc_lo s_waitcnt vmcnt(0) global_store_b32 v[0:1], v3, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 304 .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 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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .Lfunc_end0-_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .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 - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .offset: 40 .size: 4 .value_kind: by_value - .offset: 44 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: hidden_block_count_x - .offset: 52 .size: 4 .value_kind: hidden_block_count_y - .offset: 56 .size: 4 .value_kind: hidden_block_count_z - .offset: 60 .size: 2 .value_kind: hidden_group_size_x - .offset: 62 .size: 2 .value_kind: hidden_group_size_y - .offset: 64 .size: 2 .value_kind: hidden_group_size_z - .offset: 66 .size: 2 .value_kind: hidden_remainder_x - .offset: 68 .size: 2 .value_kind: hidden_remainder_y - .offset: 70 .size: 2 .value_kind: hidden_remainder_z - .offset: 88 .size: 8 .value_kind: hidden_global_offset_x - .offset: 96 .size: 8 .value_kind: hidden_global_offset_y - .offset: 104 .size: 8 .value_kind: hidden_global_offset_z - .offset: 112 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 304 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void Matrix_getRow_FloatPointer_naive(const float * A , int Acount, int Acols, const float * rowId , int empty_par1, int empty_par2, float * out0 , int out0count, int out0cols) { int id = blockDim.x*blockIdx.y*gridDim.x + blockDim.x*blockIdx.x + threadIdx.x; if (id<Acols) { out0[id] = A[id + (int)rowId[0]*Acols]; } }
.text .file "Matrix_getRow_FloatPointer_naive.hip" .globl _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii # -- Begin function _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 4, 0x90 .type _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@function _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: # @_Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) leaq 168(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%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 $_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .Lfunc_end0-_Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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 $_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, %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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@object # @_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .rodata,"a",@progbits .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 3, 0x0 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: .quad _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .size _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii" .size .L__unnamed_1, 51 .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 _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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 : _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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.Y ; /* 0x0000000000007919 */ /* 0x000e280000002600 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e280000002500 */ /*0030*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e620000002100 */ /*0040*/ IMAD R0, R0, c[0x0][0xc], R3 ; /* 0x0000030000007a24 */ /* 0x001fc800078e0203 */ /*0050*/ IMAD R0, R0, c[0x0][0x0], R5 ; /* 0x0000000000007a24 */ /* 0x002fca00078e0205 */ /*0060*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x16c], PT ; /* 0x00005b0000007a0c */ /* 0x000fda0003f06270 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ MOV R2, c[0x0][0x170] ; /* 0x00005c0000027a02 */ /* 0x000fe20000000f00 */ /*0090*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*00a0*/ MOV R3, c[0x0][0x174] ; /* 0x00005d0000037a02 */ /* 0x000fca0000000f00 */ /*00b0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fe200000001ff */ /*00d0*/ F2I.TRUNC.NTZ R5, R2 ; /* 0x0000000200057305 */ /* 0x004e24000020f100 */ /*00e0*/ IMAD R5, R5, c[0x0][0x16c], R0 ; /* 0x00005b0005057a24 */ /* 0x001fce00078e0200 */ /*00f0*/ IMAD.WIDE R4, R5, R6, c[0x0][0x160] ; /* 0x0000580005047625 */ /* 0x000fcc00078e0206 */ /*0100*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*0110*/ IMAD.WIDE R6, R0, R6, c[0x0][0x180] ; /* 0x0000600000067625 */ /* 0x000fca00078e0206 */ /*0120*/ STG.E [R6.64], R5 ; /* 0x0000000506007986 */ /* 0x004fe2000c101904 */ /*0130*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0140*/ BRA 0x140; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 8 .type _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@function _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: s_clause 0x2 s_load_b32 s3, s[0:1], 0x30 s_load_b32 s4, s[0:1], 0x3c s_load_b32 s2, s[0:1], 0xc s_waitcnt lgkmcnt(0) s_mul_i32 s3, s3, s15 s_and_b32 s4, s4, 0xffff s_add_i32 s3, s3, s14 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s3, s4, v[0:1] s_mov_b32 s3, exec_lo v_cmpx_gt_i32_e64 s2, v1 s_cbranch_execz .LBB0_2 s_clause 0x2 s_load_b64 s[4:5], s[0:1], 0x10 s_load_b64 s[6:7], s[0:1], 0x20 s_load_b64 s[0:1], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_load_b32 s3, s[4:5], 0x0 s_waitcnt lgkmcnt(0) v_cvt_i32_f32_e32 v0, s3 s_delay_alu instid0(VALU_DEP_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[2:3], 2, v[2:3] v_add_co_u32 v2, vcc_lo, s0, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v3, vcc_lo, s1, v3, vcc_lo global_load_b32 v3, v[2:3], off v_ashrrev_i32_e32 v2, 31, v1 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 v0, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v1, vcc_lo s_waitcnt vmcnt(0) global_store_b32 v[0:1], v3, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 304 .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 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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .Lfunc_end0-_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .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 - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .offset: 40 .size: 4 .value_kind: by_value - .offset: 44 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: hidden_block_count_x - .offset: 52 .size: 4 .value_kind: hidden_block_count_y - .offset: 56 .size: 4 .value_kind: hidden_block_count_z - .offset: 60 .size: 2 .value_kind: hidden_group_size_x - .offset: 62 .size: 2 .value_kind: hidden_group_size_y - .offset: 64 .size: 2 .value_kind: hidden_group_size_z - .offset: 66 .size: 2 .value_kind: hidden_remainder_x - .offset: 68 .size: 2 .value_kind: hidden_remainder_y - .offset: 70 .size: 2 .value_kind: hidden_remainder_z - .offset: 88 .size: 8 .value_kind: hidden_global_offset_x - .offset: 96 .size: 8 .value_kind: hidden_global_offset_y - .offset: 104 .size: 8 .value_kind: hidden_global_offset_z - .offset: 112 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 304 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0001f87a_00000000-6_Matrix_getRow_FloatPointer_naive.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 _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii .type _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii, @function _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movq %rcx, 24(%rsp) movl %r8d, 20(%rsp) movl %r9d, 16(%rsp) movq 208(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 8(%rsp), %rax movq %rax, 160(%rsp) leaq 216(%rsp), %rax movq %rax, 168(%rsp) leaq 224(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii(%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 _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii, .-_Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .type _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, @function _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 40(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 pushq 40(%rsp) .cfi_def_cfa_offset 48 call _Z64__device_stub__Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfiiPKfiiS0_iiPfii addq $40, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .-_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii" .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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii(%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 "Matrix_getRow_FloatPointer_naive.hip" .globl _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii # -- Begin function _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 4, 0x90 .type _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@function _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: # @_Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) leaq 168(%rsp), %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%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 $_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, .Lfunc_end0-_Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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 $_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, %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 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii,@object # @_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .section .rodata,"a",@progbits .globl _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .p2align 3, 0x0 _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii: .quad _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .size _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii" .size .L__unnamed_1, 51 .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 _Z47__device_stub__Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z32Matrix_getRow_FloatPointer_naivePKfiiS0_iiPfii .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.
// compute.cu // // driver and kernel call #include <stdio.h> #define THREADS_PER_BLOCK 128 // __global__ void compute_2d (int secondArrSize, float *arr[]) __global__ void compute_2d ( int firstArrSize, int secondArrSize, float **arr) { int x = blockIdx.x * blockDim.x + threadIdx.x; int y = blockIdx.y * blockDim.y + threadIdx.y; if (x == 0 && x < firstArrSize && y == 0 && y < firstArrSize) { printf("Hello. I'm a thread %d in block %d \n", threadIdx.x, blockIdx.x); // printf("%lf \n", arr[x][y]); } // if (x <= arrSize) { // if (x % 2 == timeStep % 2 && x <= timeStep) // { // if (timeStep > timeSteps && x <= (timeStep - timeSteps - arrSize)) // { // } else // { // if (x == 0) // { // c_d[x] = (100.0 + c_d[x + 1]) / 2.0; // } else if (x == arrSize - 1) // { // c_d[x] = (c_d[x - 1] + c_d[x]) / 2.0; // } else // { // c_d[x] = (c_d[x - 1] + c_d[x + 1]) / 2.0; // } // } // } __syncthreads(); // } } extern "C" void compute2DArr (int firstArrSize, int secondArrSize, float *metalRod, int timeSteps) { int i = 0, j = 0; int size=firstArrSize*secondArrSize*sizeof(float); //allocate resources float **cell=(float**)malloc(size * 2); float **cell2=(float**)malloc(size * 2); for (i = 0; i < firstArrSize; i ++) { cell[i] = (float*)malloc(size); cell2[i] = (float*)malloc(size); for (j = 0; j < secondArrSize; j ++) { cell[i][j] = 23.0; } } size_t pitch; float **d_cell; cudaMallocPitch((void**) &d_cell, &pitch, secondArrSize * sizeof(float), firstArrSize); cudaError_t tmp = cudaMemcpy2D(d_cell, pitch, cell, secondArrSize * sizeof(float), secondArrSize * sizeof(float), firstArrSize, cudaMemcpyHostToDevice); if (cudaSuccess != tmp) { printf("\n copy to GPU \n"); printf(cudaGetErrorString(tmp)); } dim3 dimBlock(8,8); dim3 dimGrid(1,1); compute_2d<<<dimGrid, dimBlock>>>( firstArrSize, secondArrSize, d_cell); if (cudaSuccess != tmp) { printf("\n compute \n"); printf(cudaGetErrorString(tmp)); } tmp = cudaMemcpy2D(cell2, secondArrSize * sizeof(float), d_cell, pitch, secondArrSize * sizeof(float), firstArrSize, cudaMemcpyDeviceToHost); if (cudaSuccess != tmp) { printf("\n copy to CPU \n"); printf(cudaGetErrorString(tmp)); } for (i = 0; i < firstArrSize; i++) { for (j = 0; j < secondArrSize; j ++) { printf("\n %lf ", cell2[i][j]); } } // for (i = 0; i < (2*(timeSteps - 1)) + secondArrSize; i ++) // { // //compute_2d <<< ceil((float) secondArrSize/THREADS_PER_BLOCK), THREADS_PER_BLOCK >>> (c_d, secondArrSize, i, timeSteps); // } cudaError_t err = cudaGetLastError(); if (err != cudaSuccess) printf ("CUDA error: %s\n", cudaGetErrorString(err)); }
code for sm_80 Function : _Z10compute_2diiPPf .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fc800078e00ff */ /*0010*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e220000002600 */ /*0020*/ IADD3 R1, R1, -0x8, RZ ; /* 0xfffffff801017810 */ /* 0x000fc60007ffe0ff */ /*0030*/ S2R R0, SR_TID.Y ; /* 0x0000000000007919 */ /* 0x000e220000002200 */ /*0040*/ IADD3 R6, P1, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fc60007f3e0ff */ /*0050*/ S2R R9, SR_CTAID.X ; /* 0x0000000000097919 */ /* 0x000e640000002500 */ /*0060*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P1 ; /* 0x00000900ff077624 */ /* 0x000fe400008e06ff */ /*0070*/ S2R R8, SR_TID.X ; /* 0x0000000000087919 */ /* 0x000e620000002100 */ /*0080*/ IMAD R3, R3, c[0x0][0x4], R0 ; /* 0x0000010003037a24 */ /* 0x001fe400078e0200 */ /*0090*/ IMAD R0, R9, c[0x0][0x0], R8 ; /* 0x0000000009007a24 */ /* 0x002fca00078e0208 */ /*00a0*/ LOP3.LUT P0, RZ, R0, R3, RZ, 0xfc, !PT ; /* 0x0000000300ff7212 */ /* 0x000fc8000780fcff */ /*00b0*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x160], P0 ; /* 0x0000580000007a0c */ /* 0x000fc80000706670 */ /*00c0*/ ISETP.GE.OR P0, PT, R3, c[0x0][0x160], P0 ; /* 0x0000580003007a0c */ /* 0x000fda0000706670 */ /*00d0*/ @P0 BRA 0x1b0 ; /* 0x000000d000000947 */ /* 0x000fea0003800000 */ /*00e0*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe20000000f00 */ /*00f0*/ STL.64 [R1], R8 ; /* 0x0000000801007387 */ /* 0x0001e20000100a00 */ /*0100*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff047624 */ /* 0x000fe400078e00ff */ /*0110*/ LDC.64 R2, c[0x4][R0] ; /* 0x0100000000027b82 */ /* 0x0000620000000a00 */ /*0120*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff057624 */ /* 0x000fce00078e00ff */ /*0130*/ LEPC R8 ; /* 0x000000000008734e */ /* 0x001fe40000000000 */ /*0140*/ MOV R11, 0x1b0 ; /* 0x000001b0000b7802 */ /* 0x000fe40000000f00 */ /*0150*/ MOV R20, 0x130 ; /* 0x0000013000147802 */ /* 0x000fe40000000f00 */ /*0160*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fe40000000f00 */ /*0170*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fc40000000f00 */ /*0180*/ IADD3 R20, P0, P1, -R20, R11, R8 ; /* 0x0000000b14147210 */ /* 0x000fc8000791e108 */ /*0190*/ IADD3.X R21, ~R0, R21, R9, P0, P1 ; /* 0x0000001500157210 */ /* 0x000fc800007e2509 */ /*01a0*/ CALL.ABS.NOINC R2 ; /* 0x0000000002007343 */ /* 0x002fea0003c00000 */ /*01b0*/ WARPSYNC 0xffffffff ; /* 0xffffffff00007948 */ /* 0x000fe20003800000 */ /*01c0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*01d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*01e0*/ BRA 0x1e0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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.
// compute.cu // // driver and kernel call #include <stdio.h> #define THREADS_PER_BLOCK 128 // __global__ void compute_2d (int secondArrSize, float *arr[]) __global__ void compute_2d ( int firstArrSize, int secondArrSize, float **arr) { int x = blockIdx.x * blockDim.x + threadIdx.x; int y = blockIdx.y * blockDim.y + threadIdx.y; if (x == 0 && x < firstArrSize && y == 0 && y < firstArrSize) { printf("Hello. I'm a thread %d in block %d \n", threadIdx.x, blockIdx.x); // printf("%lf \n", arr[x][y]); } // if (x <= arrSize) { // if (x % 2 == timeStep % 2 && x <= timeStep) // { // if (timeStep > timeSteps && x <= (timeStep - timeSteps - arrSize)) // { // } else // { // if (x == 0) // { // c_d[x] = (100.0 + c_d[x + 1]) / 2.0; // } else if (x == arrSize - 1) // { // c_d[x] = (c_d[x - 1] + c_d[x]) / 2.0; // } else // { // c_d[x] = (c_d[x - 1] + c_d[x + 1]) / 2.0; // } // } // } __syncthreads(); // } } extern "C" void compute2DArr (int firstArrSize, int secondArrSize, float *metalRod, int timeSteps) { int i = 0, j = 0; int size=firstArrSize*secondArrSize*sizeof(float); //allocate resources float **cell=(float**)malloc(size * 2); float **cell2=(float**)malloc(size * 2); for (i = 0; i < firstArrSize; i ++) { cell[i] = (float*)malloc(size); cell2[i] = (float*)malloc(size); for (j = 0; j < secondArrSize; j ++) { cell[i][j] = 23.0; } } size_t pitch; float **d_cell; cudaMallocPitch((void**) &d_cell, &pitch, secondArrSize * sizeof(float), firstArrSize); cudaError_t tmp = cudaMemcpy2D(d_cell, pitch, cell, secondArrSize * sizeof(float), secondArrSize * sizeof(float), firstArrSize, cudaMemcpyHostToDevice); if (cudaSuccess != tmp) { printf("\n copy to GPU \n"); printf(cudaGetErrorString(tmp)); } dim3 dimBlock(8,8); dim3 dimGrid(1,1); compute_2d<<<dimGrid, dimBlock>>>( firstArrSize, secondArrSize, d_cell); if (cudaSuccess != tmp) { printf("\n compute \n"); printf(cudaGetErrorString(tmp)); } tmp = cudaMemcpy2D(cell2, secondArrSize * sizeof(float), d_cell, pitch, secondArrSize * sizeof(float), firstArrSize, cudaMemcpyDeviceToHost); if (cudaSuccess != tmp) { printf("\n copy to CPU \n"); printf(cudaGetErrorString(tmp)); } for (i = 0; i < firstArrSize; i++) { for (j = 0; j < secondArrSize; j ++) { printf("\n %lf ", cell2[i][j]); } } // for (i = 0; i < (2*(timeSteps - 1)) + secondArrSize; i ++) // { // //compute_2d <<< ceil((float) secondArrSize/THREADS_PER_BLOCK), THREADS_PER_BLOCK >>> (c_d, secondArrSize, i, timeSteps); // } cudaError_t err = cudaGetLastError(); if (err != cudaSuccess) printf ("CUDA error: %s\n", cudaGetErrorString(err)); }
.file "tmpxft_0010f075_00000000-6_compute_2D.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 _Z33__device_stub__Z10compute_2diiPPfiiPPf .type _Z33__device_stub__Z10compute_2diiPPfiiPPf, @function _Z33__device_stub__Z10compute_2diiPPfiiPPf: .LFB2082: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movq %rdx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10compute_2diiPPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z33__device_stub__Z10compute_2diiPPfiiPPf, .-_Z33__device_stub__Z10compute_2diiPPfiiPPf .globl _Z10compute_2diiPPf .type _Z10compute_2diiPPf, @function _Z10compute_2diiPPf: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z33__device_stub__Z10compute_2diiPPfiiPPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z10compute_2diiPPf, .-_Z10compute_2diiPPf .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "\n copy to GPU \n" .LC2: .string "\n compute \n" .LC3: .string "\n copy to CPU \n" .LC4: .string "\n %lf " .LC5: .string "CUDA error: %s\n" .text .globl compute2DArr .type compute2DArr, @function compute2DArr: .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 movl %edi, %ebp movl %edi, 28(%rsp) movl %esi, %r15d movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax movl %edi, %eax imull %esi, %eax leal 0(,%rax,4), %ecx movl %ecx, 24(%rsp) sall $3, %eax movslq %eax, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r13 movq %rax, 40(%rsp) movq %rbx, %rdi call malloc@PLT movq %rax, 32(%rsp) testl %ebp, %ebp jle .L12 movq %rax, %r14 movslq %ebp, %rax leaq 0(%r13,%rax,8), %rax movq %rax, 16(%rsp) movslq %r15d, %rbp salq $2, %rbp movl .LC0(%rip), %r12d .L15: movslq 24(%rsp), %rax movq %rax, 8(%rsp) movq %rax, %rdi call malloc@PLT movq %r13, %rbx movq %rax, 0(%r13) movq 8(%rsp), %rdi call malloc@PLT movq %rax, (%r14) testl %r15d, %r15d jle .L13 movl $0, %eax .L14: movq (%rbx), %rdx movl %r12d, (%rdx,%rax) addq $4, %rax cmpq %rbp, %rax jne .L14 .L13: addq $8, %r13 addq $8, %r14 movq 16(%rsp), %rax cmpq %rax, %r13 jne .L15 .L12: movslq 28(%rsp), %rbx movslq %r15d, %rbp salq $2, %rbp leaq 48(%rsp), %rsi leaq 56(%rsp), %rdi movq %rbx, %rcx movq %rbp, %rdx call cudaMallocPitch@PLT subq $8, %rsp .cfi_def_cfa_offset 168 pushq $1 .cfi_def_cfa_offset 176 movq %rbx, %r9 movq %rbp, %r8 movq %rbp, %rcx movq 56(%rsp), %rdx movq 64(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpy2D@PLT movl %eax, %r12d addq $16, %rsp .cfi_def_cfa_offset 160 testl %eax, %eax jne .L35 movl $8, 64(%rsp) movl $8, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movl $1, %ecx movq 76(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L36 .L24: subq $8, %rsp .cfi_def_cfa_offset 168 pushq $2 .cfi_def_cfa_offset 176 movq %rbx, %r9 movq %rbp, %r8 movq 64(%rsp), %rcx movq 72(%rsp), %rdx movq %rbp, %rsi movq 48(%rsp), %rdi call cudaMemcpy2D@PLT movl %eax, %r12d addq $16, %rsp .cfi_def_cfa_offset 160 testl %eax, %eax jne .L37 .L18: cmpl $0, 28(%rsp) jle .L19 movq 32(%rsp), %rax movq %rax, %r12 leaq (%rax,%rbx,8), %r14 leaq .LC4(%rip), %r13 jmp .L20 .L35: leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $8, 64(%rsp) movl $8, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movl $1, %ecx movq 76(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L17 .L26: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L24 .L37: leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L18 .L21: movq (%r12), %rax pxor %xmm0, %xmm0 cvtss2sd (%rax,%rbx), %xmm0 movq %r13, %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %rbx cmpq %rbx, %rbp jne .L21 .L22: addq $8, %r12 cmpq %r12, %r14 je .L19 .L20: movl $0, %ebx testl %r15d, %r15d jg .L21 jmp .L22 .L19: call cudaGetLastError@PLT testl %eax, %eax jne .L38 .L11: movq 88(%rsp), %rax subq %fs:40, %rax jne .L39 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 .L38: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L11 .L17: movq 56(%rsp), %rdx movl %r15d, %esi movl 28(%rsp), %edi call _Z33__device_stub__Z10compute_2diiPPfiiPPf jmp .L26 .L36: movq 56(%rsp), %rdx movl %r15d, %esi movl 28(%rsp), %edi call _Z33__device_stub__Z10compute_2diiPPfiiPPf jmp .L24 .L39: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size compute2DArr, .-compute2DArr .section .rodata.str1.1 .LC6: .string "_Z10compute_2diiPPf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z10compute_2diiPPf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1102577664 .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.
// compute.cu // // driver and kernel call #include <stdio.h> #define THREADS_PER_BLOCK 128 // __global__ void compute_2d (int secondArrSize, float *arr[]) __global__ void compute_2d ( int firstArrSize, int secondArrSize, float **arr) { int x = blockIdx.x * blockDim.x + threadIdx.x; int y = blockIdx.y * blockDim.y + threadIdx.y; if (x == 0 && x < firstArrSize && y == 0 && y < firstArrSize) { printf("Hello. I'm a thread %d in block %d \n", threadIdx.x, blockIdx.x); // printf("%lf \n", arr[x][y]); } // if (x <= arrSize) { // if (x % 2 == timeStep % 2 && x <= timeStep) // { // if (timeStep > timeSteps && x <= (timeStep - timeSteps - arrSize)) // { // } else // { // if (x == 0) // { // c_d[x] = (100.0 + c_d[x + 1]) / 2.0; // } else if (x == arrSize - 1) // { // c_d[x] = (c_d[x - 1] + c_d[x]) / 2.0; // } else // { // c_d[x] = (c_d[x - 1] + c_d[x + 1]) / 2.0; // } // } // } __syncthreads(); // } } extern "C" void compute2DArr (int firstArrSize, int secondArrSize, float *metalRod, int timeSteps) { int i = 0, j = 0; int size=firstArrSize*secondArrSize*sizeof(float); //allocate resources float **cell=(float**)malloc(size * 2); float **cell2=(float**)malloc(size * 2); for (i = 0; i < firstArrSize; i ++) { cell[i] = (float*)malloc(size); cell2[i] = (float*)malloc(size); for (j = 0; j < secondArrSize; j ++) { cell[i][j] = 23.0; } } size_t pitch; float **d_cell; cudaMallocPitch((void**) &d_cell, &pitch, secondArrSize * sizeof(float), firstArrSize); cudaError_t tmp = cudaMemcpy2D(d_cell, pitch, cell, secondArrSize * sizeof(float), secondArrSize * sizeof(float), firstArrSize, cudaMemcpyHostToDevice); if (cudaSuccess != tmp) { printf("\n copy to GPU \n"); printf(cudaGetErrorString(tmp)); } dim3 dimBlock(8,8); dim3 dimGrid(1,1); compute_2d<<<dimGrid, dimBlock>>>( firstArrSize, secondArrSize, d_cell); if (cudaSuccess != tmp) { printf("\n compute \n"); printf(cudaGetErrorString(tmp)); } tmp = cudaMemcpy2D(cell2, secondArrSize * sizeof(float), d_cell, pitch, secondArrSize * sizeof(float), firstArrSize, cudaMemcpyDeviceToHost); if (cudaSuccess != tmp) { printf("\n copy to CPU \n"); printf(cudaGetErrorString(tmp)); } for (i = 0; i < firstArrSize; i++) { for (j = 0; j < secondArrSize; j ++) { printf("\n %lf ", cell2[i][j]); } } // for (i = 0; i < (2*(timeSteps - 1)) + secondArrSize; i ++) // { // //compute_2d <<< ceil((float) secondArrSize/THREADS_PER_BLOCK), THREADS_PER_BLOCK >>> (c_d, secondArrSize, i, timeSteps); // } cudaError_t err = cudaGetLastError(); if (err != cudaSuccess) printf ("CUDA error: %s\n", cudaGetErrorString(err)); }
// compute.cu // // driver and kernel call #include <hip/hip_runtime.h> #include <stdio.h> #define THREADS_PER_BLOCK 128 // __global__ void compute_2d (int secondArrSize, float *arr[]) __global__ void compute_2d ( int firstArrSize, int secondArrSize, float **arr) { int x = blockIdx.x * blockDim.x + threadIdx.x; int y = blockIdx.y * blockDim.y + threadIdx.y; if (x == 0 && x < firstArrSize && y == 0 && y < firstArrSize) { printf("Hello. I'm a thread %d in block %d \n", threadIdx.x, blockIdx.x); // printf("%lf \n", arr[x][y]); } // if (x <= arrSize) { // if (x % 2 == timeStep % 2 && x <= timeStep) // { // if (timeStep > timeSteps && x <= (timeStep - timeSteps - arrSize)) // { // } else // { // if (x == 0) // { // c_d[x] = (100.0 + c_d[x + 1]) / 2.0; // } else if (x == arrSize - 1) // { // c_d[x] = (c_d[x - 1] + c_d[x]) / 2.0; // } else // { // c_d[x] = (c_d[x - 1] + c_d[x + 1]) / 2.0; // } // } // } __syncthreads(); // } } extern "C" void compute2DArr (int firstArrSize, int secondArrSize, float *metalRod, int timeSteps) { int i = 0, j = 0; int size=firstArrSize*secondArrSize*sizeof(float); //allocate resources float **cell=(float**)malloc(size * 2); float **cell2=(float**)malloc(size * 2); for (i = 0; i < firstArrSize; i ++) { cell[i] = (float*)malloc(size); cell2[i] = (float*)malloc(size); for (j = 0; j < secondArrSize; j ++) { cell[i][j] = 23.0; } } size_t pitch; float **d_cell; hipMallocPitch((void**) &d_cell, &pitch, secondArrSize * sizeof(float), firstArrSize); hipError_t tmp = hipMemcpy2D(d_cell, pitch, cell, secondArrSize * sizeof(float), secondArrSize * sizeof(float), firstArrSize, hipMemcpyHostToDevice); if (hipSuccess != tmp) { printf("\n copy to GPU \n"); printf(hipGetErrorString(tmp)); } dim3 dimBlock(8,8); dim3 dimGrid(1,1); compute_2d<<<dimGrid, dimBlock>>>( firstArrSize, secondArrSize, d_cell); if (hipSuccess != tmp) { printf("\n compute \n"); printf(hipGetErrorString(tmp)); } tmp = hipMemcpy2D(cell2, secondArrSize * sizeof(float), d_cell, pitch, secondArrSize * sizeof(float), firstArrSize, hipMemcpyDeviceToHost); if (hipSuccess != tmp) { printf("\n copy to CPU \n"); printf(hipGetErrorString(tmp)); } for (i = 0; i < firstArrSize; i++) { for (j = 0; j < secondArrSize; j ++) { printf("\n %lf ", cell2[i][j]); } } // for (i = 0; i < (2*(timeSteps - 1)) + secondArrSize; i ++) // { // //compute_2d <<< ceil((float) secondArrSize/THREADS_PER_BLOCK), THREADS_PER_BLOCK >>> (c_d, secondArrSize, i, timeSteps); // } hipError_t err = hipGetLastError(); if (err != hipSuccess) printf ("CUDA error: %s\n", hipGetErrorString(err)); }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
// compute.cu // // driver and kernel call #include <hip/hip_runtime.h> #include <stdio.h> #define THREADS_PER_BLOCK 128 // __global__ void compute_2d (int secondArrSize, float *arr[]) __global__ void compute_2d ( int firstArrSize, int secondArrSize, float **arr) { int x = blockIdx.x * blockDim.x + threadIdx.x; int y = blockIdx.y * blockDim.y + threadIdx.y; if (x == 0 && x < firstArrSize && y == 0 && y < firstArrSize) { printf("Hello. I'm a thread %d in block %d \n", threadIdx.x, blockIdx.x); // printf("%lf \n", arr[x][y]); } // if (x <= arrSize) { // if (x % 2 == timeStep % 2 && x <= timeStep) // { // if (timeStep > timeSteps && x <= (timeStep - timeSteps - arrSize)) // { // } else // { // if (x == 0) // { // c_d[x] = (100.0 + c_d[x + 1]) / 2.0; // } else if (x == arrSize - 1) // { // c_d[x] = (c_d[x - 1] + c_d[x]) / 2.0; // } else // { // c_d[x] = (c_d[x - 1] + c_d[x + 1]) / 2.0; // } // } // } __syncthreads(); // } } extern "C" void compute2DArr (int firstArrSize, int secondArrSize, float *metalRod, int timeSteps) { int i = 0, j = 0; int size=firstArrSize*secondArrSize*sizeof(float); //allocate resources float **cell=(float**)malloc(size * 2); float **cell2=(float**)malloc(size * 2); for (i = 0; i < firstArrSize; i ++) { cell[i] = (float*)malloc(size); cell2[i] = (float*)malloc(size); for (j = 0; j < secondArrSize; j ++) { cell[i][j] = 23.0; } } size_t pitch; float **d_cell; hipMallocPitch((void**) &d_cell, &pitch, secondArrSize * sizeof(float), firstArrSize); hipError_t tmp = hipMemcpy2D(d_cell, pitch, cell, secondArrSize * sizeof(float), secondArrSize * sizeof(float), firstArrSize, hipMemcpyHostToDevice); if (hipSuccess != tmp) { printf("\n copy to GPU \n"); printf(hipGetErrorString(tmp)); } dim3 dimBlock(8,8); dim3 dimGrid(1,1); compute_2d<<<dimGrid, dimBlock>>>( firstArrSize, secondArrSize, d_cell); if (hipSuccess != tmp) { printf("\n compute \n"); printf(hipGetErrorString(tmp)); } tmp = hipMemcpy2D(cell2, secondArrSize * sizeof(float), d_cell, pitch, secondArrSize * sizeof(float), firstArrSize, hipMemcpyDeviceToHost); if (hipSuccess != tmp) { printf("\n copy to CPU \n"); printf(hipGetErrorString(tmp)); } for (i = 0; i < firstArrSize; i++) { for (j = 0; j < secondArrSize; j ++) { printf("\n %lf ", cell2[i][j]); } } // for (i = 0; i < (2*(timeSteps - 1)) + secondArrSize; i ++) // { // //compute_2d <<< ceil((float) secondArrSize/THREADS_PER_BLOCK), THREADS_PER_BLOCK >>> (c_d, secondArrSize, i, timeSteps); // } hipError_t err = hipGetLastError(); if (err != hipSuccess) printf ("CUDA error: %s\n", hipGetErrorString(err)); }
.text .file "compute_2D.hip" .globl _Z25__device_stub__compute_2diiPPf # -- Begin function _Z25__device_stub__compute_2diiPPf .p2align 4, 0x90 .type _Z25__device_stub__compute_2diiPPf,@function _Z25__device_stub__compute_2diiPPf: # @_Z25__device_stub__compute_2diiPPf .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movl %edi, 4(%rsp) movl %esi, (%rsp) movq %rdx, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 64(%rsp) movq %rsp, %rax movq %rax, 72(%rsp) leaq 56(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10compute_2diiPPf, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z25__device_stub__compute_2diiPPf, .Lfunc_end0-_Z25__device_stub__compute_2diiPPf .cfi_endproc # -- End function .globl compute2DArr # -- Begin function compute2DArr .p2align 4, 0x90 .type compute2DArr,@function compute2DArr: # @compute2DArr .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 %esi, %ebp movl %edi, %r15d movl %esi, %ebx imull %edi, %ebx leal (,%rbx,8), %eax movslq %eax, %r14 movq %r14, %rdi callq malloc movq %rax, %r13 movq %r14, %rdi callq malloc movq %rax, %r14 movl %r15d, %eax movq %rax, 56(%rsp) # 8-byte Spill movl %ebp, 28(%rsp) # 4-byte Spill movl %ebp, %ebp movl %r15d, 44(%rsp) # 4-byte Spill testl %r15d, %r15d jle .LBB1_6 # %bb.1: # %.lr.ph70 shll $2, %ebx movslq %ebx, %r15 xorl %ebx, %ebx jmp .LBB1_2 .p2align 4, 0x90 .LBB1_5: # %._crit_edge # in Loop: Header=BB1_2 Depth=1 incq %rbx cmpq 56(%rsp), %rbx # 8-byte Folded Reload je .LBB1_6 .LBB1_2: # =>This Loop Header: Depth=1 # Child Loop BB1_4 Depth 2 movq %r15, %rdi callq malloc movq %rax, %r12 movq %rax, (%r13,%rbx,8) movq %r15, %rdi callq malloc movq %rax, (%r14,%rbx,8) cmpl $0, 28(%rsp) # 4-byte Folded Reload jle .LBB1_5 # %bb.3: # %.lr.ph.preheader # in Loop: Header=BB1_2 Depth=1 xorl %eax, %eax .p2align 4, 0x90 .LBB1_4: # %.lr.ph # Parent Loop BB1_2 Depth=1 # => This Inner Loop Header: Depth=2 movl $1102577664, (%r12,%rax,4) # imm = 0x41B80000 incq %rax cmpq %rax, %rbp jne .LBB1_4 jmp .LBB1_5 .LBB1_6: # %._crit_edge71 movslq 28(%rsp), %r15 # 4-byte Folded Reload shlq $2, %r15 movl 44(%rsp), %ebx # 4-byte Reload movslq %ebx, %r12 leaq 32(%rsp), %rdi leaq 64(%rsp), %rsi movq %r15, %rdx movq %r12, %rcx callq hipMallocPitch movq 32(%rsp), %rdi movq 64(%rsp), %rsi movl $1, (%rsp) movq %r13, %rdx movq %r15, %rcx movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D movl %eax, %r13d testl %eax, %eax je .LBB1_8 # %bb.7: movl $.Lstr, %edi callq puts@PLT movl %r13d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_8: movabsq $4294967297, %rdi # imm = 0x100000001 movabsq $34359738376, %rdx # imm = 0x800000008 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_10 # %bb.9: movq 32(%rsp), %rax movl %ebx, 52(%rsp) movl 28(%rsp), %ecx # 4-byte Reload movl %ecx, 48(%rsp) movq %rax, 120(%rsp) leaq 52(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 120(%rsp), %rax movq %rax, 144(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 80(%rsp), %rax movq 72(%rsp), %rdi movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d movq %rdi, 8(%rsp) movq %rax, (%rsp) leaq 128(%rsp), %r9 movl $_Z10compute_2diiPPf, %edi callq hipLaunchKernel .LBB1_10: testl %r13d, %r13d je .LBB1_12 # %bb.11: movl $.Lstr.1, %edi callq puts@PLT movl %r13d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_12: movq 32(%rsp), %rdx movq 64(%rsp), %rcx movl $2, (%rsp) movq %r14, %rdi movq %r15, %rsi movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D testl %eax, %eax je .LBB1_14 # %bb.13: movl %eax, %r15d movl $.Lstr.2, %edi callq puts@PLT movl %r15d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_14: testl %ebx, %ebx jle .LBB1_20 # %bb.15: # %.preheader.lr.ph xorl %ebx, %ebx jmp .LBB1_16 .p2align 4, 0x90 .LBB1_19: # %._crit_edge74 # in Loop: Header=BB1_16 Depth=1 incq %rbx cmpq 56(%rsp), %rbx # 8-byte Folded Reload je .LBB1_20 .LBB1_16: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_18 Depth 2 cmpl $0, 28(%rsp) # 4-byte Folded Reload jle .LBB1_19 # %bb.17: # %.lr.ph73 # in Loop: Header=BB1_16 Depth=1 xorl %r15d, %r15d .p2align 4, 0x90 .LBB1_18: # Parent Loop BB1_16 Depth=1 # => This Inner Loop Header: Depth=2 movq (%r14,%rbx,8), %rax movss (%rax,%r15,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str.3, %edi movb $1, %al callq printf incq %r15 cmpq %r15, %rbp jne .LBB1_18 jmp .LBB1_19 .LBB1_20: # %._crit_edge76 callq hipGetLastError testl %eax, %eax je .LBB1_22 # %bb.21: movl %eax, %edi callq hipGetErrorString movl $.L.str.4, %edi movq %rax, %rsi xorl %eax, %eax callq printf .LBB1_22: 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 compute2DArr, .Lfunc_end1-compute2DArr .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 $_Z10compute_2diiPPf, %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 _Z10compute_2diiPPf,@object # @_Z10compute_2diiPPf .section .rodata,"a",@progbits .globl _Z10compute_2diiPPf .p2align 3, 0x0 _Z10compute_2diiPPf: .quad _Z25__device_stub__compute_2diiPPf .size _Z10compute_2diiPPf, 8 .type .L.str.3,@object # @.str.3 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.3: .asciz "\n %lf " .size .L.str.3, 7 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "CUDA error: %s\n" .size .L.str.4, 16 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10compute_2diiPPf" .size .L__unnamed_1, 20 .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 "\n copy to GPU " .size .Lstr, 15 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "\n compute " .size .Lstr.1, 11 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "\n copy to CPU " .size .Lstr.2, 15 .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__compute_2diiPPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10compute_2diiPPf .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_0010f075_00000000-6_compute_2D.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 _Z33__device_stub__Z10compute_2diiPPfiiPPf .type _Z33__device_stub__Z10compute_2diiPPfiiPPf, @function _Z33__device_stub__Z10compute_2diiPPfiiPPf: .LFB2082: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movl %esi, 8(%rsp) movq %rdx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10compute_2diiPPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z33__device_stub__Z10compute_2diiPPfiiPPf, .-_Z33__device_stub__Z10compute_2diiPPfiiPPf .globl _Z10compute_2diiPPf .type _Z10compute_2diiPPf, @function _Z10compute_2diiPPf: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z33__device_stub__Z10compute_2diiPPfiiPPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z10compute_2diiPPf, .-_Z10compute_2diiPPf .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "\n copy to GPU \n" .LC2: .string "\n compute \n" .LC3: .string "\n copy to CPU \n" .LC4: .string "\n %lf " .LC5: .string "CUDA error: %s\n" .text .globl compute2DArr .type compute2DArr, @function compute2DArr: .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 movl %edi, %ebp movl %edi, 28(%rsp) movl %esi, %r15d movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax movl %edi, %eax imull %esi, %eax leal 0(,%rax,4), %ecx movl %ecx, 24(%rsp) sall $3, %eax movslq %eax, %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r13 movq %rax, 40(%rsp) movq %rbx, %rdi call malloc@PLT movq %rax, 32(%rsp) testl %ebp, %ebp jle .L12 movq %rax, %r14 movslq %ebp, %rax leaq 0(%r13,%rax,8), %rax movq %rax, 16(%rsp) movslq %r15d, %rbp salq $2, %rbp movl .LC0(%rip), %r12d .L15: movslq 24(%rsp), %rax movq %rax, 8(%rsp) movq %rax, %rdi call malloc@PLT movq %r13, %rbx movq %rax, 0(%r13) movq 8(%rsp), %rdi call malloc@PLT movq %rax, (%r14) testl %r15d, %r15d jle .L13 movl $0, %eax .L14: movq (%rbx), %rdx movl %r12d, (%rdx,%rax) addq $4, %rax cmpq %rbp, %rax jne .L14 .L13: addq $8, %r13 addq $8, %r14 movq 16(%rsp), %rax cmpq %rax, %r13 jne .L15 .L12: movslq 28(%rsp), %rbx movslq %r15d, %rbp salq $2, %rbp leaq 48(%rsp), %rsi leaq 56(%rsp), %rdi movq %rbx, %rcx movq %rbp, %rdx call cudaMallocPitch@PLT subq $8, %rsp .cfi_def_cfa_offset 168 pushq $1 .cfi_def_cfa_offset 176 movq %rbx, %r9 movq %rbp, %r8 movq %rbp, %rcx movq 56(%rsp), %rdx movq 64(%rsp), %rsi movq 72(%rsp), %rdi call cudaMemcpy2D@PLT movl %eax, %r12d addq $16, %rsp .cfi_def_cfa_offset 160 testl %eax, %eax jne .L35 movl $8, 64(%rsp) movl $8, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movl $1, %ecx movq 76(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L36 .L24: subq $8, %rsp .cfi_def_cfa_offset 168 pushq $2 .cfi_def_cfa_offset 176 movq %rbx, %r9 movq %rbp, %r8 movq 64(%rsp), %rcx movq 72(%rsp), %rdx movq %rbp, %rsi movq 48(%rsp), %rdi call cudaMemcpy2D@PLT movl %eax, %r12d addq $16, %rsp .cfi_def_cfa_offset 160 testl %eax, %eax jne .L37 .L18: cmpl $0, 28(%rsp) jle .L19 movq 32(%rsp), %rax movq %rax, %r12 leaq (%rax,%rbx,8), %r14 leaq .LC4(%rip), %r13 jmp .L20 .L35: leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $8, 64(%rsp) movl $8, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movl $1, %ecx movq 76(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L17 .L26: leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L24 .L37: leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %r12d, %edi call cudaGetErrorString@PLT movq %rax, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L18 .L21: movq (%r12), %rax pxor %xmm0, %xmm0 cvtss2sd (%rax,%rbx), %xmm0 movq %r13, %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %rbx cmpq %rbx, %rbp jne .L21 .L22: addq $8, %r12 cmpq %r12, %r14 je .L19 .L20: movl $0, %ebx testl %r15d, %r15d jg .L21 jmp .L22 .L19: call cudaGetLastError@PLT testl %eax, %eax jne .L38 .L11: movq 88(%rsp), %rax subq %fs:40, %rax jne .L39 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 .L38: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L11 .L17: movq 56(%rsp), %rdx movl %r15d, %esi movl 28(%rsp), %edi call _Z33__device_stub__Z10compute_2diiPPfiiPPf jmp .L26 .L36: movq 56(%rsp), %rdx movl %r15d, %esi movl 28(%rsp), %edi call _Z33__device_stub__Z10compute_2diiPPfiiPPf jmp .L24 .L39: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size compute2DArr, .-compute2DArr .section .rodata.str1.1 .LC6: .string "_Z10compute_2diiPPf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z10compute_2diiPPf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1102577664 .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 "compute_2D.hip" .globl _Z25__device_stub__compute_2diiPPf # -- Begin function _Z25__device_stub__compute_2diiPPf .p2align 4, 0x90 .type _Z25__device_stub__compute_2diiPPf,@function _Z25__device_stub__compute_2diiPPf: # @_Z25__device_stub__compute_2diiPPf .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movl %edi, 4(%rsp) movl %esi, (%rsp) movq %rdx, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 64(%rsp) movq %rsp, %rax movq %rax, 72(%rsp) leaq 56(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10compute_2diiPPf, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z25__device_stub__compute_2diiPPf, .Lfunc_end0-_Z25__device_stub__compute_2diiPPf .cfi_endproc # -- End function .globl compute2DArr # -- Begin function compute2DArr .p2align 4, 0x90 .type compute2DArr,@function compute2DArr: # @compute2DArr .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 %esi, %ebp movl %edi, %r15d movl %esi, %ebx imull %edi, %ebx leal (,%rbx,8), %eax movslq %eax, %r14 movq %r14, %rdi callq malloc movq %rax, %r13 movq %r14, %rdi callq malloc movq %rax, %r14 movl %r15d, %eax movq %rax, 56(%rsp) # 8-byte Spill movl %ebp, 28(%rsp) # 4-byte Spill movl %ebp, %ebp movl %r15d, 44(%rsp) # 4-byte Spill testl %r15d, %r15d jle .LBB1_6 # %bb.1: # %.lr.ph70 shll $2, %ebx movslq %ebx, %r15 xorl %ebx, %ebx jmp .LBB1_2 .p2align 4, 0x90 .LBB1_5: # %._crit_edge # in Loop: Header=BB1_2 Depth=1 incq %rbx cmpq 56(%rsp), %rbx # 8-byte Folded Reload je .LBB1_6 .LBB1_2: # =>This Loop Header: Depth=1 # Child Loop BB1_4 Depth 2 movq %r15, %rdi callq malloc movq %rax, %r12 movq %rax, (%r13,%rbx,8) movq %r15, %rdi callq malloc movq %rax, (%r14,%rbx,8) cmpl $0, 28(%rsp) # 4-byte Folded Reload jle .LBB1_5 # %bb.3: # %.lr.ph.preheader # in Loop: Header=BB1_2 Depth=1 xorl %eax, %eax .p2align 4, 0x90 .LBB1_4: # %.lr.ph # Parent Loop BB1_2 Depth=1 # => This Inner Loop Header: Depth=2 movl $1102577664, (%r12,%rax,4) # imm = 0x41B80000 incq %rax cmpq %rax, %rbp jne .LBB1_4 jmp .LBB1_5 .LBB1_6: # %._crit_edge71 movslq 28(%rsp), %r15 # 4-byte Folded Reload shlq $2, %r15 movl 44(%rsp), %ebx # 4-byte Reload movslq %ebx, %r12 leaq 32(%rsp), %rdi leaq 64(%rsp), %rsi movq %r15, %rdx movq %r12, %rcx callq hipMallocPitch movq 32(%rsp), %rdi movq 64(%rsp), %rsi movl $1, (%rsp) movq %r13, %rdx movq %r15, %rcx movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D movl %eax, %r13d testl %eax, %eax je .LBB1_8 # %bb.7: movl $.Lstr, %edi callq puts@PLT movl %r13d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_8: movabsq $4294967297, %rdi # imm = 0x100000001 movabsq $34359738376, %rdx # imm = 0x800000008 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_10 # %bb.9: movq 32(%rsp), %rax movl %ebx, 52(%rsp) movl 28(%rsp), %ecx # 4-byte Reload movl %ecx, 48(%rsp) movq %rax, 120(%rsp) leaq 52(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 120(%rsp), %rax movq %rax, 144(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 80(%rsp), %rax movq 72(%rsp), %rdi movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d movq %rdi, 8(%rsp) movq %rax, (%rsp) leaq 128(%rsp), %r9 movl $_Z10compute_2diiPPf, %edi callq hipLaunchKernel .LBB1_10: testl %r13d, %r13d je .LBB1_12 # %bb.11: movl $.Lstr.1, %edi callq puts@PLT movl %r13d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_12: movq 32(%rsp), %rdx movq 64(%rsp), %rcx movl $2, (%rsp) movq %r14, %rdi movq %r15, %rsi movq %r15, %r8 movq %r12, %r9 callq hipMemcpy2D testl %eax, %eax je .LBB1_14 # %bb.13: movl %eax, %r15d movl $.Lstr.2, %edi callq puts@PLT movl %r15d, %edi callq hipGetErrorString movq %rax, %rdi xorl %eax, %eax callq printf .LBB1_14: testl %ebx, %ebx jle .LBB1_20 # %bb.15: # %.preheader.lr.ph xorl %ebx, %ebx jmp .LBB1_16 .p2align 4, 0x90 .LBB1_19: # %._crit_edge74 # in Loop: Header=BB1_16 Depth=1 incq %rbx cmpq 56(%rsp), %rbx # 8-byte Folded Reload je .LBB1_20 .LBB1_16: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_18 Depth 2 cmpl $0, 28(%rsp) # 4-byte Folded Reload jle .LBB1_19 # %bb.17: # %.lr.ph73 # in Loop: Header=BB1_16 Depth=1 xorl %r15d, %r15d .p2align 4, 0x90 .LBB1_18: # Parent Loop BB1_16 Depth=1 # => This Inner Loop Header: Depth=2 movq (%r14,%rbx,8), %rax movss (%rax,%r15,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str.3, %edi movb $1, %al callq printf incq %r15 cmpq %r15, %rbp jne .LBB1_18 jmp .LBB1_19 .LBB1_20: # %._crit_edge76 callq hipGetLastError testl %eax, %eax je .LBB1_22 # %bb.21: movl %eax, %edi callq hipGetErrorString movl $.L.str.4, %edi movq %rax, %rsi xorl %eax, %eax callq printf .LBB1_22: 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 compute2DArr, .Lfunc_end1-compute2DArr .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 $_Z10compute_2diiPPf, %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 _Z10compute_2diiPPf,@object # @_Z10compute_2diiPPf .section .rodata,"a",@progbits .globl _Z10compute_2diiPPf .p2align 3, 0x0 _Z10compute_2diiPPf: .quad _Z25__device_stub__compute_2diiPPf .size _Z10compute_2diiPPf, 8 .type .L.str.3,@object # @.str.3 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.3: .asciz "\n %lf " .size .L.str.3, 7 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "CUDA error: %s\n" .size .L.str.4, 16 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10compute_2diiPPf" .size .L__unnamed_1, 20 .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 "\n copy to GPU " .size .Lstr, 15 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "\n compute " .size .Lstr.1, 11 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "\n copy to CPU " .size .Lstr.2, 15 .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__compute_2diiPPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10compute_2diiPPf .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 "includes.h" __device__ void softmax_device(float *input, int n, float temp, int stride, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i*stride]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = expf(input[i*stride]/temp - largest/temp); sum += e; output[i*stride] = e; } for(i = 0; i < n; ++i){ output[i*stride] /= sum; } } __device__ void softmax_device(int n, float *input, float temp, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = exp(input[i]/temp - largest/temp); sum += e; output[i] = e; } for(i = 0; i < n; ++i){ output[i] /= sum; } } __global__ void softmax_kernel(float *input, int n, int batch, int batch_offset, int groups, int group_offset, int stride, float temp, float *output) { int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x; if (id >= batch*groups) return; int b = id / groups; int g = id % groups; softmax_device(input + b*batch_offset + g*group_offset, n, temp, stride, output + b*batch_offset + g*group_offset); }
.file "tmpxft_0009d388_00000000-6_softmax_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2031: .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 .LFE2031: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z14softmax_devicePfifiS_ .type _Z14softmax_devicePfifiS_, @function _Z14softmax_devicePfifiS_: .LFB2027: .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 .LFE2027: .size _Z14softmax_devicePfifiS_, .-_Z14softmax_devicePfifiS_ .globl _Z14softmax_deviceiPffS_ .type _Z14softmax_deviceiPffS_, @function _Z14softmax_deviceiPffS_: .LFB2028: .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 .LFE2028: .size _Z14softmax_deviceiPffS_, .-_Z14softmax_deviceiPffS_ .globl _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ .type _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_, @function _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_: .LFB2053: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movl %r9d, 20(%rsp) movss %xmm0, 16(%rsp) movq 216(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 28(%rsp), %rax movq %rax, 136(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) leaq 20(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 184(%rsp), %rax subq %fs:40, %rax jne .L12 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14softmax_kernelPfiiiiiifS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_, .-_Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ .globl _Z14softmax_kernelPfiiiiiifS_ .type _Z14softmax_kernelPfiiiiiifS_, @function _Z14softmax_kernelPfiiiiiifS_: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _Z14softmax_kernelPfiiiiiifS_, .-_Z14softmax_kernelPfiiiiiifS_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14softmax_kernelPfiiiiiifS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2056: .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 _Z14softmax_kernelPfiiiiiifS_(%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 .LFE2056: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __device__ void softmax_device(float *input, int n, float temp, int stride, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i*stride]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = expf(input[i*stride]/temp - largest/temp); sum += e; output[i*stride] = e; } for(i = 0; i < n; ++i){ output[i*stride] /= sum; } } __device__ void softmax_device(int n, float *input, float temp, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = exp(input[i]/temp - largest/temp); sum += e; output[i] = e; } for(i = 0; i < n; ++i){ output[i] /= sum; } } __global__ void softmax_kernel(float *input, int n, int batch, int batch_offset, int groups, int group_offset, int stride, float temp, float *output) { int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x; if (id >= batch*groups) return; int b = id / groups; int g = id % groups; softmax_device(input + b*batch_offset + g*group_offset, n, temp, stride, output + b*batch_offset + g*group_offset); }
#include <hip/hip_runtime.h> #include "includes.h" __device__ void softmax_device(float *input, int n, float temp, int stride, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i*stride]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = expf(input[i*stride]/temp - largest/temp); sum += e; output[i*stride] = e; } for(i = 0; i < n; ++i){ output[i*stride] /= sum; } } __device__ void softmax_device(int n, float *input, float temp, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = exp(input[i]/temp - largest/temp); sum += e; output[i] = e; } for(i = 0; i < n; ++i){ output[i] /= sum; } } __global__ void softmax_kernel(float *input, int n, int batch, int batch_offset, int groups, int group_offset, int stride, float temp, float *output) { int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x; if (id >= batch*groups) return; int b = id / groups; int g = id % groups; softmax_device(input + b*batch_offset + g*group_offset, n, temp, stride, output + b*batch_offset + g*group_offset); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __device__ void softmax_device(float *input, int n, float temp, int stride, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i*stride]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = expf(input[i*stride]/temp - largest/temp); sum += e; output[i*stride] = e; } for(i = 0; i < n; ++i){ output[i*stride] /= sum; } } __device__ void softmax_device(int n, float *input, float temp, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = exp(input[i]/temp - largest/temp); sum += e; output[i] = e; } for(i = 0; i < n; ++i){ output[i] /= sum; } } __global__ void softmax_kernel(float *input, int n, int batch, int batch_offset, int groups, int group_offset, int stride, float temp, float *output) { int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x; if (id >= batch*groups) return; int b = id / groups; int g = id % groups; softmax_device(input + b*batch_offset + g*group_offset, n, temp, stride, output + b*batch_offset + g*group_offset); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14softmax_kernelPfiiiiiifS_ .globl _Z14softmax_kernelPfiiiiiifS_ .p2align 8 .type _Z14softmax_kernelPfiiiiiifS_,@function _Z14softmax_kernelPfiiiiiifS_: s_clause 0x3 s_load_b32 s2, s[0:1], 0x30 s_load_b32 s3, s[0:1], 0x3c s_load_b32 s5, s[0:1], 0xc s_load_b32 s4, s[0:1], 0x14 s_waitcnt lgkmcnt(0) s_mul_i32 s2, s2, s15 s_and_b32 s3, s3, 0xffff s_add_i32 s2, s2, s14 s_delay_alu instid0(SALU_CYCLE_1) v_mad_u64_u32 v[1:2], null, s2, s3, v[0:1] s_mul_i32 s2, s4, s5 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_cmp_gt_i32_e32 vcc_lo, s2, v1 s_and_saveexec_b32 s2, vcc_lo s_cbranch_execz .LBB0_12 s_ashr_i32 s2, s4, 31 v_ashrrev_i32_e32 v3, 31, v1 s_add_i32 s3, s4, s2 s_load_b32 s10, s[0:1], 0x8 s_xor_b32 s3, s3, s2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, v1, v3 v_cvt_f32_u32_e32 v0, s3 s_sub_i32 s5, 0, s3 v_xor_b32_e32 v4, v4, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_rcp_iflag_f32_e32 v0, v0 v_xor_b32_e32 v3, s2, v3 s_waitcnt_depctr 0xfff v_mul_f32_e32 v0, 0x4f7ffffe, v0 v_cvt_u32_f32_e32 v0, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_lo_u32 v2, s5, v0 v_mul_hi_u32 v2, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v0, v0, v2 v_mul_hi_u32 v0, v4, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_lo_u32 v2, v0, s3 v_sub_nc_u32_e32 v2, v4, v2 v_add_nc_u32_e32 v4, 1, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_subrev_nc_u32_e32 v5, s3, v2 v_cmp_le_u32_e32 vcc_lo, s3, v2 v_cndmask_b32_e32 v0, v0, v4, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v2, v2, v5, vcc_lo v_add_nc_u32_e32 v4, 1, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_cmp_le_u32_e32 vcc_lo, s3, v2 s_clause 0x1 s_load_b32 s6, s[0:1], 0x10 s_load_b64 s[2:3], s[0:1], 0x18 v_cndmask_b32_e32 v0, v0, v4, vcc_lo v_xor_b32_e32 v0, v0, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_sub_nc_u32_e32 v0, v0, v3 v_mul_lo_u32 v2, v0, s4 s_load_b64 s[4:5], s[0:1], 0x0 s_waitcnt lgkmcnt(0) v_mul_lo_u32 v0, v0, s6 s_cmp_gt_i32 s10, 0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_sub_nc_u32_e32 v1, v1, v2 v_mul_lo_u32 v2, v1, s2 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v1, 31, v0 s_cselect_b32 s2, -1, 0 s_cmp_lt_i32 s10, 1 v_ashrrev_i32_e32 v3, 31, v2 s_cbranch_scc1 .LBB0_4 v_lshlrev_b64 v[4:5], 2, v[0:1] s_delay_alu instid0(VALU_DEP_2) v_lshlrev_b64 v[6:7], 2, v[2:3] s_ashr_i32 s7, s3, 31 s_mov_b32 s6, s3 s_mov_b32 s8, s10 s_lshl_b64 s[6:7], s[6:7], 2 v_add_co_u32 v4, vcc_lo, s4, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v4, vcc_lo, v4, v6 v_add_co_ci_u32_e32 v5, vcc_lo, v5, v7, vcc_lo v_mov_b32_e32 v6, 0xff800000 .LBB0_3: global_load_b32 v7, v[4:5], off v_add_co_u32 v4, vcc_lo, v4, s6 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v5, vcc_lo s_add_i32 s8, s8, -1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) s_cmp_eq_u32 s8, 0 s_waitcnt vmcnt(0) v_cvt_i32_f32_e32 v7, v7 v_cvt_f32_i32_e32 v7, v7 s_delay_alu instid0(VALU_DEP_1) v_cmp_lt_f32_e32 vcc_lo, v6, v7 v_cndmask_b32_e32 v6, v6, v7, vcc_lo s_cbranch_scc0 .LBB0_3 s_branch .LBB0_5 .LBB0_4: v_mov_b32_e32 v6, 0xff800000 .LBB0_5: s_load_b64 s[6:7], s[0:1], 0x28 v_cndmask_b32_e64 v5, 0, 1, s2 s_and_not1_b32 vcc_lo, exec_lo, s2 s_cbranch_vccnz .LBB0_8 s_load_b32 s1, s[0:1], 0x20 v_lshlrev_b64 v[9:10], 2, v[0:1] s_ashr_i32 s9, s3, 31 s_mov_b32 s8, s3 s_mov_b32 s2, s10 s_lshl_b64 s[8:9], s[8:9], 2 s_waitcnt lgkmcnt(0) v_div_scale_f32 v4, null, s1, s1, v6 v_div_scale_f32 v12, vcc_lo, v6, s1, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f32_e32 v11, v4 s_waitcnt_depctr 0xfff v_fma_f32 v7, -v4, v11, 1.0 v_fmac_f32_e32 v11, v7, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f32_e32 v13, v12, v11 v_fma_f32 v7, -v4, v13, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_fmac_f32_e32 v13, v7, v11 v_lshlrev_b64 v[7:8], 2, v[2:3] v_fma_f32 v4, -v4, v13, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_div_fmas_f32 v4, v4, v11, v13 v_add_co_u32 v7, vcc_lo, v7, v9 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_ci_u32_e32 v8, vcc_lo, v8, v10, vcc_lo v_div_fixup_f32 v6, v4, s1, v6 v_mov_b32_e32 v4, 0 .LBB0_7: s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_u32 v9, vcc_lo, s4, v7 v_add_co_ci_u32_e32 v10, vcc_lo, s5, v8, vcc_lo s_add_i32 s2, s2, -1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_2) s_cmp_eq_u32 s2, 0 global_load_b32 v9, v[9:10], off s_waitcnt vmcnt(0) v_div_scale_f32 v10, null, s1, s1, v9 v_div_scale_f32 v13, vcc_lo, v9, s1, v9 v_rcp_f32_e32 v11, v10 s_waitcnt_depctr 0xfff v_fma_f32 v12, -v10, v11, 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v11, v12, v11 v_mul_f32_e32 v12, v13, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v14, -v10, v12, v13 v_fmac_f32_e32 v12, v14, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v10, -v10, v12, v13 v_div_fmas_f32 v10, v10, v11, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fixup_f32 v9, v10, s1, v9 v_sub_f32_e32 v11, v9, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_f32_e32 v9, 0x3fb8aa3b, v11 v_cmp_ngt_f32_e64 s0, 0xc2ce8ed0, v11 v_fma_f32 v10, v11, 0x3fb8aa3b, -v9 v_rndne_f32_e32 v12, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_fmac_f32 v10, 0x32a5705f, v11 :: v_dual_sub_f32 v9, v9, v12 v_add_f32_e32 v9, v9, v10 v_cvt_i32_f32_e32 v10, v12 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_exp_f32_e32 v9, v9 s_waitcnt_depctr 0xfff v_ldexp_f32 v10, v9, v10 v_add_co_u32 v9, vcc_lo, s6, v7 v_cndmask_b32_e64 v12, 0, v10, s0 v_add_co_ci_u32_e32 v10, vcc_lo, s7, v8, vcc_lo v_cmp_nlt_f32_e32 vcc_lo, 0x42b17218, v11 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v11, 0x7f800000, v12, vcc_lo v_add_co_u32 v7, vcc_lo, v7, s8 v_add_co_ci_u32_e32 v8, vcc_lo, s9, v8, vcc_lo v_add_f32_e32 v4, v4, v11 global_store_b32 v[9:10], v11, off s_cbranch_scc0 .LBB0_7 s_branch .LBB0_9 .LBB0_8: v_mov_b32_e32 v4, 0 .LBB0_9: s_delay_alu instid0(VALU_DEP_2) v_cmp_ne_u32_e32 vcc_lo, 1, v5 s_cbranch_vccnz .LBB0_12 v_lshlrev_b64 v[0:1], 2, v[0:1] v_lshlrev_b64 v[2:3], 2, v[2:3] s_ashr_i32 s1, s3, 31 s_mov_b32 s0, s3 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) s_lshl_b64 s[0:1], s[0:1], 2 s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, v0, v2 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, v1, v3, vcc_lo .p2align 6 .LBB0_11: global_load_b32 v2, v[0:1], off s_add_i32 s10, s10, -1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) s_cmp_lg_u32 s10, 0 s_waitcnt vmcnt(0) v_div_scale_f32 v3, null, v4, v4, v2 v_div_scale_f32 v7, vcc_lo, v2, v4, v2 v_rcp_f32_e32 v5, v3 s_waitcnt_depctr 0xfff v_fma_f32 v6, -v3, v5, 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v5, v6, v5 v_mul_f32_e32 v6, v7, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v8, -v3, v6, v7 v_fmac_f32_e32 v6, v8, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v3, -v3, v6, v7 v_div_fmas_f32 v3, v3, v5, v6 s_delay_alu instid0(VALU_DEP_1) v_div_fixup_f32 v2, v3, v4, v2 global_store_b32 v[0:1], v2, off v_add_co_u32 v0, vcc_lo, v0, s0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_cbranch_scc1 .LBB0_11 .LBB0_12: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14softmax_kernelPfiiiiiifS_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 304 .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 0 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z14softmax_kernelPfiiiiiifS_, .Lfunc_end0-_Z14softmax_kernelPfiiiiiifS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .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: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .offset: 48 .size: 4 .value_kind: hidden_block_count_x - .offset: 52 .size: 4 .value_kind: hidden_block_count_y - .offset: 56 .size: 4 .value_kind: hidden_block_count_z - .offset: 60 .size: 2 .value_kind: hidden_group_size_x - .offset: 62 .size: 2 .value_kind: hidden_group_size_y - .offset: 64 .size: 2 .value_kind: hidden_group_size_z - .offset: 66 .size: 2 .value_kind: hidden_remainder_x - .offset: 68 .size: 2 .value_kind: hidden_remainder_y - .offset: 70 .size: 2 .value_kind: hidden_remainder_z - .offset: 88 .size: 8 .value_kind: hidden_global_offset_x - .offset: 96 .size: 8 .value_kind: hidden_global_offset_y - .offset: 104 .size: 8 .value_kind: hidden_global_offset_z - .offset: 112 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 304 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14softmax_kernelPfiiiiiifS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z14softmax_kernelPfiiiiiifS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __device__ void softmax_device(float *input, int n, float temp, int stride, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i*stride]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = expf(input[i*stride]/temp - largest/temp); sum += e; output[i*stride] = e; } for(i = 0; i < n; ++i){ output[i*stride] /= sum; } } __device__ void softmax_device(int n, float *input, float temp, float *output) { int i; float sum = 0; float largest = -INFINITY; for(i = 0; i < n; ++i){ int val = input[i]; largest = (val>largest) ? val : largest; } for(i = 0; i < n; ++i){ float e = exp(input[i]/temp - largest/temp); sum += e; output[i] = e; } for(i = 0; i < n; ++i){ output[i] /= sum; } } __global__ void softmax_kernel(float *input, int n, int batch, int batch_offset, int groups, int group_offset, int stride, float temp, float *output) { int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x; if (id >= batch*groups) return; int b = id / groups; int g = id % groups; softmax_device(input + b*batch_offset + g*group_offset, n, temp, stride, output + b*batch_offset + g*group_offset); }
.text .file "softmax_kernel.hip" .globl _Z29__device_stub__softmax_kernelPfiiiiiifS_ # -- Begin function _Z29__device_stub__softmax_kernelPfiiiiiifS_ .p2align 4, 0x90 .type _Z29__device_stub__softmax_kernelPfiiiiiifS_,@function _Z29__device_stub__softmax_kernelPfiiiiiifS_: # @_Z29__device_stub__softmax_kernelPfiiiiiifS_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movl %ecx, 12(%rsp) movl %r8d, 8(%rsp) movl %r9d, 4(%rsp) movss %xmm0, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 20(%rsp), %rax movq %rax, 88(%rsp) leaq 16(%rsp), %rax movq %rax, 96(%rsp) leaq 12(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 168(%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 80(%rsp), %r9 movl $_Z14softmax_kernelPfiiiiiifS_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z29__device_stub__softmax_kernelPfiiiiiifS_, .Lfunc_end0-_Z29__device_stub__softmax_kernelPfiiiiiifS_ .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 $_Z14softmax_kernelPfiiiiiifS_, %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 _Z14softmax_kernelPfiiiiiifS_,@object # @_Z14softmax_kernelPfiiiiiifS_ .section .rodata,"a",@progbits .globl _Z14softmax_kernelPfiiiiiifS_ .p2align 3, 0x0 _Z14softmax_kernelPfiiiiiifS_: .quad _Z29__device_stub__softmax_kernelPfiiiiiifS_ .size _Z14softmax_kernelPfiiiiiifS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14softmax_kernelPfiiiiiifS_" .size .L__unnamed_1, 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 _Z29__device_stub__softmax_kernelPfiiiiiifS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14softmax_kernelPfiiiiiifS_ .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_0009d388_00000000-6_softmax_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2031: .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 .LFE2031: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z14softmax_devicePfifiS_ .type _Z14softmax_devicePfifiS_, @function _Z14softmax_devicePfifiS_: .LFB2027: .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 .LFE2027: .size _Z14softmax_devicePfifiS_, .-_Z14softmax_devicePfifiS_ .globl _Z14softmax_deviceiPffS_ .type _Z14softmax_deviceiPffS_, @function _Z14softmax_deviceiPffS_: .LFB2028: .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 .LFE2028: .size _Z14softmax_deviceiPffS_, .-_Z14softmax_deviceiPffS_ .globl _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ .type _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_, @function _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_: .LFB2053: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movl %r9d, 20(%rsp) movss %xmm0, 16(%rsp) movq 216(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 28(%rsp), %rax movq %rax, 136(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) leaq 20(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 184(%rsp), %rax subq %fs:40, %rax jne .L12 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z14softmax_kernelPfiiiiiifS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2053: .size _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_, .-_Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ .globl _Z14softmax_kernelPfiiiiiifS_ .type _Z14softmax_kernelPfiiiiiifS_, @function _Z14softmax_kernelPfiiiiiifS_: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z43__device_stub__Z14softmax_kernelPfiiiiiifS_PfiiiiiifS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _Z14softmax_kernelPfiiiiiifS_, .-_Z14softmax_kernelPfiiiiiifS_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z14softmax_kernelPfiiiiiifS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2056: .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 _Z14softmax_kernelPfiiiiiifS_(%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 .LFE2056: .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 "softmax_kernel.hip" .globl _Z29__device_stub__softmax_kernelPfiiiiiifS_ # -- Begin function _Z29__device_stub__softmax_kernelPfiiiiiifS_ .p2align 4, 0x90 .type _Z29__device_stub__softmax_kernelPfiiiiiifS_,@function _Z29__device_stub__softmax_kernelPfiiiiiifS_: # @_Z29__device_stub__softmax_kernelPfiiiiiifS_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movl %ecx, 12(%rsp) movl %r8d, 8(%rsp) movl %r9d, 4(%rsp) movss %xmm0, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 20(%rsp), %rax movq %rax, 88(%rsp) leaq 16(%rsp), %rax movq %rax, 96(%rsp) leaq 12(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 168(%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 80(%rsp), %r9 movl $_Z14softmax_kernelPfiiiiiifS_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z29__device_stub__softmax_kernelPfiiiiiifS_, .Lfunc_end0-_Z29__device_stub__softmax_kernelPfiiiiiifS_ .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 $_Z14softmax_kernelPfiiiiiifS_, %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 _Z14softmax_kernelPfiiiiiifS_,@object # @_Z14softmax_kernelPfiiiiiifS_ .section .rodata,"a",@progbits .globl _Z14softmax_kernelPfiiiiiifS_ .p2align 3, 0x0 _Z14softmax_kernelPfiiiiiifS_: .quad _Z29__device_stub__softmax_kernelPfiiiiiifS_ .size _Z14softmax_kernelPfiiiiiifS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z14softmax_kernelPfiiiiiifS_" .size .L__unnamed_1, 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 _Z29__device_stub__softmax_kernelPfiiiiiifS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14softmax_kernelPfiiiiiifS_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
code for sm_80 Function : _Z4fillPffi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x16c], PT ; /* 0x00005b0002007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0070*/ MOV R5, c[0x0][0x168] ; /* 0x00005a0000057a02 */ /* 0x000fe20000000f00 */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*00a0*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
.file "tmpxft_00013e51_00000000-6_fill.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 _Z25__device_stub__Z4fillPffiPffi .type _Z25__device_stub__Z4fillPffiPffi, @function _Z25__device_stub__Z4fillPffiPffi: .LFB2051: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movss %xmm0, 4(%rsp) movl %esi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z4fillPffi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z25__device_stub__Z4fillPffiPffi, .-_Z25__device_stub__Z4fillPffiPffi .globl _Z4fillPffi .type _Z4fillPffi, @function _Z4fillPffi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z4fillPffiPffi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4fillPffi, .-_Z4fillPffi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4fillPffi" .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 _Z4fillPffi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4fillPffi .globl _Z4fillPffi .p2align 8 .type _Z4fillPffi,@function _Z4fillPffi: s_clause 0x1 s_load_b32 s2, s[0:1], 0x1c s_load_b32 s3, s[0:1], 0xc 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_2 s_clause 0x1 s_load_b64 s[2:3], s[0:1], 0x0 s_load_b32 s0, s[0:1], 0x8 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[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo v_mov_b32_e32 v2, s0 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 _Z4fillPffi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .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 _Z4fillPffi, .Lfunc_end0-_Z4fillPffi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4fillPffi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4fillPffi.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void fill(float * w, float val, int size) { const int tid = (blockIdx.x * blockDim.x) + threadIdx.x; if (tid < size) w[tid] = val; }
.text .file "fill.hip" .globl _Z19__device_stub__fillPffi # -- Begin function _Z19__device_stub__fillPffi .p2align 4, 0x90 .type _Z19__device_stub__fillPffi,@function _Z19__device_stub__fillPffi: # @_Z19__device_stub__fillPffi .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movss %xmm0, 4(%rsp) movl %esi, (%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 4(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z4fillPffi, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z19__device_stub__fillPffi, .Lfunc_end0-_Z19__device_stub__fillPffi .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 $_Z4fillPffi, %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 _Z4fillPffi,@object # @_Z4fillPffi .section .rodata,"a",@progbits .globl _Z4fillPffi .p2align 3, 0x0 _Z4fillPffi: .quad _Z19__device_stub__fillPffi .size _Z4fillPffi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4fillPffi" .size .L__unnamed_1, 12 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__fillPffi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4fillPffi .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 : _Z4fillPffi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x16c], PT ; /* 0x00005b0002007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0070*/ MOV R5, c[0x0][0x168] ; /* 0x00005a0000057a02 */ /* 0x000fe20000000f00 */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*00a0*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4fillPffi .globl _Z4fillPffi .p2align 8 .type _Z4fillPffi,@function _Z4fillPffi: s_clause 0x1 s_load_b32 s2, s[0:1], 0x1c s_load_b32 s3, s[0:1], 0xc 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_2 s_clause 0x1 s_load_b64 s[2:3], s[0:1], 0x0 s_load_b32 s0, s[0:1], 0x8 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[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s2, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo v_mov_b32_e32 v2, s0 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 _Z4fillPffi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .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 _Z4fillPffi, .Lfunc_end0-_Z4fillPffi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4fillPffi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4fillPffi.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00013e51_00000000-6_fill.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 _Z25__device_stub__Z4fillPffiPffi .type _Z25__device_stub__Z4fillPffiPffi, @function _Z25__device_stub__Z4fillPffiPffi: .LFB2051: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movss %xmm0, 4(%rsp) movl %esi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z4fillPffi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z25__device_stub__Z4fillPffiPffi, .-_Z25__device_stub__Z4fillPffiPffi .globl _Z4fillPffi .type _Z4fillPffi, @function _Z4fillPffi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z4fillPffiPffi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4fillPffi, .-_Z4fillPffi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4fillPffi" .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 _Z4fillPffi(%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 "fill.hip" .globl _Z19__device_stub__fillPffi # -- Begin function _Z19__device_stub__fillPffi .p2align 4, 0x90 .type _Z19__device_stub__fillPffi,@function _Z19__device_stub__fillPffi: # @_Z19__device_stub__fillPffi .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movss %xmm0, 4(%rsp) movl %esi, (%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 4(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z4fillPffi, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z19__device_stub__fillPffi, .Lfunc_end0-_Z19__device_stub__fillPffi .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 $_Z4fillPffi, %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 _Z4fillPffi,@object # @_Z4fillPffi .section .rodata,"a",@progbits .globl _Z4fillPffi .p2align 3, 0x0 _Z4fillPffi: .quad _Z19__device_stub__fillPffi .size _Z4fillPffi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4fillPffi" .size .L__unnamed_1, 12 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__fillPffi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4fillPffi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
code for sm_80 Function : _Z26merge_without_blend_kernelPfS_S_Piiii .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ IMAD.MOV.U32 R0, RZ, RZ, c[0x0][0x180] ; /* 0x00006000ff007624 */ /* 0x000fca00078e00ff */ /*0020*/ ISETP.GE.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fda0003f06270 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ IADD3 R4, R0.reuse, -0x1, RZ ; /* 0xffffffff00047810 */ /* 0x040fe20007ffe0ff */ /*0050*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000e220000002600 */ /*0060*/ LOP3.LUT R0, R0, 0x3, RZ, 0xc0, !PT ; /* 0x0000000300007812 */ /* 0x000fe200078ec0ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0080*/ ISETP.GE.U32.AND P1, PT, R4, 0x3, PT ; /* 0x000000030400780c */ /* 0x000fe20003f26070 */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000e620000002200 */ /*00a0*/ HFMA2.MMA R20, -RZ, RZ, 0, 0 ; /* 0x00000000ff147435 */ /* 0x000fe200000001ff */ /*00b0*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe40003f05270 */ /*00c0*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000ea80000002500 */ /*00d0*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000ee80000002100 */ /*00e0*/ @!P1 BRA 0x4c0 ; /* 0x000003d000009947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R7, R3, c[0x0][0x188], RZ ; /* 0x0000620003077a10 */ /* 0x002fe20007ffe0ff */ /*0100*/ IMAD R6, R2, c[0x0][0x4], R3 ; /* 0x0000010002067a24 */ /* 0x001fe200078e0203 */ /*0110*/ IADD3 R11, R0, -c[0x0][0x180], RZ ; /* 0x80006000000b7a10 */ /* 0x000fe20007ffe0ff */ /*0120*/ IMAD.MOV.U32 R13, RZ, RZ, c[0x0][0x188] ; /* 0x00006200ff0d7624 */ /* 0x000fc400078e00ff */ /*0130*/ IMAD R9, R4, c[0x0][0x0], R5 ; /* 0x0000000004097a24 */ /* 0x00cfe400078e0205 */ /*0140*/ IMAD R8, R2, c[0x0][0x4], R7 ; /* 0x0000010002087a24 */ /* 0x000fe200078e0207 */ /*0150*/ LEA R10, R13.reuse, R6, 0x1 ; /* 0x000000060d0a7211 */ /* 0x040fe200078e08ff */ /*0160*/ IMAD R12, R13, 0x3, R6 ; /* 0x000000030d0c7824 */ /* 0x000fe400078e0206 */ /*0170*/ IMAD R7, R8, c[0x0][0x184], R9.reuse ; /* 0x0000610008077a24 */ /* 0x100fe400078e0209 */ /*0180*/ IMAD R6, R6, c[0x0][0x184], R9.reuse ; /* 0x0000610006067a24 */ /* 0x100fe400078e0209 */ /*0190*/ IMAD R8, R10, c[0x0][0x184], R9 ; /* 0x000061000a087a24 */ /* 0x000fc400078e0209 */ /*01a0*/ IMAD R9, R12, c[0x0][0x184], R9 ; /* 0x000061000c097a24 */ /* 0x000fe400078e0209 */ /*01b0*/ IMAD.MOV.U32 R20, RZ, RZ, RZ ; /* 0x000000ffff147224 */ /* 0x000fe400078e00ff */ /*01c0*/ IMAD R10, R13, c[0x0][0x184], RZ ; /* 0x000061000d0a7a24 */ /* 0x000fe400078e02ff */ /*01d0*/ MOV R15, 0x4 ; /* 0x00000004000f7802 */ /* 0x001fca0000000f00 */ /*01e0*/ IMAD.WIDE R16, R6, R15, c[0x0][0x178] ; /* 0x00005e0006107625 */ /* 0x000fca00078e020f */ /*01f0*/ LDG.E R12, [R16.64] ; /* 0x00000004100c7981 */ /* 0x000ea2000c1e1900 */ /*0200*/ IMAD.MOV.U32 R22, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff167624 */ /* 0x000fe200078e00ff */ /*0210*/ MOV R21, c[0x0][0x16c] ; /* 0x00005b0000157a02 */ /* 0x000fe40000000f00 */ /*0220*/ ISETP.NE.AND P1, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x004fc80003f25270 */ /*0230*/ SEL R12, R22, c[0x0][0x160], !P1 ; /* 0x00005800160c7a07 */ /* 0x000fe40004800000 */ /*0240*/ SEL R13, R21, c[0x0][0x164], !P1 ; /* 0x00005900150d7a07 */ /* 0x000fca0004800000 */ /*0250*/ IMAD.WIDE R18, R6, 0x4, R12 ; /* 0x0000000406127825 */ /* 0x000fca00078e020c */ /*0260*/ LDG.E R23, [R18.64] ; /* 0x0000000412177981 */ /* 0x000ea2000c1e1900 */ /*0270*/ IMAD.WIDE R14, R6, R15, c[0x0][0x170] ; /* 0x00005c00060e7625 */ /* 0x000fc800078e020f */ /*0280*/ IMAD.WIDE R12, R10, 0x4, R16 ; /* 0x000000040a0c7825 */ /* 0x000fe200078e0210 */ /*0290*/ STG.E [R14.64], R23 ; /* 0x000000170e007986 */ /* 0x0041e8000c101904 */ /*02a0*/ LDG.E R16, [R12.64] ; /* 0x000000040c107981 */ /* 0x000ea4000c1e1900 */ /*02b0*/ ISETP.NE.AND P1, PT, R16, RZ, PT ; /* 0x000000ff1000720c */ /* 0x004fc80003f25270 */ /*02c0*/ SEL R16, R22, c[0x0][0x160], !P1 ; /* 0x0000580016107a07 */ /* 0x000fe40004800000 */ /*02d0*/ SEL R17, R21, c[0x0][0x164], !P1 ; /* 0x0000590015117a07 */ /* 0x000fca0004800000 */ /*02e0*/ IMAD.WIDE R24, R7, 0x4, R16 ; /* 0x0000000407187825 */ /* 0x000fca00078e0210 */ /*02f0*/ LDG.E R27, [R24.64] ; /* 0x00000004181b7981 */ /* 0x000ea2000c1e1900 */ /*0300*/ IMAD.WIDE R16, R10, 0x4, R14 ; /* 0x000000040a107825 */ /* 0x000fc800078e020e */ /*0310*/ IMAD.WIDE R18, R10, 0x4, R12 ; /* 0x000000040a127825 */ /* 0x000fe200078e020c */ /*0320*/ STG.E [R16.64], R27 ; /* 0x0000001b10007986 */ /* 0x0043e8000c101904 */ /*0330*/ LDG.E R26, [R18.64] ; /* 0x00000004121a7981 */ /* 0x000ea4000c1e1900 */ /*0340*/ ISETP.NE.AND P1, PT, R26, RZ, PT ; /* 0x000000ff1a00720c */ /* 0x004fc80003f25270 */ /*0350*/ SEL R12, R22, c[0x0][0x160], !P1 ; /* 0x00005800160c7a07 */ /* 0x000fe40004800000 */ /*0360*/ SEL R13, R21, c[0x0][0x164], !P1 ; /* 0x00005900150d7a07 */ /* 0x000fca0004800000 */ /*0370*/ IMAD.WIDE R12, R8, 0x4, R12 ; /* 0x00000004080c7825 */ /* 0x000fcc00078e020c */ /*0380*/ LDG.E R13, [R12.64] ; /* 0x000000040c0d7981 */ /* 0x000ea2000c1e1900 */ /*0390*/ IMAD.WIDE R14, R10, 0x4, R16 ; /* 0x000000040a0e7825 */ /* 0x001fc800078e0210 */ /*03a0*/ IMAD.WIDE R24, R10, 0x4, R18 ; /* 0x000000040a187825 */ /* 0x000fe200078e0212 */ /*03b0*/ STG.E [R14.64], R13 ; /* 0x0000000d0e007986 */ /* 0x0041ea000c101904 */ /*03c0*/ LDG.E R24, [R24.64] ; /* 0x0000000418187981 */ /* 0x000ea2000c1e1900 */ /*03d0*/ IADD3 R20, R20, 0x4, RZ ; /* 0x0000000414147810 */ /* 0x000fe40007ffe0ff */ /*03e0*/ ISETP.NE.AND P1, PT, R24, RZ, PT ; /* 0x000000ff1800720c */ /* 0x004fc80003f25270 */ /*03f0*/ SEL R22, R22, c[0x0][0x160], !P1 ; /* 0x0000580016167a07 */ /* 0x000fe40004800000 */ /*0400*/ SEL R23, R21, c[0x0][0x164], !P1 ; /* 0x0000590015177a07 */ /* 0x000fca0004800000 */ /*0410*/ IMAD.WIDE R22, R9, 0x4, R22 ; /* 0x0000000409167825 */ /* 0x000fcc00078e0216 */ /*0420*/ LDG.E R23, [R22.64] ; /* 0x0000000416177981 */ /* 0x000ea2000c1e1900 */ /*0430*/ IMAD.IADD R12, R11, 0x1, R20 ; /* 0x000000010b0c7824 */ /* 0x000fe200078e0214 */ /*0440*/ LEA R6, R10.reuse, R6, 0x2 ; /* 0x000000060a067211 */ /* 0x040fe200078e10ff */ /*0450*/ IMAD.WIDE R16, R10.reuse, 0x4, R14 ; /* 0x000000040a107825 */ /* 0x042fe200078e020e */ /*0460*/ LEA R7, R10, R7, 0x2 ; /* 0x000000070a077211 */ /* 0x000fe400078e10ff */ /*0470*/ ISETP.NE.AND P1, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x000fe20003f25270 */ /*0480*/ IMAD R8, R10.reuse, 0x4, R8 ; /* 0x000000040a087824 */ /* 0x040fe200078e0208 */ /*0490*/ LEA R9, R10, R9, 0x2 ; /* 0x000000090a097211 */ /* 0x000fe200078e10ff */ /*04a0*/ STG.E [R16.64], R23 ; /* 0x0000001710007986 */ /* 0x0041f4000c101904 */ /*04b0*/ @P1 BRA 0x1d0 ; /* 0xfffffd1000001947 */ /* 0x000fea000383ffff */ /*04c0*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*04d0*/ IMAD R3, R2, c[0x0][0x4], R3 ; /* 0x0000010002037a24 */ /* 0x003fe400078e0203 */ /*04e0*/ IMAD R5, R4, c[0x0][0x0], R5 ; /* 0x0000000004057a24 */ /* 0x00cfc400078e0205 */ /*04f0*/ IMAD R20, R20, c[0x0][0x188], R3 ; /* 0x0000620014147a24 */ /* 0x000fc800078e0203 */ /*0500*/ IMAD R9, R20, c[0x0][0x184], R5 ; /* 0x0000610014097a24 */ /* 0x000fe400078e0205 */ /*0510*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x001fd400000001ff */ /*0520*/ IMAD.WIDE R2, R9, R6, c[0x0][0x178] ; /* 0x00005e0009027625 */ /* 0x000fcc00078e0206 */ /*0530*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*0540*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff047624 */ /* 0x000fe200078e00ff */ /*0550*/ MOV R5, c[0x0][0x16c] ; /* 0x00005b0000057a02 */ /* 0x000fe40000000f00 */ /*0560*/ ISETP.NE.AND P0, PT, R2, RZ, PT ; /* 0x000000ff0200720c */ /* 0x004fc80003f05270 */ /*0570*/ SEL R4, R4, c[0x0][0x160], !P0 ; /* 0x0000580004047a07 */ /* 0x000fe40004000000 */ /*0580*/ SEL R5, R5, c[0x0][0x164], !P0 ; /* 0x0000590005057a07 */ /* 0x000fca0004000000 */ /*0590*/ IMAD.WIDE R4, R9, 0x4, R4 ; /* 0x0000000409047825 */ /* 0x000fcc00078e0204 */ /*05a0*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*05b0*/ IADD3 R0, R0, -0x1, RZ ; /* 0xffffffff00007810 */ /* 0x000fe20007ffe0ff */ /*05c0*/ IMAD.WIDE R6, R9, R6, c[0x0][0x170] ; /* 0x00005c0009067625 */ /* 0x000fe200078e0206 */ /*05d0*/ MOV R2, c[0x0][0x188] ; /* 0x0000620000027a02 */ /* 0x000fe40000000f00 */ /*05e0*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fc60003f05270 */ /*05f0*/ IMAD R9, R2, c[0x0][0x184], R9 ; /* 0x0000610002097a24 */ /* 0x000fe200078e0209 */ /*0600*/ STG.E [R6.64], R5 ; /* 0x0000000506007986 */ /* 0x0041f2000c101904 */ /*0610*/ @P0 BRA 0x510 ; /* 0xfffffef000000947 */ /* 0x000fea000383ffff */ /*0620*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0630*/ BRA 0x630; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0640*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0650*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0660*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0670*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0680*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0690*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
.file "tmpxft_000d0fd9_00000000-6_merge_without_blend_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 _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii .type _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii, @function _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii: .LFB2051: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z26merge_without_blend_kernelPfS_S_Piiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii, .-_Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii .globl _Z26merge_without_blend_kernelPfS_S_Piiii .type _Z26merge_without_blend_kernelPfS_S_Piiii, @function _Z26merge_without_blend_kernelPfS_S_Piiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z26merge_without_blend_kernelPfS_S_Piiii, .-_Z26merge_without_blend_kernelPfS_S_Piiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z26merge_without_blend_kernelPfS_S_Piiii" .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 _Z26merge_without_blend_kernelPfS_S_Piiii(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
#include <hip/hip_runtime.h> #include "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z26merge_without_blend_kernelPfS_S_Piiii .globl _Z26merge_without_blend_kernelPfS_S_Piiii .p2align 8 .type _Z26merge_without_blend_kernelPfS_S_Piiii,@function _Z26merge_without_blend_kernelPfS_S_Piiii: s_load_b32 s8, s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s8, 1 s_cbranch_scc1 .LBB0_3 s_clause 0x1 s_load_b32 s9, s[0:1], 0x3c s_load_b64 s[10:11], s[0:1], 0x24 v_bfe_u32 v1, v0, 10, 10 s_load_b256 s[0:7], s[0:1], 0x0 v_and_b32_e32 v0, 0x3ff, v0 s_waitcnt lgkmcnt(0) s_lshr_b32 s12, s9, 16 s_and_b32 s9, s9, 0xffff v_mad_u64_u32 v[2:3], null, s15, s12, v[1:2] v_mov_b32_e32 v3, s2 s_mul_i32 s14, s14, s9 s_mul_i32 s2, s11, s10 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_lo_u32 v1, s10, v2 v_mov_b32_e32 v2, s3 v_add3_u32 v0, v0, v1, s14 .p2align 6 .LBB0_2: s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_ashrrev_i32_e32 v1, 31, v0 s_add_i32 s8, s8, -1 s_cmp_eq_u32 s8, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_lshlrev_b64 v[4:5], 2, v[0:1] v_add_nc_u32_e32 v0, s2, v0 v_add_co_u32 v6, vcc_lo, s6, v4 s_delay_alu instid0(VALU_DEP_3) v_add_co_ci_u32_e32 v7, vcc_lo, s7, v5, vcc_lo global_load_b32 v1, v[6:7], off s_waitcnt vmcnt(0) v_cmp_eq_u32_e32 vcc_lo, 0, v1 v_cndmask_b32_e32 v6, s0, v3, vcc_lo v_cndmask_b32_e32 v1, s1, v2, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v6, vcc_lo, v6, v4 v_add_co_ci_u32_e32 v7, vcc_lo, v1, v5, vcc_lo v_add_co_u32 v4, vcc_lo, s4, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo global_load_b32 v1, v[6:7], off s_waitcnt vmcnt(0) global_store_b32 v[4:5], v1, off s_cbranch_scc0 .LBB0_2 .LBB0_3: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z26merge_without_blend_kernelPfS_S_Piiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 304 .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 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_end0: .size _Z26merge_without_blend_kernelPfS_S_Piiii, .Lfunc_end0-_Z26merge_without_blend_kernelPfS_S_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 - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: hidden_block_count_x - .offset: 52 .size: 4 .value_kind: hidden_block_count_y - .offset: 56 .size: 4 .value_kind: hidden_block_count_z - .offset: 60 .size: 2 .value_kind: hidden_group_size_x - .offset: 62 .size: 2 .value_kind: hidden_group_size_y - .offset: 64 .size: 2 .value_kind: hidden_group_size_z - .offset: 66 .size: 2 .value_kind: hidden_remainder_x - .offset: 68 .size: 2 .value_kind: hidden_remainder_y - .offset: 70 .size: 2 .value_kind: hidden_remainder_z - .offset: 88 .size: 8 .value_kind: hidden_global_offset_x - .offset: 96 .size: 8 .value_kind: hidden_global_offset_y - .offset: 104 .size: 8 .value_kind: hidden_global_offset_z - .offset: 112 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 304 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z26merge_without_blend_kernelPfS_S_Piiii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z26merge_without_blend_kernelPfS_S_Piiii.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 "includes.h" using namespace std; #define ITERATIONS 40000 enum pixel_position {INSIDE_MASK, BOUNDRY, OUTSIDE}; __global__ void merge_without_blend_kernel(float *srcimg, float *targetimg, float *outimg, int *boundary_array,int source_nchannel, int source_width, int source_height){ int x = threadIdx.x + blockIdx.x * blockDim.x; int y = threadIdx.y + blockIdx.y * blockDim.y; for(int channel = 0; channel < source_nchannel; channel++){ int id = x + y*source_width + channel * source_width * source_height; if(boundary_array[id] == INSIDE_MASK){ outimg[id] = targetimg[id]; } else{ outimg[id] = srcimg[id]; } } }
.text .file "merge_without_blend_kernel.hip" .globl _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii # -- Begin function _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .p2align 4, 0x90 .type _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii,@function _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii: # @_Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%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 $_Z26merge_without_blend_kernelPfS_S_Piiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii, .Lfunc_end0-_Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .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 $_Z26merge_without_blend_kernelPfS_S_Piiii, %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 _Z26merge_without_blend_kernelPfS_S_Piiii,@object # @_Z26merge_without_blend_kernelPfS_S_Piiii .section .rodata,"a",@progbits .globl _Z26merge_without_blend_kernelPfS_S_Piiii .p2align 3, 0x0 _Z26merge_without_blend_kernelPfS_S_Piiii: .quad _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .size _Z26merge_without_blend_kernelPfS_S_Piiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z26merge_without_blend_kernelPfS_S_Piiii" .size .L__unnamed_1, 42 .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 _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z26merge_without_blend_kernelPfS_S_Piiii .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 : _Z26merge_without_blend_kernelPfS_S_Piiii .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ IMAD.MOV.U32 R0, RZ, RZ, c[0x0][0x180] ; /* 0x00006000ff007624 */ /* 0x000fca00078e00ff */ /*0020*/ ISETP.GE.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fda0003f06270 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ IADD3 R4, R0.reuse, -0x1, RZ ; /* 0xffffffff00047810 */ /* 0x040fe20007ffe0ff */ /*0050*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000e220000002600 */ /*0060*/ LOP3.LUT R0, R0, 0x3, RZ, 0xc0, !PT ; /* 0x0000000300007812 */ /* 0x000fe200078ec0ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0080*/ ISETP.GE.U32.AND P1, PT, R4, 0x3, PT ; /* 0x000000030400780c */ /* 0x000fe20003f26070 */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000e620000002200 */ /*00a0*/ HFMA2.MMA R20, -RZ, RZ, 0, 0 ; /* 0x00000000ff147435 */ /* 0x000fe200000001ff */ /*00b0*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe40003f05270 */ /*00c0*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000ea80000002500 */ /*00d0*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000ee80000002100 */ /*00e0*/ @!P1 BRA 0x4c0 ; /* 0x000003d000009947 */ /* 0x000fea0003800000 */ /*00f0*/ IADD3 R7, R3, c[0x0][0x188], RZ ; /* 0x0000620003077a10 */ /* 0x002fe20007ffe0ff */ /*0100*/ IMAD R6, R2, c[0x0][0x4], R3 ; /* 0x0000010002067a24 */ /* 0x001fe200078e0203 */ /*0110*/ IADD3 R11, R0, -c[0x0][0x180], RZ ; /* 0x80006000000b7a10 */ /* 0x000fe20007ffe0ff */ /*0120*/ IMAD.MOV.U32 R13, RZ, RZ, c[0x0][0x188] ; /* 0x00006200ff0d7624 */ /* 0x000fc400078e00ff */ /*0130*/ IMAD R9, R4, c[0x0][0x0], R5 ; /* 0x0000000004097a24 */ /* 0x00cfe400078e0205 */ /*0140*/ IMAD R8, R2, c[0x0][0x4], R7 ; /* 0x0000010002087a24 */ /* 0x000fe200078e0207 */ /*0150*/ LEA R10, R13.reuse, R6, 0x1 ; /* 0x000000060d0a7211 */ /* 0x040fe200078e08ff */ /*0160*/ IMAD R12, R13, 0x3, R6 ; /* 0x000000030d0c7824 */ /* 0x000fe400078e0206 */ /*0170*/ IMAD R7, R8, c[0x0][0x184], R9.reuse ; /* 0x0000610008077a24 */ /* 0x100fe400078e0209 */ /*0180*/ IMAD R6, R6, c[0x0][0x184], R9.reuse ; /* 0x0000610006067a24 */ /* 0x100fe400078e0209 */ /*0190*/ IMAD R8, R10, c[0x0][0x184], R9 ; /* 0x000061000a087a24 */ /* 0x000fc400078e0209 */ /*01a0*/ IMAD R9, R12, c[0x0][0x184], R9 ; /* 0x000061000c097a24 */ /* 0x000fe400078e0209 */ /*01b0*/ IMAD.MOV.U32 R20, RZ, RZ, RZ ; /* 0x000000ffff147224 */ /* 0x000fe400078e00ff */ /*01c0*/ IMAD R10, R13, c[0x0][0x184], RZ ; /* 0x000061000d0a7a24 */ /* 0x000fe400078e02ff */ /*01d0*/ MOV R15, 0x4 ; /* 0x00000004000f7802 */ /* 0x001fca0000000f00 */ /*01e0*/ IMAD.WIDE R16, R6, R15, c[0x0][0x178] ; /* 0x00005e0006107625 */ /* 0x000fca00078e020f */ /*01f0*/ LDG.E R12, [R16.64] ; /* 0x00000004100c7981 */ /* 0x000ea2000c1e1900 */ /*0200*/ IMAD.MOV.U32 R22, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff167624 */ /* 0x000fe200078e00ff */ /*0210*/ MOV R21, c[0x0][0x16c] ; /* 0x00005b0000157a02 */ /* 0x000fe40000000f00 */ /*0220*/ ISETP.NE.AND P1, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x004fc80003f25270 */ /*0230*/ SEL R12, R22, c[0x0][0x160], !P1 ; /* 0x00005800160c7a07 */ /* 0x000fe40004800000 */ /*0240*/ SEL R13, R21, c[0x0][0x164], !P1 ; /* 0x00005900150d7a07 */ /* 0x000fca0004800000 */ /*0250*/ IMAD.WIDE R18, R6, 0x4, R12 ; /* 0x0000000406127825 */ /* 0x000fca00078e020c */ /*0260*/ LDG.E R23, [R18.64] ; /* 0x0000000412177981 */ /* 0x000ea2000c1e1900 */ /*0270*/ IMAD.WIDE R14, R6, R15, c[0x0][0x170] ; /* 0x00005c00060e7625 */ /* 0x000fc800078e020f */ /*0280*/ IMAD.WIDE R12, R10, 0x4, R16 ; /* 0x000000040a0c7825 */ /* 0x000fe200078e0210 */ /*0290*/ STG.E [R14.64], R23 ; /* 0x000000170e007986 */ /* 0x0041e8000c101904 */ /*02a0*/ LDG.E R16, [R12.64] ; /* 0x000000040c107981 */ /* 0x000ea4000c1e1900 */ /*02b0*/ ISETP.NE.AND P1, PT, R16, RZ, PT ; /* 0x000000ff1000720c */ /* 0x004fc80003f25270 */ /*02c0*/ SEL R16, R22, c[0x0][0x160], !P1 ; /* 0x0000580016107a07 */ /* 0x000fe40004800000 */ /*02d0*/ SEL R17, R21, c[0x0][0x164], !P1 ; /* 0x0000590015117a07 */ /* 0x000fca0004800000 */ /*02e0*/ IMAD.WIDE R24, R7, 0x4, R16 ; /* 0x0000000407187825 */ /* 0x000fca00078e0210 */ /*02f0*/ LDG.E R27, [R24.64] ; /* 0x00000004181b7981 */ /* 0x000ea2000c1e1900 */ /*0300*/ IMAD.WIDE R16, R10, 0x4, R14 ; /* 0x000000040a107825 */ /* 0x000fc800078e020e */ /*0310*/ IMAD.WIDE R18, R10, 0x4, R12 ; /* 0x000000040a127825 */ /* 0x000fe200078e020c */ /*0320*/ STG.E [R16.64], R27 ; /* 0x0000001b10007986 */ /* 0x0043e8000c101904 */ /*0330*/ LDG.E R26, [R18.64] ; /* 0x00000004121a7981 */ /* 0x000ea4000c1e1900 */ /*0340*/ ISETP.NE.AND P1, PT, R26, RZ, PT ; /* 0x000000ff1a00720c */ /* 0x004fc80003f25270 */ /*0350*/ SEL R12, R22, c[0x0][0x160], !P1 ; /* 0x00005800160c7a07 */ /* 0x000fe40004800000 */ /*0360*/ SEL R13, R21, c[0x0][0x164], !P1 ; /* 0x00005900150d7a07 */ /* 0x000fca0004800000 */ /*0370*/ IMAD.WIDE R12, R8, 0x4, R12 ; /* 0x00000004080c7825 */ /* 0x000fcc00078e020c */ /*0380*/ LDG.E R13, [R12.64] ; /* 0x000000040c0d7981 */ /* 0x000ea2000c1e1900 */ /*0390*/ IMAD.WIDE R14, R10, 0x4, R16 ; /* 0x000000040a0e7825 */ /* 0x001fc800078e0210 */ /*03a0*/ IMAD.WIDE R24, R10, 0x4, R18 ; /* 0x000000040a187825 */ /* 0x000fe200078e0212 */ /*03b0*/ STG.E [R14.64], R13 ; /* 0x0000000d0e007986 */ /* 0x0041ea000c101904 */ /*03c0*/ LDG.E R24, [R24.64] ; /* 0x0000000418187981 */ /* 0x000ea2000c1e1900 */ /*03d0*/ IADD3 R20, R20, 0x4, RZ ; /* 0x0000000414147810 */ /* 0x000fe40007ffe0ff */ /*03e0*/ ISETP.NE.AND P1, PT, R24, RZ, PT ; /* 0x000000ff1800720c */ /* 0x004fc80003f25270 */ /*03f0*/ SEL R22, R22, c[0x0][0x160], !P1 ; /* 0x0000580016167a07 */ /* 0x000fe40004800000 */ /*0400*/ SEL R23, R21, c[0x0][0x164], !P1 ; /* 0x0000590015177a07 */ /* 0x000fca0004800000 */ /*0410*/ IMAD.WIDE R22, R9, 0x4, R22 ; /* 0x0000000409167825 */ /* 0x000fcc00078e0216 */ /*0420*/ LDG.E R23, [R22.64] ; /* 0x0000000416177981 */ /* 0x000ea2000c1e1900 */ /*0430*/ IMAD.IADD R12, R11, 0x1, R20 ; /* 0x000000010b0c7824 */ /* 0x000fe200078e0214 */ /*0440*/ LEA R6, R10.reuse, R6, 0x2 ; /* 0x000000060a067211 */ /* 0x040fe200078e10ff */ /*0450*/ IMAD.WIDE R16, R10.reuse, 0x4, R14 ; /* 0x000000040a107825 */ /* 0x042fe200078e020e */ /*0460*/ LEA R7, R10, R7, 0x2 ; /* 0x000000070a077211 */ /* 0x000fe400078e10ff */ /*0470*/ ISETP.NE.AND P1, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x000fe20003f25270 */ /*0480*/ IMAD R8, R10.reuse, 0x4, R8 ; /* 0x000000040a087824 */ /* 0x040fe200078e0208 */ /*0490*/ LEA R9, R10, R9, 0x2 ; /* 0x000000090a097211 */ /* 0x000fe200078e10ff */ /*04a0*/ STG.E [R16.64], R23 ; /* 0x0000001710007986 */ /* 0x0041f4000c101904 */ /*04b0*/ @P1 BRA 0x1d0 ; /* 0xfffffd1000001947 */ /* 0x000fea000383ffff */ /*04c0*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*04d0*/ IMAD R3, R2, c[0x0][0x4], R3 ; /* 0x0000010002037a24 */ /* 0x003fe400078e0203 */ /*04e0*/ IMAD R5, R4, c[0x0][0x0], R5 ; /* 0x0000000004057a24 */ /* 0x00cfc400078e0205 */ /*04f0*/ IMAD R20, R20, c[0x0][0x188], R3 ; /* 0x0000620014147a24 */ /* 0x000fc800078e0203 */ /*0500*/ IMAD R9, R20, c[0x0][0x184], R5 ; /* 0x0000610014097a24 */ /* 0x000fe400078e0205 */ /*0510*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x001fd400000001ff */ /*0520*/ IMAD.WIDE R2, R9, R6, c[0x0][0x178] ; /* 0x00005e0009027625 */ /* 0x000fcc00078e0206 */ /*0530*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1900 */ /*0540*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff047624 */ /* 0x000fe200078e00ff */ /*0550*/ MOV R5, c[0x0][0x16c] ; /* 0x00005b0000057a02 */ /* 0x000fe40000000f00 */ /*0560*/ ISETP.NE.AND P0, PT, R2, RZ, PT ; /* 0x000000ff0200720c */ /* 0x004fc80003f05270 */ /*0570*/ SEL R4, R4, c[0x0][0x160], !P0 ; /* 0x0000580004047a07 */ /* 0x000fe40004000000 */ /*0580*/ SEL R5, R5, c[0x0][0x164], !P0 ; /* 0x0000590005057a07 */ /* 0x000fca0004000000 */ /*0590*/ IMAD.WIDE R4, R9, 0x4, R4 ; /* 0x0000000409047825 */ /* 0x000fcc00078e0204 */ /*05a0*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*05b0*/ IADD3 R0, R0, -0x1, RZ ; /* 0xffffffff00007810 */ /* 0x000fe20007ffe0ff */ /*05c0*/ IMAD.WIDE R6, R9, R6, c[0x0][0x170] ; /* 0x00005c0009067625 */ /* 0x000fe200078e0206 */ /*05d0*/ MOV R2, c[0x0][0x188] ; /* 0x0000620000027a02 */ /* 0x000fe40000000f00 */ /*05e0*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fc60003f05270 */ /*05f0*/ IMAD R9, R2, c[0x0][0x184], R9 ; /* 0x0000610002097a24 */ /* 0x000fe200078e0209 */ /*0600*/ STG.E [R6.64], R5 ; /* 0x0000000506007986 */ /* 0x0041f2000c101904 */ /*0610*/ @P0 BRA 0x510 ; /* 0xfffffef000000947 */ /* 0x000fea000383ffff */ /*0620*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0630*/ BRA 0x630; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0640*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0650*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0660*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0670*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0680*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0690*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z26merge_without_blend_kernelPfS_S_Piiii .globl _Z26merge_without_blend_kernelPfS_S_Piiii .p2align 8 .type _Z26merge_without_blend_kernelPfS_S_Piiii,@function _Z26merge_without_blend_kernelPfS_S_Piiii: s_load_b32 s8, s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s8, 1 s_cbranch_scc1 .LBB0_3 s_clause 0x1 s_load_b32 s9, s[0:1], 0x3c s_load_b64 s[10:11], s[0:1], 0x24 v_bfe_u32 v1, v0, 10, 10 s_load_b256 s[0:7], s[0:1], 0x0 v_and_b32_e32 v0, 0x3ff, v0 s_waitcnt lgkmcnt(0) s_lshr_b32 s12, s9, 16 s_and_b32 s9, s9, 0xffff v_mad_u64_u32 v[2:3], null, s15, s12, v[1:2] v_mov_b32_e32 v3, s2 s_mul_i32 s14, s14, s9 s_mul_i32 s2, s11, s10 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_lo_u32 v1, s10, v2 v_mov_b32_e32 v2, s3 v_add3_u32 v0, v0, v1, s14 .p2align 6 .LBB0_2: s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_ashrrev_i32_e32 v1, 31, v0 s_add_i32 s8, s8, -1 s_cmp_eq_u32 s8, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_lshlrev_b64 v[4:5], 2, v[0:1] v_add_nc_u32_e32 v0, s2, v0 v_add_co_u32 v6, vcc_lo, s6, v4 s_delay_alu instid0(VALU_DEP_3) v_add_co_ci_u32_e32 v7, vcc_lo, s7, v5, vcc_lo global_load_b32 v1, v[6:7], off s_waitcnt vmcnt(0) v_cmp_eq_u32_e32 vcc_lo, 0, v1 v_cndmask_b32_e32 v6, s0, v3, vcc_lo v_cndmask_b32_e32 v1, s1, v2, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v6, vcc_lo, v6, v4 v_add_co_ci_u32_e32 v7, vcc_lo, v1, v5, vcc_lo v_add_co_u32 v4, vcc_lo, s4, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo global_load_b32 v1, v[6:7], off s_waitcnt vmcnt(0) global_store_b32 v[4:5], v1, off s_cbranch_scc0 .LBB0_2 .LBB0_3: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z26merge_without_blend_kernelPfS_S_Piiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 304 .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 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_end0: .size _Z26merge_without_blend_kernelPfS_S_Piiii, .Lfunc_end0-_Z26merge_without_blend_kernelPfS_S_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 - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: hidden_block_count_x - .offset: 52 .size: 4 .value_kind: hidden_block_count_y - .offset: 56 .size: 4 .value_kind: hidden_block_count_z - .offset: 60 .size: 2 .value_kind: hidden_group_size_x - .offset: 62 .size: 2 .value_kind: hidden_group_size_y - .offset: 64 .size: 2 .value_kind: hidden_group_size_z - .offset: 66 .size: 2 .value_kind: hidden_remainder_x - .offset: 68 .size: 2 .value_kind: hidden_remainder_y - .offset: 70 .size: 2 .value_kind: hidden_remainder_z - .offset: 88 .size: 8 .value_kind: hidden_global_offset_x - .offset: 96 .size: 8 .value_kind: hidden_global_offset_y - .offset: 104 .size: 8 .value_kind: hidden_global_offset_z - .offset: 112 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 304 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z26merge_without_blend_kernelPfS_S_Piiii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z26merge_without_blend_kernelPfS_S_Piiii.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_000d0fd9_00000000-6_merge_without_blend_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 _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii .type _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii, @function _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii: .LFB2051: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movq %rcx, 16(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 16(%rsp), %rax movq %rax, 136(%rsp) leaq 12(%rsp), %rax movq %rax, 144(%rsp) leaq 8(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z26merge_without_blend_kernelPfS_S_Piiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii, .-_Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii .globl _Z26merge_without_blend_kernelPfS_S_Piiii .type _Z26merge_without_blend_kernelPfS_S_Piiii, @function _Z26merge_without_blend_kernelPfS_S_Piiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z55__device_stub__Z26merge_without_blend_kernelPfS_S_PiiiiPfS_S_Piiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z26merge_without_blend_kernelPfS_S_Piiii, .-_Z26merge_without_blend_kernelPfS_S_Piiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z26merge_without_blend_kernelPfS_S_Piiii" .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 _Z26merge_without_blend_kernelPfS_S_Piiii(%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 "merge_without_blend_kernel.hip" .globl _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii # -- Begin function _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .p2align 4, 0x90 .type _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii,@function _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii: # @_Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movl %r8d, 12(%rsp) movl %r9d, 8(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 12(%rsp), %rax movq %rax, 128(%rsp) leaq 8(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%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 $_Z26merge_without_blend_kernelPfS_S_Piiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii, .Lfunc_end0-_Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .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 $_Z26merge_without_blend_kernelPfS_S_Piiii, %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 _Z26merge_without_blend_kernelPfS_S_Piiii,@object # @_Z26merge_without_blend_kernelPfS_S_Piiii .section .rodata,"a",@progbits .globl _Z26merge_without_blend_kernelPfS_S_Piiii .p2align 3, 0x0 _Z26merge_without_blend_kernelPfS_S_Piiii: .quad _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .size _Z26merge_without_blend_kernelPfS_S_Piiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z26merge_without_blend_kernelPfS_S_Piiii" .size .L__unnamed_1, 42 .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 _Z41__device_stub__merge_without_blend_kernelPfS_S_Piiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z26merge_without_blend_kernelPfS_S_Piiii .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/* This file shows how to use the "gather" pattern in CUDA. This specific example is using the Black-Scholes equation, where we parallelize over securities, but gather parameters for each security. */ #include "cuda_runtime.h" #include "device_launch_parameters.h" #include "curand.h" // CUDA random number generators #include <math.h> // standard mathematical operations (log, erf, sqrt) #define _USE_MATH_DEFINES // common mathematical constants (`M_SQRT1_2`, used below, is equal to sqrt(1/2)) #include <cstdio> // cumulative distribution function (CDF) of a standard normal distribution // note the `__inline__`, which tells the compiler to just inline the function at compile time for // performance (at the cost of larger binary size) __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5 * erff(x * M_SQRT1_2); } // options are a right to buy (call) or sell (put) an asset at a specific price/date // k = strike price, s = underlying asset price, t = time until option expires, // r = rate at which money can be borrowed, v = volatility of option // c = call price, p = put price // this kernel actually does all the calculations for each security __device__ __host__ void price(float k, float s, float t, float r, float v, float* c, float* p) { float srt = v * sqrtf(t); float d1 = (logf(s/k) + (r+0.5*v*v)*t) / srt; float d2 = d1 - srt; float kert = k * expf(-r*t); *c = erff(d1)*s - erff(d2)*kert; *p = kert - s + *c; } // intermediate kernel which selects the index of the appropriate security and passes on // computation to the function defined above __global__ void price(float* k, float* s, float* t, float* r, float* v, float* c, float* p) { int idx = threadIdx.x; price(k[idx], s[idx], t[idx], r[idx], v[idx], &c[idx], &p[idx]); } int main() { const int count = 512; // number of securities to analyze const int size = count * sizeof(float); float *args[5]; // array of arrays of parameters for each security (k, s, t, r, v) // random generator for security parameters curandGenerator_t gen; curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_MTGP32); // generate values for the parameters directly on device for (int i=0; i<5; i++) { cudaMalloc(&args[i], size); curandGenerateUniform(gen, args[i], count); } float *dc, *dp; // call and put arrays cudaMalloc(&dc, size); // allocate space on device cudaMalloc(&dp, size); // just realized - `&dc` and `&dp` are pointers to pointers... passing them to `cudaMalloc` allows // the function to overwrite the pointers `dc` and `dp` to point at _new_ memory locations on the // device, rather than the original locations on the host they were pointed at when declared // calculate call/put values price<<<1,count>>>(args[0], args[1], args[2], args[3], args[4], dc, dp); // copy memory from device to host float hc[count], hp[count]; cudaMemcpy(hc, dc, size, cudaMemcpyDeviceToHost); cudaMemcpy(hp, dp, size, cudaMemcpyDeviceToHost); // free memory on device cudaFree(&dc); cudaFree(&dp); dc = NULL; dp = NULL; // make sure pointers can't be used or freed a second time for (int i=0; i<5; i++) { cudaFree(&args[i]); args[i] = NULL; } // print out some values for (int s=0; s<10; s++) { for (int i=0; i<5; i++) { printf("Call price: $%.2f, put price: $%.2f\n", hc[i], hp[i]); } } return 0; }
.file "tmpxft_000aff06_00000000-6_gather.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2062: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z5pricefffffPfS_ .type _Z5pricefffffPfS_, @function _Z5pricefffffPfS_: .LFB2058: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $40, %rsp .cfi_def_cfa_offset 64 movss %xmm0, 16(%rsp) movss %xmm1, 20(%rsp) movss %xmm2, 24(%rsp) movss %xmm3, 28(%rsp) movss %xmm4, 12(%rsp) movq %rdi, %rbp movq %rsi, %rbx pxor %xmm0, %xmm0 ucomiss %xmm2, %xmm0 ja .L8 sqrtss %xmm2, %xmm2 movaps %xmm2, %xmm0 .L6: mulss 12(%rsp), %xmm0 movss %xmm0, 8(%rsp) movss 20(%rsp), %xmm0 divss 16(%rsp), %xmm0 call logf@PLT movaps %xmm0, %xmm2 pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 movapd %xmm0, %xmm1 mulsd .LC1(%rip), %xmm1 mulsd %xmm0, %xmm1 movss 28(%rsp), %xmm3 pxor %xmm0, %xmm0 cvtss2sd %xmm3, %xmm0 addsd %xmm0, %xmm1 movss 24(%rsp), %xmm8 pxor %xmm0, %xmm0 cvtss2sd %xmm8, %xmm0 mulsd %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtss2sd %xmm2, %xmm0 addsd %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 divsd %xmm0, %xmm1 pxor %xmm6, %xmm6 cvtsd2ss %xmm1, %xmm6 movss %xmm6, 12(%rsp) xorps .LC2(%rip), %xmm3 movaps %xmm3, %xmm0 mulss %xmm8, %xmm0 call expf@PLT movaps %xmm0, %xmm6 mulss 16(%rsp), %xmm6 movss %xmm6, 24(%rsp) movss 12(%rsp), %xmm0 call erff@PLT movss %xmm0, 16(%rsp) movss 12(%rsp), %xmm4 subss 8(%rsp), %xmm4 movaps %xmm4, %xmm0 call erff@PLT movss 16(%rsp), %xmm1 movss 20(%rsp), %xmm5 mulss %xmm5, %xmm1 movss 24(%rsp), %xmm6 mulss %xmm6, %xmm0 subss %xmm0, %xmm1 movss %xmm1, 0(%rbp) subss %xmm5, %xmm6 movaps %xmm6, %xmm0 addss %xmm1, %xmm0 movss %xmm0, (%rbx) addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L8: .cfi_restore_state movss 24(%rsp), %xmm0 call sqrtf@PLT jmp .L6 .cfi_endproc .LFE2058: .size _Z5pricefffffPfS_, .-_Z5pricefffffPfS_ .globl _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ .type _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_, @function _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_: .LFB2084: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movq %rcx, 32(%rsp) movq %r8, 24(%rsp) movq %r9, 16(%rsp) movq 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 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) 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 .L14 .L10: movq 184(%rsp), %rax subq %fs:40, %rax jne .L15 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .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 _Z5pricePfS_S_S_S_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L10 .L15: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_, .-_Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ .globl _Z5pricePfS_S_S_S_S_S_ .type _Z5pricePfS_S_S_S_S_S_, @function _Z5pricePfS_S_S_S_S_S_: .LFB2085: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z5pricePfS_S_S_S_S_S_, .-_Z5pricePfS_S_S_S_S_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "Call price: $%.2f, put price: $%.2f\n" .text .globl main .type main, @function main: .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 subq $4096, %rsp .cfi_def_cfa_offset 4128 orq $0, (%rsp) subq $112, %rsp .cfi_def_cfa_offset 4240 movq %fs:40, %rax movq %rax, 4200(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $141, %esi call curandCreateGenerator@PLT leaq 48(%rsp), %rbx leaq 88(%rsp), %r12 movq %rbx, %rbp .L19: movl $2048, %esi movq %rbp, %rdi call cudaMalloc@PLT movq 0(%rbp), %rsi movl $512, %edx movq (%rsp), %rdi call curandGenerateUniform@PLT addq $8, %rbp cmpq %r12, %rbp jne .L19 leaq 8(%rsp), %rdi movl $2048, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $2048, %esi call cudaMalloc@PLT movl $512, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 36(%rsp), %rdx movl $1, %ecx movq 24(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L30 .L20: leaq 96(%rsp), %rdi movl $2, %ecx movl $2048, %edx movq 8(%rsp), %rsi call cudaMemcpy@PLT leaq 2144(%rsp), %rdi movl $2, %ecx movl $2048, %edx movq 16(%rsp), %rsi call cudaMemcpy@PLT leaq 8(%rsp), %rdi call cudaFree@PLT leaq 16(%rsp), %rdi call cudaFree@PLT movq $0, 8(%rsp) movq $0, 16(%rsp) .L21: movq %rbx, %rdi call cudaFree@PLT movq $0, (%rbx) addq $8, %rbx cmpq %r12, %rbx jne .L21 movl $10, %r12d leaq .LC3(%rip), %rbp jmp .L22 .L30: subq $8, %rsp .cfi_def_cfa_offset 4248 pushq 24(%rsp) .cfi_def_cfa_offset 4256 movq 24(%rsp), %r9 movq 96(%rsp), %r8 movq 88(%rsp), %rcx movq 80(%rsp), %rdx movq 72(%rsp), %rsi movq 64(%rsp), %rdi call _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ addq $16, %rsp .cfi_def_cfa_offset 4240 jmp .L20 .L31: subl $1, %r12d je .L24 .L22: movl $0, %ebx .L23: pxor %xmm0, %xmm0 cvtss2sd 96(%rsp,%rbx), %xmm0 pxor %xmm1, %xmm1 cvtss2sd 2144(%rsp,%rbx), %xmm1 movq %rbp, %rsi movl $2, %edi movl $2, %eax call __printf_chk@PLT addq $4, %rbx cmpq $20, %rbx jne .L23 jmp .L31 .L24: movq 4200(%rsp), %rax subq %fs:40, %rax jne .L32 movl $0, %eax addq $4208, %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 .L32: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC4: .string "_Z5pricePfS_S_S_S_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z5pricePfS_S_S_S_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1071644672 .section .rodata.cst16,"aM",@progbits,16 .align 16 .LC2: .long -2147483648 .long 0 .long 0 .long 0 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
/* This file shows how to use the "gather" pattern in CUDA. This specific example is using the Black-Scholes equation, where we parallelize over securities, but gather parameters for each security. */ #include "cuda_runtime.h" #include "device_launch_parameters.h" #include "curand.h" // CUDA random number generators #include <math.h> // standard mathematical operations (log, erf, sqrt) #define _USE_MATH_DEFINES // common mathematical constants (`M_SQRT1_2`, used below, is equal to sqrt(1/2)) #include <cstdio> // cumulative distribution function (CDF) of a standard normal distribution // note the `__inline__`, which tells the compiler to just inline the function at compile time for // performance (at the cost of larger binary size) __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5 * erff(x * M_SQRT1_2); } // options are a right to buy (call) or sell (put) an asset at a specific price/date // k = strike price, s = underlying asset price, t = time until option expires, // r = rate at which money can be borrowed, v = volatility of option // c = call price, p = put price // this kernel actually does all the calculations for each security __device__ __host__ void price(float k, float s, float t, float r, float v, float* c, float* p) { float srt = v * sqrtf(t); float d1 = (logf(s/k) + (r+0.5*v*v)*t) / srt; float d2 = d1 - srt; float kert = k * expf(-r*t); *c = erff(d1)*s - erff(d2)*kert; *p = kert - s + *c; } // intermediate kernel which selects the index of the appropriate security and passes on // computation to the function defined above __global__ void price(float* k, float* s, float* t, float* r, float* v, float* c, float* p) { int idx = threadIdx.x; price(k[idx], s[idx], t[idx], r[idx], v[idx], &c[idx], &p[idx]); } int main() { const int count = 512; // number of securities to analyze const int size = count * sizeof(float); float *args[5]; // array of arrays of parameters for each security (k, s, t, r, v) // random generator for security parameters curandGenerator_t gen; curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_MTGP32); // generate values for the parameters directly on device for (int i=0; i<5; i++) { cudaMalloc(&args[i], size); curandGenerateUniform(gen, args[i], count); } float *dc, *dp; // call and put arrays cudaMalloc(&dc, size); // allocate space on device cudaMalloc(&dp, size); // just realized - `&dc` and `&dp` are pointers to pointers... passing them to `cudaMalloc` allows // the function to overwrite the pointers `dc` and `dp` to point at _new_ memory locations on the // device, rather than the original locations on the host they were pointed at when declared // calculate call/put values price<<<1,count>>>(args[0], args[1], args[2], args[3], args[4], dc, dp); // copy memory from device to host float hc[count], hp[count]; cudaMemcpy(hc, dc, size, cudaMemcpyDeviceToHost); cudaMemcpy(hp, dp, size, cudaMemcpyDeviceToHost); // free memory on device cudaFree(&dc); cudaFree(&dp); dc = NULL; dp = NULL; // make sure pointers can't be used or freed a second time for (int i=0; i<5; i++) { cudaFree(&args[i]); args[i] = NULL; } // print out some values for (int s=0; s<10; s++) { for (int i=0; i<5; i++) { printf("Call price: $%.2f, put price: $%.2f\n", hc[i], hp[i]); } } return 0; }
/* This file shows how to use the "gather" pattern in CUDA. This specific example is using the Black-Scholes equation, where we parallelize over securities, but gather parameters for each security. */ #include "hip/hip_runtime.h" #include "hiprand/hiprand.h" // CUDA random number generators #include <math.h> // standard mathematical operations (log, erf, sqrt) #define _USE_MATH_DEFINES // common mathematical constants (`M_SQRT1_2`, used below, is equal to sqrt(1/2)) #include <cstdio> // cumulative distribution function (CDF) of a standard normal distribution // note the `__inline__`, which tells the compiler to just inline the function at compile time for // performance (at the cost of larger binary size) __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5 * erff(x * M_SQRT1_2); } // options are a right to buy (call) or sell (put) an asset at a specific price/date // k = strike price, s = underlying asset price, t = time until option expires, // r = rate at which money can be borrowed, v = volatility of option // c = call price, p = put price // this kernel actually does all the calculations for each security __device__ __host__ void price(float k, float s, float t, float r, float v, float* c, float* p) { float srt = v * sqrtf(t); float d1 = (logf(s/k) + (r+0.5*v*v)*t) / srt; float d2 = d1 - srt; float kert = k * expf(-r*t); *c = erff(d1)*s - erff(d2)*kert; *p = kert - s + *c; } // intermediate kernel which selects the index of the appropriate security and passes on // computation to the function defined above __global__ void price(float* k, float* s, float* t, float* r, float* v, float* c, float* p) { int idx = threadIdx.x; price(k[idx], s[idx], t[idx], r[idx], v[idx], &c[idx], &p[idx]); } int main() { const int count = 512; // number of securities to analyze const int size = count * sizeof(float); float *args[5]; // array of arrays of parameters for each security (k, s, t, r, v) // random generator for security parameters hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_MTGP32); // generate values for the parameters directly on device for (int i=0; i<5; i++) { hipMalloc(&args[i], size); hiprandGenerateUniform(gen, args[i], count); } float *dc, *dp; // call and put arrays hipMalloc(&dc, size); // allocate space on device hipMalloc(&dp, size); // just realized - `&dc` and `&dp` are pointers to pointers... passing them to `cudaMalloc` allows // the function to overwrite the pointers `dc` and `dp` to point at _new_ memory locations on the // device, rather than the original locations on the host they were pointed at when declared // calculate call/put values price<<<1,count>>>(args[0], args[1], args[2], args[3], args[4], dc, dp); // copy memory from device to host float hc[count], hp[count]; hipMemcpy(hc, dc, size, hipMemcpyDeviceToHost); hipMemcpy(hp, dp, size, hipMemcpyDeviceToHost); // free memory on device hipFree(&dc); hipFree(&dp); dc = NULL; dp = NULL; // make sure pointers can't be used or freed a second time for (int i=0; i<5; i++) { hipFree(&args[i]); args[i] = NULL; } // print out some values for (int s=0; s<10; s++) { for (int i=0; i<5; i++) { printf("Call price: $%.2f, put price: $%.2f\n", hc[i], hp[i]); } } return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/* This file shows how to use the "gather" pattern in CUDA. This specific example is using the Black-Scholes equation, where we parallelize over securities, but gather parameters for each security. */ #include "hip/hip_runtime.h" #include "hiprand/hiprand.h" // CUDA random number generators #include <math.h> // standard mathematical operations (log, erf, sqrt) #define _USE_MATH_DEFINES // common mathematical constants (`M_SQRT1_2`, used below, is equal to sqrt(1/2)) #include <cstdio> // cumulative distribution function (CDF) of a standard normal distribution // note the `__inline__`, which tells the compiler to just inline the function at compile time for // performance (at the cost of larger binary size) __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5 * erff(x * M_SQRT1_2); } // options are a right to buy (call) or sell (put) an asset at a specific price/date // k = strike price, s = underlying asset price, t = time until option expires, // r = rate at which money can be borrowed, v = volatility of option // c = call price, p = put price // this kernel actually does all the calculations for each security __device__ __host__ void price(float k, float s, float t, float r, float v, float* c, float* p) { float srt = v * sqrtf(t); float d1 = (logf(s/k) + (r+0.5*v*v)*t) / srt; float d2 = d1 - srt; float kert = k * expf(-r*t); *c = erff(d1)*s - erff(d2)*kert; *p = kert - s + *c; } // intermediate kernel which selects the index of the appropriate security and passes on // computation to the function defined above __global__ void price(float* k, float* s, float* t, float* r, float* v, float* c, float* p) { int idx = threadIdx.x; price(k[idx], s[idx], t[idx], r[idx], v[idx], &c[idx], &p[idx]); } int main() { const int count = 512; // number of securities to analyze const int size = count * sizeof(float); float *args[5]; // array of arrays of parameters for each security (k, s, t, r, v) // random generator for security parameters hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_MTGP32); // generate values for the parameters directly on device for (int i=0; i<5; i++) { hipMalloc(&args[i], size); hiprandGenerateUniform(gen, args[i], count); } float *dc, *dp; // call and put arrays hipMalloc(&dc, size); // allocate space on device hipMalloc(&dp, size); // just realized - `&dc` and `&dp` are pointers to pointers... passing them to `cudaMalloc` allows // the function to overwrite the pointers `dc` and `dp` to point at _new_ memory locations on the // device, rather than the original locations on the host they were pointed at when declared // calculate call/put values price<<<1,count>>>(args[0], args[1], args[2], args[3], args[4], dc, dp); // copy memory from device to host float hc[count], hp[count]; hipMemcpy(hc, dc, size, hipMemcpyDeviceToHost); hipMemcpy(hp, dp, size, hipMemcpyDeviceToHost); // free memory on device hipFree(&dc); hipFree(&dp); dc = NULL; dp = NULL; // make sure pointers can't be used or freed a second time for (int i=0; i<5; i++) { hipFree(&args[i]); args[i] = NULL; } // print out some values for (int s=0; s<10; s++) { for (int i=0; i<5; i++) { printf("Call price: $%.2f, put price: $%.2f\n", hc[i], hp[i]); } } return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5pricePfS_S_S_S_S_S_ .globl _Z5pricePfS_S_S_S_S_S_ .p2align 8 .type _Z5pricePfS_S_S_S_S_S_,@function _Z5pricePfS_S_S_S_S_S_: s_load_b256 s[4:11], s[0:1], 0x0 v_lshlrev_b32_e32 v5, 2, v0 s_load_b64 s[2:3], s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_clause 0x4 global_load_b32 v1, v5, s[4:5] global_load_b32 v2, v5, s[6:7] global_load_b32 v3, v5, s[8:9] global_load_b32 v4, v5, s[10:11] global_load_b32 v13, v5, s[2:3] s_waitcnt vmcnt(2) v_cmp_gt_f32_e64 s2, 0xf800000, v3 v_div_scale_f32 v7, null, v1, v1, v2 v_div_scale_f32 v9, vcc_lo, v2, v1, v2 v_mul_f32_e32 v15, 0x4f800000, v3 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_rcp_f32_e32 v8, v7 v_cndmask_b32_e64 v15, v3, v15, s2 s_waitcnt_depctr 0xfff v_fma_f32 v5, -v7, v8, 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_fmac_f32_e32 v8, v5, v8 s_waitcnt vmcnt(0) v_cvt_f64_f32_e32 v[5:6], v13 v_mul_f32_e32 v10, v9, v8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v11, -v7, v10, v9 v_fmac_f32_e32 v10, v11, v8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v7, -v7, v10, v9 v_div_fmas_f32 v7, v7, v8, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fixup_f32 v7, v7, v1, v2 v_cmp_gt_f32_e32 vcc_lo, 0x800000, v7 v_mul_f64 v[9:10], v[5:6], 0.5 v_cndmask_b32_e64 v8, 1.0, 0x4f800000, vcc_lo v_cndmask_b32_e64 v16, 0, 0x41b17218, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f32_e32 v7, v7, v8 v_log_f32_e32 v11, v7 v_cvt_f64_f32_e32 v[7:8], v4 s_waitcnt_depctr 0xfff v_mul_f32_e32 v12, 0x3f317217, v11 v_cmp_gt_f32_e64 vcc_lo, 0x7f800000, |v11| s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v14, v11, 0x3f317217, -v12 v_fmamk_f32 v14, v11, 0x3377d1cf, v14 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_f32_e32 v12, v12, v14 v_sqrt_f32_e32 v14, v15 s_waitcnt_depctr 0xfff v_dual_cndmask_b32 v11, v11, v12 :: v_dual_add_nc_u32 v18, 1, v14 v_dual_sub_f32 v16, v11, v16 :: v_dual_add_nc_u32 v17, -1, v14 v_fma_f64 v[5:6], v[9:10], v[5:6], v[7:8] v_cvt_f64_f32_e32 v[11:12], v3 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_fma_f32 v10, -v18, v14, v15 v_fma_f32 v7, -v17, v14, v15 v_cvt_f64_f32_e32 v[8:9], v16 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_cmp_ge_f32_e32 vcc_lo, 0, v7 v_cndmask_b32_e32 v7, v14, v17, vcc_lo v_cmp_lt_f32_e32 vcc_lo, 0, v10 v_cndmask_b32_e32 v7, v7, v18, vcc_lo v_cmp_class_f32_e64 vcc_lo, v15, 0x260 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f32_e32 v10, 0x37800000, v7 v_cndmask_b32_e64 v7, v7, v10, s2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e32 v7, v7, v15, vcc_lo v_mul_f32_e32 v7, v7, v13 v_fma_f64 v[5:6], v[5:6], v[11:12], v[8:9] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_f64_f32_e32 v[13:14], v7 v_div_scale_f64 v[8:9], null, v[13:14], v[13:14], v[5:6] s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f64_e32 v[10:11], v[8:9] s_waitcnt_depctr 0xfff v_fma_f64 v[15:16], -v[8:9], v[10:11], 1.0 v_fma_f64 v[10:11], v[10:11], v[15:16], v[10:11] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[15:16], -v[8:9], v[10:11], 1.0 v_fma_f64 v[10:11], v[10:11], v[15:16], v[10:11] v_div_scale_f64 v[15:16], vcc_lo, v[5:6], v[13:14], v[5:6] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[17:18], v[15:16], v[10:11] v_fma_f64 v[8:9], -v[8:9], v[17:18], v[15:16] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fmas_f64 v[8:9], v[8:9], v[10:11], v[17:18] v_div_fixup_f64 v[5:6], v[8:9], v[13:14], v[5:6] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_f32_f64_e32 v5, v[5:6] v_cmp_nlt_f32_e64 s2, |v5|, 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_xor_b32 s2, exec_lo, s3 s_cbranch_execz .LBB0_2 s_mov_b32 s3, 0x378e98ab s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, |v5|, s3, 0xb9c68948 v_fma_f32 v6, |v5|, v6, 0x3b7cd369 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, |v5|, v6, 0xbcc618b2 v_fma_f32 v6, |v5|, v6, 0x3dda74e4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, |v5|, v6, 0x3f228afd v_fma_f32 v6, |v5|, v6, 0x3e03c728 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, |v5|, v6, |v5| v_mul_f32_e32 v8, 0xbfb8aa3b, v6 v_cmp_nlt_f32_e32 vcc_lo, 0x42ce8ed0, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_fma_f32 v9, v6, 0xbfb8aa3b, -v8 v_rndne_f32_e32 v10, v8 v_dual_fmamk_f32 v9, v6, 0xb2a5705f, v9 :: v_dual_sub_f32 v8, v8, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_add_f32_e32 v8, v8, v9 v_cvt_i32_f32_e32 v9, v10 v_exp_f32_e32 v8, v8 s_waitcnt_depctr 0xfff v_ldexp_f32 v8, v8, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v8, 0, v8, vcc_lo v_cmp_ngt_f32_e32 vcc_lo, 0xc2b17218, v6 v_cndmask_b32_e32 v6, 0x7f800000, v8, vcc_lo s_delay_alu instid0(VALU_DEP_1) v_sub_f32_e32 v6, 1.0, v6 .LBB0_2: s_and_not1_saveexec_b32 s2, s2 v_mul_f32_e32 v6, v5, v5 s_mov_b32 s3, 0xba1345e1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fmaak_f32 v8, s3, v6, 0x3ba10414 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmaak_f32 v8, v6, v8, 0xbcdac9b8 v_fmaak_f32 v8, v6, v8, 0x3de703be s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmaak_f32 v8, v6, v8, 0xbec09330 v_fmaak_f32 v6, v6, v8, 0x3e0375d0 s_delay_alu instid0(VALU_DEP_1) v_fma_f32 v6, |v5|, v6, |v5| s_or_b32 exec_lo, exec_lo, s2 v_sub_f32_e32 v7, v5, v7 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_nlt_f32_e64 s2, |v7|, 1.0 s_and_saveexec_b32 s3, s2 s_delay_alu instid0(SALU_CYCLE_1) s_xor_b32 s2, exec_lo, s3 s_cbranch_execz .LBB0_6 s_mov_b32 s3, 0x378e98ab s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v8, |v7|, s3, 0xb9c68948 v_fma_f32 v8, |v7|, v8, 0x3b7cd369 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v8, |v7|, v8, 0xbcc618b2 v_fma_f32 v8, |v7|, v8, 0x3dda74e4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v8, |v7|, v8, 0x3f228afd v_fma_f32 v8, |v7|, v8, 0x3e03c728 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v8, |v7|, v8, |v7| v_mul_f32_e32 v9, 0xbfb8aa3b, v8 v_cmp_nlt_f32_e32 vcc_lo, 0x42ce8ed0, v8 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_fma_f32 v10, v8, 0xbfb8aa3b, -v9 v_rndne_f32_e32 v11, v9 v_dual_fmamk_f32 v10, v8, 0xb2a5705f, v10 :: v_dual_sub_f32 v9, v9, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_add_f32_e32 v9, v9, v10 v_cvt_i32_f32_e32 v10, v11 v_exp_f32_e32 v9, v9 s_waitcnt_depctr 0xfff v_ldexp_f32 v9, v9, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v9, 0, v9, vcc_lo v_cmp_ngt_f32_e32 vcc_lo, 0xc2b17218, v8 v_cndmask_b32_e32 v8, 0x7f800000, v9, vcc_lo s_delay_alu instid0(VALU_DEP_1) v_sub_f32_e32 v8, 1.0, v8 .LBB0_6: s_and_not1_saveexec_b32 s2, s2 v_mul_f32_e32 v8, v7, v7 s_mov_b32 s3, 0xba1345e1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_fmaak_f32 v9, s3, v8, 0x3ba10414 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmaak_f32 v9, v8, v9, 0xbcdac9b8 v_fmaak_f32 v9, v8, v9, 0x3de703be s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmaak_f32 v9, v8, v9, 0xbec09330 v_fmaak_f32 v8, v8, v9, 0x3e0375d0 s_delay_alu instid0(VALU_DEP_1) v_fma_f32 v8, |v7|, v8, |v7| s_or_b32 exec_lo, exec_lo, s2 v_mul_f32_e64 v3, v3, -v4 s_load_b128 s[0:3], s[0:1], 0x28 v_bfi_b32 v5, 0x7fffffff, v6, v5 v_lshlrev_b32_e32 v0, 2, v0 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mul_f32_e32 v4, 0x3fb8aa3b, v3 v_cmp_ngt_f32_e32 vcc_lo, 0xc2ce8ed0, v3 v_fma_f32 v9, v3, 0x3fb8aa3b, -v4 v_rndne_f32_e32 v10, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_fmamk_f32 v9, v3, 0x32a5705f, v9 :: v_dual_sub_f32 v4, v4, v10 v_add_f32_e32 v4, v4, v9 v_cvt_i32_f32_e32 v9, v10 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_exp_f32_e32 v4, v4 s_waitcnt_depctr 0xfff v_ldexp_f32 v4, v4, v9 v_cndmask_b32_e32 v4, 0, v4, vcc_lo v_cmp_nlt_f32_e32 vcc_lo, 0x42b17218, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v3, 0x7f800000, v4, vcc_lo v_bfi_b32 v4, 0x7fffffff, v8, v7 v_mul_f32_e32 v7, v1, v3 v_fma_f32 v1, v1, v3, -v2 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f32_e32 v4, v7, v4 v_fma_f32 v4, v2, v5, -v4 s_delay_alu instid0(VALU_DEP_1) v_add_f32_e32 v1, v1, v4 s_waitcnt lgkmcnt(0) s_clause 0x1 global_store_b32 v0, v4, s[0:1] global_store_b32 v0, v1, s[2:3] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5pricePfS_S_S_S_S_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 56 .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 19 .amdhsa_next_free_sgpr 12 .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 _Z5pricePfS_S_S_S_S_S_, .Lfunc_end0-_Z5pricePfS_S_S_S_S_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 48 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 56 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5pricePfS_S_S_S_S_S_ .private_segment_fixed_size: 0 .sgpr_count: 14 .sgpr_spill_count: 0 .symbol: _Z5pricePfS_S_S_S_S_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 19 .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.
/* This file shows how to use the "gather" pattern in CUDA. This specific example is using the Black-Scholes equation, where we parallelize over securities, but gather parameters for each security. */ #include "hip/hip_runtime.h" #include "hiprand/hiprand.h" // CUDA random number generators #include <math.h> // standard mathematical operations (log, erf, sqrt) #define _USE_MATH_DEFINES // common mathematical constants (`M_SQRT1_2`, used below, is equal to sqrt(1/2)) #include <cstdio> // cumulative distribution function (CDF) of a standard normal distribution // note the `__inline__`, which tells the compiler to just inline the function at compile time for // performance (at the cost of larger binary size) __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5 * erff(x * M_SQRT1_2); } // options are a right to buy (call) or sell (put) an asset at a specific price/date // k = strike price, s = underlying asset price, t = time until option expires, // r = rate at which money can be borrowed, v = volatility of option // c = call price, p = put price // this kernel actually does all the calculations for each security __device__ __host__ void price(float k, float s, float t, float r, float v, float* c, float* p) { float srt = v * sqrtf(t); float d1 = (logf(s/k) + (r+0.5*v*v)*t) / srt; float d2 = d1 - srt; float kert = k * expf(-r*t); *c = erff(d1)*s - erff(d2)*kert; *p = kert - s + *c; } // intermediate kernel which selects the index of the appropriate security and passes on // computation to the function defined above __global__ void price(float* k, float* s, float* t, float* r, float* v, float* c, float* p) { int idx = threadIdx.x; price(k[idx], s[idx], t[idx], r[idx], v[idx], &c[idx], &p[idx]); } int main() { const int count = 512; // number of securities to analyze const int size = count * sizeof(float); float *args[5]; // array of arrays of parameters for each security (k, s, t, r, v) // random generator for security parameters hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_MTGP32); // generate values for the parameters directly on device for (int i=0; i<5; i++) { hipMalloc(&args[i], size); hiprandGenerateUniform(gen, args[i], count); } float *dc, *dp; // call and put arrays hipMalloc(&dc, size); // allocate space on device hipMalloc(&dp, size); // just realized - `&dc` and `&dp` are pointers to pointers... passing them to `cudaMalloc` allows // the function to overwrite the pointers `dc` and `dp` to point at _new_ memory locations on the // device, rather than the original locations on the host they were pointed at when declared // calculate call/put values price<<<1,count>>>(args[0], args[1], args[2], args[3], args[4], dc, dp); // copy memory from device to host float hc[count], hp[count]; hipMemcpy(hc, dc, size, hipMemcpyDeviceToHost); hipMemcpy(hp, dp, size, hipMemcpyDeviceToHost); // free memory on device hipFree(&dc); hipFree(&dp); dc = NULL; dp = NULL; // make sure pointers can't be used or freed a second time for (int i=0; i<5; i++) { hipFree(&args[i]); args[i] = NULL; } // print out some values for (int s=0; s<10; s++) { for (int i=0; i<5; i++) { printf("Call price: $%.2f, put price: $%.2f\n", hc[i], hp[i]); } } return 0; }
.text .file "gather.hip" .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z5pricefffffPfS_ .LCPI0_0: .quad 0x3fe0000000000000 # double 0.5 .section .rodata.cst16,"aM",@progbits,16 .p2align 4, 0x0 .LCPI0_1: .long 0x80000000 # float -0 .long 0x80000000 # float -0 .long 0x80000000 # float -0 .long 0x80000000 # float -0 .text .globl _Z5pricefffffPfS_ .p2align 4, 0x90 .type _Z5pricefffffPfS_,@function _Z5pricefffffPfS_: # @_Z5pricefffffPfS_ .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $56, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movq %rsi, %rbx movq %rdi, %r14 movss %xmm4, 8(%rsp) # 4-byte Spill movaps %xmm3, 32(%rsp) # 16-byte Spill movss %xmm1, 16(%rsp) # 4-byte Spill movss %xmm0, 12(%rsp) # 4-byte Spill xorps %xmm0, %xmm0 ucomiss %xmm0, %xmm2 movss %xmm2, 28(%rsp) # 4-byte Spill jb .LBB0_2 # %bb.1: xorps %xmm0, %xmm0 sqrtss %xmm2, %xmm0 jmp .LBB0_3 .LBB0_2: # %call.sqrt movaps %xmm2, %xmm0 callq sqrtf .LBB0_3: # %.split mulss 8(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 24(%rsp) # 4-byte Spill movss 16(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero divss 12(%rsp), %xmm0 # 4-byte Folded Reload callq logf xorps %xmm4, %xmm4 cvtss2sd %xmm0, %xmm4 movaps 32(%rsp), %xmm0 # 16-byte Reload xorps %xmm1, %xmm1 cvtss2sd %xmm0, %xmm1 movss 8(%rsp), %xmm2 # 4-byte Reload # xmm2 = mem[0],zero,zero,zero cvtss2sd %xmm2, %xmm2 movsd .LCPI0_0(%rip), %xmm3 # xmm3 = mem[0],zero mulsd %xmm2, %xmm3 mulsd %xmm2, %xmm3 addsd %xmm1, %xmm3 movss 28(%rsp), %xmm5 # 4-byte Reload # xmm5 = mem[0],zero,zero,zero xorps %xmm1, %xmm1 cvtss2sd %xmm5, %xmm1 mulsd %xmm3, %xmm1 addsd %xmm4, %xmm1 movss 24(%rsp), %xmm3 # 4-byte Reload # xmm3 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm3, %xmm2 divsd %xmm2, %xmm1 cvtsd2ss %xmm1, %xmm1 movss %xmm1, 20(%rsp) # 4-byte Spill subss %xmm3, %xmm1 movss %xmm1, 8(%rsp) # 4-byte Spill xorps .LCPI0_1(%rip), %xmm0 mulss %xmm5, %xmm0 callq expf mulss 12(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 12(%rsp) # 4-byte Spill movss 20(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero callq erff mulss 16(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 32(%rsp) # 4-byte Spill movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero callq erff movss 12(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero mulss %xmm1, %xmm0 movss 32(%rsp), %xmm2 # 4-byte Reload # xmm2 = mem[0],zero,zero,zero subss %xmm0, %xmm2 movss %xmm2, (%r14) subss 16(%rsp), %xmm1 # 4-byte Folded Reload addss %xmm2, %xmm1 movss %xmm1, (%rbx) addq $56, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z5pricefffffPfS_, .Lfunc_end0-_Z5pricefffffPfS_ .cfi_endproc # -- End function .globl _Z20__device_stub__pricePfS_S_S_S_S_S_ # -- Begin function _Z20__device_stub__pricePfS_S_S_S_S_S_ .p2align 4, 0x90 .type _Z20__device_stub__pricePfS_S_S_S_S_S_,@function _Z20__device_stub__pricePfS_S_S_S_S_S_: # @_Z20__device_stub__pricePfS_S_S_S_S_S_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z5pricePfS_S_S_S_S_S_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end1: .size _Z20__device_stub__pricePfS_S_S_S_S_S_, .Lfunc_end1-_Z20__device_stub__pricePfS_S_S_S_S_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $4264, %rsp # imm = 0x10A8 .cfi_def_cfa_offset 4288 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 16(%rsp), %rdi movl $403, %esi # imm = 0x193 callq hiprandCreateGenerator xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_1: # =>This Inner Loop Header: Depth=1 leaq (%rsp,%rbx), %rdi addq $112, %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movq 16(%rsp), %rdi movq 112(%rsp,%rbx), %rsi movl $512, %edx # imm = 0x200 callq hiprandGenerateUniform addq $8, %rbx cmpq $40, %rbx jne .LBB2_1 # %bb.2: leaq 8(%rsp), %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movq %rsp, %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movabsq $4294967297, %rdi # imm = 0x100000001 leaq 511(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 112(%rsp), %rax movq 120(%rsp), %rcx movq 128(%rsp), %rdx movq 136(%rsp), %rsi movq 144(%rsp), %rdi movq 8(%rsp), %r8 movq (%rsp), %r9 movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movq %rsi, 80(%rsp) movq %rdi, 72(%rsp) movq %r8, 64(%rsp) movq %r9, 56(%rsp) leaq 104(%rsp), %rax movq %rax, 160(%rsp) leaq 96(%rsp), %rax movq %rax, 168(%rsp) leaq 88(%rsp), %rax movq %rax, 176(%rsp) leaq 80(%rsp), %rax movq %rax, 184(%rsp) leaq 72(%rsp), %rax movq %rax, 192(%rsp) leaq 64(%rsp), %rax movq %rax, 200(%rsp) leaq 56(%rsp), %rax movq %rax, 208(%rsp) leaq 2208(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 2208(%rsp), %rsi movl 2216(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 160(%rsp), %r9 movl $_Z5pricePfS_S_S_S_S_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_4: movq 8(%rsp), %rsi leaq 160(%rsp), %rdi movl $2048, %edx # imm = 0x800 movl $2, %ecx callq hipMemcpy movq (%rsp), %rsi leaq 2208(%rsp), %rdi movl $2048, %edx # imm = 0x800 movl $2, %ecx callq hipMemcpy leaq 8(%rsp), %rdi callq hipFree movq %rsp, %rdi callq hipFree movq $0, 8(%rsp) movq $0, (%rsp) xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_5: # =>This Inner Loop Header: Depth=1 leaq (%rsp,%rbx), %rdi addq $112, %rdi callq hipFree movq $0, 112(%rsp,%rbx) addq $8, %rbx cmpq $40, %rbx jne .LBB2_5 # %bb.6: # %.preheader.preheader xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_7: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB2_8 Depth 2 xorl %r14d, %r14d .p2align 4, 0x90 .LBB2_8: # Parent Loop BB2_7 Depth=1 # => This Inner Loop Header: Depth=2 movss 160(%rsp,%r14,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movss 2208(%rsp,%r14,4), %xmm1 # xmm1 = mem[0],zero,zero,zero cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi movb $2, %al callq printf incq %r14 cmpq $5, %r14 jne .LBB2_8 # %bb.9: # in Loop: Header=BB2_7 Depth=1 incl %ebx cmpl $10, %ebx jne .LBB2_7 # %bb.10: xorl %eax, %eax addq $4264, %rsp # imm = 0x10A8 .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5pricePfS_S_S_S_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_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 _Z5pricePfS_S_S_S_S_S_,@object # @_Z5pricePfS_S_S_S_S_S_ .section .rodata,"a",@progbits .globl _Z5pricePfS_S_S_S_S_S_ .p2align 3, 0x0 _Z5pricePfS_S_S_S_S_S_: .quad _Z20__device_stub__pricePfS_S_S_S_S_S_ .size _Z5pricePfS_S_S_S_S_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Call price: $%.2f, put price: $%.2f\n" .size .L.str, 37 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5pricePfS_S_S_S_S_S_" .size .L__unnamed_1, 23 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__pricePfS_S_S_S_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5pricePfS_S_S_S_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000aff06_00000000-6_gather.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2062: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2062: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z5pricefffffPfS_ .type _Z5pricefffffPfS_, @function _Z5pricefffffPfS_: .LFB2058: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 subq $40, %rsp .cfi_def_cfa_offset 64 movss %xmm0, 16(%rsp) movss %xmm1, 20(%rsp) movss %xmm2, 24(%rsp) movss %xmm3, 28(%rsp) movss %xmm4, 12(%rsp) movq %rdi, %rbp movq %rsi, %rbx pxor %xmm0, %xmm0 ucomiss %xmm2, %xmm0 ja .L8 sqrtss %xmm2, %xmm2 movaps %xmm2, %xmm0 .L6: mulss 12(%rsp), %xmm0 movss %xmm0, 8(%rsp) movss 20(%rsp), %xmm0 divss 16(%rsp), %xmm0 call logf@PLT movaps %xmm0, %xmm2 pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 movapd %xmm0, %xmm1 mulsd .LC1(%rip), %xmm1 mulsd %xmm0, %xmm1 movss 28(%rsp), %xmm3 pxor %xmm0, %xmm0 cvtss2sd %xmm3, %xmm0 addsd %xmm0, %xmm1 movss 24(%rsp), %xmm8 pxor %xmm0, %xmm0 cvtss2sd %xmm8, %xmm0 mulsd %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtss2sd %xmm2, %xmm0 addsd %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 divsd %xmm0, %xmm1 pxor %xmm6, %xmm6 cvtsd2ss %xmm1, %xmm6 movss %xmm6, 12(%rsp) xorps .LC2(%rip), %xmm3 movaps %xmm3, %xmm0 mulss %xmm8, %xmm0 call expf@PLT movaps %xmm0, %xmm6 mulss 16(%rsp), %xmm6 movss %xmm6, 24(%rsp) movss 12(%rsp), %xmm0 call erff@PLT movss %xmm0, 16(%rsp) movss 12(%rsp), %xmm4 subss 8(%rsp), %xmm4 movaps %xmm4, %xmm0 call erff@PLT movss 16(%rsp), %xmm1 movss 20(%rsp), %xmm5 mulss %xmm5, %xmm1 movss 24(%rsp), %xmm6 mulss %xmm6, %xmm0 subss %xmm0, %xmm1 movss %xmm1, 0(%rbp) subss %xmm5, %xmm6 movaps %xmm6, %xmm0 addss %xmm1, %xmm0 movss %xmm0, (%rbx) addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L8: .cfi_restore_state movss 24(%rsp), %xmm0 call sqrtf@PLT jmp .L6 .cfi_endproc .LFE2058: .size _Z5pricefffffPfS_, .-_Z5pricefffffPfS_ .globl _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ .type _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_, @function _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_: .LFB2084: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) movq %rdx, 40(%rsp) movq %rcx, 32(%rsp) movq %r8, 24(%rsp) movq %r9, 16(%rsp) movq 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 48(%rsp), %rax movq %rax, 136(%rsp) leaq 40(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rax movq %rax, 152(%rsp) leaq 24(%rsp), %rax movq %rax, 160(%rsp) leaq 16(%rsp), %rax movq %rax, 168(%rsp) leaq 8(%rsp), %rax movq %rax, 176(%rsp) 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 .L14 .L10: movq 184(%rsp), %rax subq %fs:40, %rax jne .L15 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .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 _Z5pricePfS_S_S_S_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L10 .L15: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_, .-_Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ .globl _Z5pricePfS_S_S_S_S_S_ .type _Z5pricePfS_S_S_S_S_S_, @function _Z5pricePfS_S_S_S_S_S_: .LFB2085: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 pushq 24(%rsp) .cfi_def_cfa_offset 32 call _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z5pricePfS_S_S_S_S_S_, .-_Z5pricePfS_S_S_S_S_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "Call price: $%.2f, put price: $%.2f\n" .text .globl main .type main, @function main: .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 subq $4096, %rsp .cfi_def_cfa_offset 4128 orq $0, (%rsp) subq $112, %rsp .cfi_def_cfa_offset 4240 movq %fs:40, %rax movq %rax, 4200(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $141, %esi call curandCreateGenerator@PLT leaq 48(%rsp), %rbx leaq 88(%rsp), %r12 movq %rbx, %rbp .L19: movl $2048, %esi movq %rbp, %rdi call cudaMalloc@PLT movq 0(%rbp), %rsi movl $512, %edx movq (%rsp), %rdi call curandGenerateUniform@PLT addq $8, %rbp cmpq %r12, %rbp jne .L19 leaq 8(%rsp), %rdi movl $2048, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $2048, %esi call cudaMalloc@PLT movl $512, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 36(%rsp), %rdx movl $1, %ecx movq 24(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L30 .L20: leaq 96(%rsp), %rdi movl $2, %ecx movl $2048, %edx movq 8(%rsp), %rsi call cudaMemcpy@PLT leaq 2144(%rsp), %rdi movl $2, %ecx movl $2048, %edx movq 16(%rsp), %rsi call cudaMemcpy@PLT leaq 8(%rsp), %rdi call cudaFree@PLT leaq 16(%rsp), %rdi call cudaFree@PLT movq $0, 8(%rsp) movq $0, 16(%rsp) .L21: movq %rbx, %rdi call cudaFree@PLT movq $0, (%rbx) addq $8, %rbx cmpq %r12, %rbx jne .L21 movl $10, %r12d leaq .LC3(%rip), %rbp jmp .L22 .L30: subq $8, %rsp .cfi_def_cfa_offset 4248 pushq 24(%rsp) .cfi_def_cfa_offset 4256 movq 24(%rsp), %r9 movq 96(%rsp), %r8 movq 88(%rsp), %rcx movq 80(%rsp), %rdx movq 72(%rsp), %rsi movq 64(%rsp), %rdi call _Z36__device_stub__Z5pricePfS_S_S_S_S_S_PfS_S_S_S_S_S_ addq $16, %rsp .cfi_def_cfa_offset 4240 jmp .L20 .L31: subl $1, %r12d je .L24 .L22: movl $0, %ebx .L23: pxor %xmm0, %xmm0 cvtss2sd 96(%rsp,%rbx), %xmm0 pxor %xmm1, %xmm1 cvtss2sd 2144(%rsp,%rbx), %xmm1 movq %rbp, %rsi movl $2, %edi movl $2, %eax call __printf_chk@PLT addq $4, %rbx cmpq $20, %rbx jne .L23 jmp .L31 .L24: movq 4200(%rsp), %rax subq %fs:40, %rax jne .L32 movl $0, %eax addq $4208, %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 .L32: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC4: .string "_Z5pricePfS_S_S_S_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z5pricePfS_S_S_S_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1071644672 .section .rodata.cst16,"aM",@progbits,16 .align 16 .LC2: .long -2147483648 .long 0 .long 0 .long 0 .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 "gather.hip" .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z5pricefffffPfS_ .LCPI0_0: .quad 0x3fe0000000000000 # double 0.5 .section .rodata.cst16,"aM",@progbits,16 .p2align 4, 0x0 .LCPI0_1: .long 0x80000000 # float -0 .long 0x80000000 # float -0 .long 0x80000000 # float -0 .long 0x80000000 # float -0 .text .globl _Z5pricefffffPfS_ .p2align 4, 0x90 .type _Z5pricefffffPfS_,@function _Z5pricefffffPfS_: # @_Z5pricefffffPfS_ .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $56, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movq %rsi, %rbx movq %rdi, %r14 movss %xmm4, 8(%rsp) # 4-byte Spill movaps %xmm3, 32(%rsp) # 16-byte Spill movss %xmm1, 16(%rsp) # 4-byte Spill movss %xmm0, 12(%rsp) # 4-byte Spill xorps %xmm0, %xmm0 ucomiss %xmm0, %xmm2 movss %xmm2, 28(%rsp) # 4-byte Spill jb .LBB0_2 # %bb.1: xorps %xmm0, %xmm0 sqrtss %xmm2, %xmm0 jmp .LBB0_3 .LBB0_2: # %call.sqrt movaps %xmm2, %xmm0 callq sqrtf .LBB0_3: # %.split mulss 8(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 24(%rsp) # 4-byte Spill movss 16(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero divss 12(%rsp), %xmm0 # 4-byte Folded Reload callq logf xorps %xmm4, %xmm4 cvtss2sd %xmm0, %xmm4 movaps 32(%rsp), %xmm0 # 16-byte Reload xorps %xmm1, %xmm1 cvtss2sd %xmm0, %xmm1 movss 8(%rsp), %xmm2 # 4-byte Reload # xmm2 = mem[0],zero,zero,zero cvtss2sd %xmm2, %xmm2 movsd .LCPI0_0(%rip), %xmm3 # xmm3 = mem[0],zero mulsd %xmm2, %xmm3 mulsd %xmm2, %xmm3 addsd %xmm1, %xmm3 movss 28(%rsp), %xmm5 # 4-byte Reload # xmm5 = mem[0],zero,zero,zero xorps %xmm1, %xmm1 cvtss2sd %xmm5, %xmm1 mulsd %xmm3, %xmm1 addsd %xmm4, %xmm1 movss 24(%rsp), %xmm3 # 4-byte Reload # xmm3 = mem[0],zero,zero,zero xorps %xmm2, %xmm2 cvtss2sd %xmm3, %xmm2 divsd %xmm2, %xmm1 cvtsd2ss %xmm1, %xmm1 movss %xmm1, 20(%rsp) # 4-byte Spill subss %xmm3, %xmm1 movss %xmm1, 8(%rsp) # 4-byte Spill xorps .LCPI0_1(%rip), %xmm0 mulss %xmm5, %xmm0 callq expf mulss 12(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 12(%rsp) # 4-byte Spill movss 20(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero callq erff mulss 16(%rsp), %xmm0 # 4-byte Folded Reload movss %xmm0, 32(%rsp) # 4-byte Spill movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero callq erff movss 12(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero mulss %xmm1, %xmm0 movss 32(%rsp), %xmm2 # 4-byte Reload # xmm2 = mem[0],zero,zero,zero subss %xmm0, %xmm2 movss %xmm2, (%r14) subss 16(%rsp), %xmm1 # 4-byte Folded Reload addss %xmm2, %xmm1 movss %xmm1, (%rbx) addq $56, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z5pricefffffPfS_, .Lfunc_end0-_Z5pricefffffPfS_ .cfi_endproc # -- End function .globl _Z20__device_stub__pricePfS_S_S_S_S_S_ # -- Begin function _Z20__device_stub__pricePfS_S_S_S_S_S_ .p2align 4, 0x90 .type _Z20__device_stub__pricePfS_S_S_S_S_S_,@function _Z20__device_stub__pricePfS_S_S_S_S_S_: # @_Z20__device_stub__pricePfS_S_S_S_S_S_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movq %rcx, 64(%rsp) movq %r8, 56(%rsp) movq %r9, 48(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rax movq %rax, 120(%rsp) leaq 56(%rsp), %rax movq %rax, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z5pricePfS_S_S_S_S_S_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end1: .size _Z20__device_stub__pricePfS_S_S_S_S_S_, .Lfunc_end1-_Z20__device_stub__pricePfS_S_S_S_S_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $4264, %rsp # imm = 0x10A8 .cfi_def_cfa_offset 4288 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 16(%rsp), %rdi movl $403, %esi # imm = 0x193 callq hiprandCreateGenerator xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_1: # =>This Inner Loop Header: Depth=1 leaq (%rsp,%rbx), %rdi addq $112, %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movq 16(%rsp), %rdi movq 112(%rsp,%rbx), %rsi movl $512, %edx # imm = 0x200 callq hiprandGenerateUniform addq $8, %rbx cmpq $40, %rbx jne .LBB2_1 # %bb.2: leaq 8(%rsp), %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movq %rsp, %rdi movl $2048, %esi # imm = 0x800 callq hipMalloc movabsq $4294967297, %rdi # imm = 0x100000001 leaq 511(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_4 # %bb.3: movq 112(%rsp), %rax movq 120(%rsp), %rcx movq 128(%rsp), %rdx movq 136(%rsp), %rsi movq 144(%rsp), %rdi movq 8(%rsp), %r8 movq (%rsp), %r9 movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movq %rsi, 80(%rsp) movq %rdi, 72(%rsp) movq %r8, 64(%rsp) movq %r9, 56(%rsp) leaq 104(%rsp), %rax movq %rax, 160(%rsp) leaq 96(%rsp), %rax movq %rax, 168(%rsp) leaq 88(%rsp), %rax movq %rax, 176(%rsp) leaq 80(%rsp), %rax movq %rax, 184(%rsp) leaq 72(%rsp), %rax movq %rax, 192(%rsp) leaq 64(%rsp), %rax movq %rax, 200(%rsp) leaq 56(%rsp), %rax movq %rax, 208(%rsp) leaq 2208(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 2208(%rsp), %rsi movl 2216(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 160(%rsp), %r9 movl $_Z5pricePfS_S_S_S_S_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB2_4: movq 8(%rsp), %rsi leaq 160(%rsp), %rdi movl $2048, %edx # imm = 0x800 movl $2, %ecx callq hipMemcpy movq (%rsp), %rsi leaq 2208(%rsp), %rdi movl $2048, %edx # imm = 0x800 movl $2, %ecx callq hipMemcpy leaq 8(%rsp), %rdi callq hipFree movq %rsp, %rdi callq hipFree movq $0, 8(%rsp) movq $0, (%rsp) xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_5: # =>This Inner Loop Header: Depth=1 leaq (%rsp,%rbx), %rdi addq $112, %rdi callq hipFree movq $0, 112(%rsp,%rbx) addq $8, %rbx cmpq $40, %rbx jne .LBB2_5 # %bb.6: # %.preheader.preheader xorl %ebx, %ebx .p2align 4, 0x90 .LBB2_7: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB2_8 Depth 2 xorl %r14d, %r14d .p2align 4, 0x90 .LBB2_8: # Parent Loop BB2_7 Depth=1 # => This Inner Loop Header: Depth=2 movss 160(%rsp,%r14,4), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movss 2208(%rsp,%r14,4), %xmm1 # xmm1 = mem[0],zero,zero,zero cvtss2sd %xmm1, %xmm1 movl $.L.str, %edi movb $2, %al callq printf incq %r14 cmpq $5, %r14 jne .LBB2_8 # %bb.9: # in Loop: Header=BB2_7 Depth=1 incl %ebx cmpl $10, %ebx jne .LBB2_7 # %bb.10: xorl %eax, %eax addq $4264, %rsp # imm = 0x10A8 .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5pricePfS_S_S_S_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_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 _Z5pricePfS_S_S_S_S_S_,@object # @_Z5pricePfS_S_S_S_S_S_ .section .rodata,"a",@progbits .globl _Z5pricePfS_S_S_S_S_S_ .p2align 3, 0x0 _Z5pricePfS_S_S_S_S_S_: .quad _Z20__device_stub__pricePfS_S_S_S_S_S_ .size _Z5pricePfS_S_S_S_S_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Call price: $%.2f, put price: $%.2f\n" .size .L.str, 37 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5pricePfS_S_S_S_S_S_" .size .L__unnamed_1, 23 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__pricePfS_S_S_S_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5pricePfS_S_S_S_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
/*#include<stdio.h> #include<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/*#include<stdio.h> #include<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
.file "tmpxft_00037cdc_00000000-6_gpuArr.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<stdio.h> #include<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
#include <hip/hip_runtime.h> /*#include<stdio.h> #include<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
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<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> /*#include<stdio.h> #include<cuda.h> #include<cuda_runtime.h> #define N 40000; //ThreadsLaunched __global__ void gpuArrSet(long *d_in1, long *d_in2) { int M = N; long tID = blockIdx.x; if(tID<M) { d_in1[tID] = (-tID); d_in2[tID] = (2 * tID); } } __global__ void gpuArrAdd(long *d_in1, long *d_in2, long*d_out) { int M = N; long tID = blockIdx.x; if (tID < M) { d_out[tID] = d_in2[tID] + d_in1[tID]; } } int main() { long* d_in1; long* d_in2; long* d_out; int M = N; cudaMalloc((void**)&d_in1, sizeof(long)*M); cudaMalloc((void**)&d_in2, sizeof(long)*M); cudaMalloc((void**)&d_out, sizeof(long)*M); long* h_out =(long*) malloc(sizeof(long)*M); gpuArrSet << <M , 1>> >(d_in1,d_in2); gpuArrAdd << <M , 1>> >(d_in1, d_in2, d_out); cudaDeviceSynchronize(); cudaMemcpy(h_out, d_out, sizeof(int)*M, cudaMemcpyDeviceToHost); printf("Output: "); for (int i = 0; i < M; i++) { printf("%ld\n", h_out[i]); } free(h_out); cudaFree(d_in1); cudaFree(d_in2); cudaFree(d_out); system("Pause"); return 911; }*/
.text .file "gpuArr.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_00037cdc_00000000-6_gpuArr.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 "gpuArr.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 <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(cudaError_t code, const char *file, int line, bool abort=true){ if (code != cudaSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; cudaMalloc(&d_x, size); cudaMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } cudaMemcpy(d_x, h_x, size, cudaMemcpyHostToDevice); cudaMemcpy(d_y, h_y, size, cudaMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( cudaPeekAtLastError() ); // cudaDeviceSynchronize(); cudaMemcpy(h_y, d_y, size, cudaMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; cudaFree(d_x); cudaFree(d_y); free(h_x); free(h_y); return 0; }
code for sm_80 Function : _Z3addiPfS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x160], PT ; /* 0x0000580002007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0070*/ MOV R5, 0x40400000 ; /* 0x4040000000057802 */ /* 0x000fe20000000f00 */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R3, c[0x0][0x170] ; /* 0x00005c0002027625 */ /* 0x000fca00078e0203 */ /*00a0*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(cudaError_t code, const char *file, int line, bool abort=true){ if (code != cudaSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; cudaMalloc(&d_x, size); cudaMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } cudaMemcpy(d_x, h_x, size, cudaMemcpyHostToDevice); cudaMemcpy(d_y, h_y, size, cudaMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( cudaPeekAtLastError() ); // cudaDeviceSynchronize(); cudaMemcpy(h_y, d_y, size, cudaMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; cudaFree(d_x); cudaFree(d_y); free(h_x); free(h_y); return 0; }
.file "tmpxft_00077741_00000000-6_main.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3673: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3673: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z25__device_stub__Z3addiPfS_iPfS_ .type _Z25__device_stub__Z3addiPfS_iPfS_, @function _Z25__device_stub__Z3addiPfS_iPfS_: .LFB3695: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movl %edi, 28(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .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 _Z3addiPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3695: .size _Z25__device_stub__Z3addiPfS_iPfS_, .-_Z25__device_stub__Z3addiPfS_iPfS_ .globl _Z3addiPfS_ .type _Z3addiPfS_, @function _Z3addiPfS_: .LFB3696: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z3addiPfS_iPfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3696: .size _Z3addiPfS_, .-_Z3addiPfS_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "/home/ubuntu/Datasets/stackv2/train-structured/akhtyamovpavel/ParallelComputationExamples/master/CUDA/01-intro/04-memcpy/main.cu" .section .rodata.str1.1,"aMS",@progbits,1 .LC4: .string "GPUassert: %s %s %d\n" .LC7: .string "Max error: " .text .globl main .type main, @function main: .LFB3670: .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 $72, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $1073741824, %edi call malloc@PLT movq %rax, %r12 movl $1073741824, %edi call malloc@PLT movq %rax, %rbp leaq 16(%rsp), %rdi movl $1073741824, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $1073741824, %esi call cudaMalloc@PLT movl $0, %eax movss .LC1(%rip), %xmm1 movss .LC2(%rip), %xmm0 .L12: movss %xmm1, (%r12,%rax) movss %xmm0, 0(%rbp,%rax) addq $4, %rax cmpq $1073741824, %rax jne .L12 movl $1, %ecx movl $1073741824, %edx movq %r12, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $1073741824, %edx movq %rbp, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 44(%rsp) movl $1, 48(%rsp) movl $262144, 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 .L20 .L13: call cudaPeekAtLastError@PLT movl %eax, %ebx testl %eax, %eax jne .L21 movl $2, %ecx movl $1073741824, %edx movq 24(%rsp), %rsi movq %rbp, %rdi call cudaMemcpy@PLT movq %rbp, %rbx leaq 1073741824(%rbp), %r13 movl $0x00000000, 12(%rsp) .L15: movss (%rbx), %xmm0 subss .LC5(%rip), %xmm0 andps .LC6(%rip), %xmm0 movss 12(%rsp), %xmm1 call fmaxf@PLT movss %xmm0, 12(%rsp) addq $4, %rbx cmpq %r13, %rbx jne .L15 leaq .LC7(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L22 movl $0, %eax addq $72, %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 .L20: .cfi_restore_state movq 24(%rsp), %rdx movq 16(%rsp), %rsi movl $268435456, %edi call _Z25__device_stub__Z3addiPfS_iPfS_ jmp .L13 .L21: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rcx movl $50, %r9d leaq .LC3(%rip), %r8 leaq .LC4(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl %ebx, %edi call exit@PLT .L22: call __stack_chk_fail@PLT .cfi_endproc .LFE3670: .size main, .-main .section .rodata.str1.1 .LC8: .string "_Z3addiPfS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3698: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _Z3addiPfS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3698: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC1: .long 1065353216 .align 4 .LC2: .long 1073741824 .align 4 .LC5: .long 1077936128 .section .rodata.cst16,"aM",@progbits,16 .align 16 .LC6: .long 2147483647 .long 0 .long 0 .long 0 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(cudaError_t code, const char *file, int line, bool abort=true){ if (code != cudaSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; cudaMalloc(&d_x, size); cudaMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } cudaMemcpy(d_x, h_x, size, cudaMemcpyHostToDevice); cudaMemcpy(d_y, h_y, size, cudaMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( cudaPeekAtLastError() ); // cudaDeviceSynchronize(); cudaMemcpy(h_y, d_y, size, cudaMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; cudaFree(d_x); cudaFree(d_y); free(h_x); free(h_y); return 0; }
#include <hip/hip_runtime.h> #include <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(hipError_t code, const char *file, int line, bool abort=true){ if (code != hipSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", hipGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; hipMalloc(&d_x, size); hipMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } hipMemcpy(d_x, h_x, size, hipMemcpyHostToDevice); hipMemcpy(d_y, h_y, size, hipMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( hipPeekAtLastError() ); // cudaDeviceSynchronize(); hipMemcpy(h_y, d_y, size, hipMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; hipFree(d_x); hipFree(d_y); free(h_x); free(h_y); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(hipError_t code, const char *file, int line, bool abort=true){ if (code != hipSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", hipGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; hipMalloc(&d_x, size); hipMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } hipMemcpy(d_x, h_x, size, hipMemcpyHostToDevice); hipMemcpy(d_y, h_y, size, hipMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( hipPeekAtLastError() ); // cudaDeviceSynchronize(); hipMemcpy(h_y, d_y, size, hipMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; hipFree(d_x); hipFree(d_y); free(h_x); free(h_y); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addiPfS_ .globl _Z3addiPfS_ .p2align 8 .type _Z3addiPfS_,@function _Z3addiPfS_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b32 s3, s[0:1], 0x0 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_2 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_2) | instid1(VALU_DEP_2) v_lshlrev_b64 v[0:1], 2, v[1:2] v_mov_b32_e32 v2, 0x40400000 s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_3) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo 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 _Z3addiPfS_ .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 3 .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 _Z3addiPfS_, .Lfunc_end0-_Z3addiPfS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addiPfS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z3addiPfS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <iostream> #include <cmath> #include <cstdio> #define cudaErrchk(ans) { GPUAssert((ans), __FILE__, __LINE__); } inline void GPUAssert(hipError_t code, const char *file, int line, bool abort=true){ if (code != hipSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", hipGetErrorString(code), file, line); if (abort) exit(code); } } __global__ void add(int n, float* x, float* y) { int index = blockIdx.x * blockDim.x + threadIdx.x; if (index < n) { y[index] = 1 + 2; } } int main() { int N = 1 << 28; size_t size = N * sizeof(float); float *h_x = (float*)malloc(size); float *h_y = (float*)malloc(size); float *d_x, *d_y; hipMalloc(&d_x, size); hipMalloc(&d_y, size); for (int i = 0; i < N; ++i) { h_x[i] = 1.0f; h_y[i] = 2.0f; } hipMemcpy(d_x, h_x, size, hipMemcpyHostToDevice); hipMemcpy(d_y, h_y, size, hipMemcpyHostToDevice); int blockSize = 1024; int numBlocks = (N + blockSize - 1) / blockSize; add<<<numBlocks, blockSize>>>(N, d_x, d_y); cudaErrchk( hipPeekAtLastError() ); // cudaDeviceSynchronize(); hipMemcpy(h_y, d_y, size, hipMemcpyDeviceToHost); float maxError = 0.0f; for (int i = 0; i < N; i++) { maxError = fmax(maxError, fabs(h_y[i]-3.0f)); } std::cout << "Max error: " << maxError << std::endl; hipFree(d_x); hipFree(d_y); free(h_x); free(h_y); return 0; }
.text .file "main.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__addiPfS_ # -- Begin function _Z18__device_stub__addiPfS_ .p2align 4, 0x90 .type _Z18__device_stub__addiPfS_,@function _Z18__device_stub__addiPfS_: # @_Z18__device_stub__addiPfS_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movl %edi, 12(%rsp) movq %rsi, 72(%rsp) movq %rdx, 64(%rsp) leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 72(%rsp), %rax movq %rax, 88(%rsp) leaq 64(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z3addiPfS_, %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 _Z18__device_stub__addiPfS_, .Lfunc_end0-_Z18__device_stub__addiPfS_ .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function main .LCPI1_0: .long 0xc0400000 # float -3 .section .rodata.cst16,"aM",@progbits,16 .p2align 4, 0x0 .LCPI1_1: .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $144, %rsp .cfi_def_cfa_offset 192 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $1073741824, %edi # imm = 0x40000000 callq malloc movq %rax, %rbx movl $1073741824, %edi # imm = 0x40000000 callq malloc movq %rax, %r14 leaq 16(%rsp), %rdi movl $1073741824, %esi # imm = 0x40000000 callq hipMalloc leaq 8(%rsp), %rdi movl $1073741824, %esi # imm = 0x40000000 callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%rbx,%rax,4) # imm = 0x3F800000 movl $1073741824, (%r14,%rax,4) # imm = 0x40000000 incq %rax cmpq $268435456, %rax # imm = 0x10000000 jne .LBB1_1 # %bb.2: movq 16(%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294968320, %rdx # imm = 0x100000400 leaq 261120(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 16(%rsp), %rax movq 8(%rsp), %rcx movl $268435456, 28(%rsp) # imm = 0x10000000 movq %rax, 88(%rsp) movq %rcx, 80(%rsp) leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 88(%rsp), %rax movq %rax, 104(%rsp) leaq 80(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z3addiPfS_, %edi pushq 32(%rsp) .cfi_adjust_cfa_offset 8 pushq 48(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: callq hipPeekAtLastError testl %eax, %eax jne .LBB1_12 # %bb.5: # %_Z9GPUAssert10hipError_tPKcib.exit movq 8(%rsp), %rsi movl $1073741824, %edx # imm = 0x40000000 movq %r14, %rdi movl $2, %ecx callq hipMemcpy xorps %xmm2, %xmm2 xorl %eax, %eax movss .LCPI1_0(%rip), %xmm0 # xmm0 = mem[0],zero,zero,zero movaps .LCPI1_1(%rip), %xmm1 # xmm1 = [NaN,NaN,NaN,NaN] movaps %xmm2, %xmm5 .p2align 4, 0x90 .LBB1_6: # =>This Inner Loop Header: Depth=1 movss (%r14,%rax,4), %xmm3 # xmm3 = mem[0],zero,zero,zero addss %xmm0, %xmm3 andps %xmm1, %xmm3 cmpunordss %xmm5, %xmm5 movaps %xmm5, %xmm4 andps %xmm3, %xmm4 maxss %xmm2, %xmm3 andnps %xmm3, %xmm5 orps %xmm4, %xmm5 incq %rax movaps %xmm5, %xmm2 cmpq $268435456, %rax # imm = 0x10000000 jne .LBB1_6 # %bb.7: movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $11, %edx movaps %xmm5, 128(%rsp) # 16-byte Spill callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movaps 128(%rsp), %xmm0 # 16-byte Reload cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %r15 testq %r15, %r15 je .LBB1_13 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r15) je .LBB1_10 # %bb.9: movzbl 67(%r15), %ecx jmp .LBB1_11 .LBB1_10: movq %r15, %rdi movq %rax, %r12 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r12, %rax .LBB1_11: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $144, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .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_12: .cfi_def_cfa_offset 192 movq stderr(%rip), %rbx movl %eax, %edi movl %eax, %ebp callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $52, %r8d xorl %eax, %eax callq fprintf movl %ebp, %edi callq exit .LBB1_13: 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 $_Z3addiPfS_, %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 _Z3addiPfS_,@object # @_Z3addiPfS_ .section .rodata,"a",@progbits .globl _Z3addiPfS_ .p2align 3, 0x0 _Z3addiPfS_: .quad _Z18__device_stub__addiPfS_ .size _Z3addiPfS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/akhtyamovpavel/ParallelComputationExamples/master/CUDA/01-intro/04-memcpy/main.hip" .size .L.str, 140 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Max error: " .size .L.str.1, 12 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "GPUassert: %s %s %d\n" .size .L.str.2, 21 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z3addiPfS_" .size .L__unnamed_1, 12 .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__addiPfS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addiPfS_ .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 : _Z3addiPfS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x160], PT ; /* 0x0000580002007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0070*/ MOV R5, 0x40400000 ; /* 0x4040000000057802 */ /* 0x000fe20000000f00 */ /*0080*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*0090*/ IMAD.WIDE R2, R2, R3, c[0x0][0x170] ; /* 0x00005c0002027625 */ /* 0x000fca00078e0203 */ /*00a0*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addiPfS_ .globl _Z3addiPfS_ .p2align 8 .type _Z3addiPfS_,@function _Z3addiPfS_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b32 s3, s[0:1], 0x0 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_2 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_2) | instid1(VALU_DEP_2) v_lshlrev_b64 v[0:1], 2, v[1:2] v_mov_b32_e32 v2, 0x40400000 s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_3) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo 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 _Z3addiPfS_ .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 3 .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 _Z3addiPfS_, .Lfunc_end0-_Z3addiPfS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addiPfS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z3addiPfS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00077741_00000000-6_main.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3673: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3673: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z25__device_stub__Z3addiPfS_iPfS_ .type _Z25__device_stub__Z3addiPfS_iPfS_, @function _Z25__device_stub__Z3addiPfS_iPfS_: .LFB3695: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movl %edi, 28(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .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 _Z3addiPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3695: .size _Z25__device_stub__Z3addiPfS_iPfS_, .-_Z25__device_stub__Z3addiPfS_iPfS_ .globl _Z3addiPfS_ .type _Z3addiPfS_, @function _Z3addiPfS_: .LFB3696: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z3addiPfS_iPfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3696: .size _Z3addiPfS_, .-_Z3addiPfS_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "/home/ubuntu/Datasets/stackv2/train-structured/akhtyamovpavel/ParallelComputationExamples/master/CUDA/01-intro/04-memcpy/main.cu" .section .rodata.str1.1,"aMS",@progbits,1 .LC4: .string "GPUassert: %s %s %d\n" .LC7: .string "Max error: " .text .globl main .type main, @function main: .LFB3670: .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 $72, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $1073741824, %edi call malloc@PLT movq %rax, %r12 movl $1073741824, %edi call malloc@PLT movq %rax, %rbp leaq 16(%rsp), %rdi movl $1073741824, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $1073741824, %esi call cudaMalloc@PLT movl $0, %eax movss .LC1(%rip), %xmm1 movss .LC2(%rip), %xmm0 .L12: movss %xmm1, (%r12,%rax) movss %xmm0, 0(%rbp,%rax) addq $4, %rax cmpq $1073741824, %rax jne .L12 movl $1, %ecx movl $1073741824, %edx movq %r12, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $1073741824, %edx movq %rbp, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 44(%rsp) movl $1, 48(%rsp) movl $262144, 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 .L20 .L13: call cudaPeekAtLastError@PLT movl %eax, %ebx testl %eax, %eax jne .L21 movl $2, %ecx movl $1073741824, %edx movq 24(%rsp), %rsi movq %rbp, %rdi call cudaMemcpy@PLT movq %rbp, %rbx leaq 1073741824(%rbp), %r13 movl $0x00000000, 12(%rsp) .L15: movss (%rbx), %xmm0 subss .LC5(%rip), %xmm0 andps .LC6(%rip), %xmm0 movss 12(%rsp), %xmm1 call fmaxf@PLT movss %xmm0, 12(%rsp) addq $4, %rbx cmpq %r13, %rbx jne .L15 leaq .LC7(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L22 movl $0, %eax addq $72, %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 .L20: .cfi_restore_state movq 24(%rsp), %rdx movq 16(%rsp), %rsi movl $268435456, %edi call _Z25__device_stub__Z3addiPfS_iPfS_ jmp .L13 .L21: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rcx movl $50, %r9d leaq .LC3(%rip), %r8 leaq .LC4(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl %ebx, %edi call exit@PLT .L22: call __stack_chk_fail@PLT .cfi_endproc .LFE3670: .size main, .-main .section .rodata.str1.1 .LC8: .string "_Z3addiPfS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3698: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _Z3addiPfS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3698: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC1: .long 1065353216 .align 4 .LC2: .long 1073741824 .align 4 .LC5: .long 1077936128 .section .rodata.cst16,"aM",@progbits,16 .align 16 .LC6: .long 2147483647 .long 0 .long 0 .long 0 .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 "main.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__addiPfS_ # -- Begin function _Z18__device_stub__addiPfS_ .p2align 4, 0x90 .type _Z18__device_stub__addiPfS_,@function _Z18__device_stub__addiPfS_: # @_Z18__device_stub__addiPfS_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movl %edi, 12(%rsp) movq %rsi, 72(%rsp) movq %rdx, 64(%rsp) leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 72(%rsp), %rax movq %rax, 88(%rsp) leaq 64(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z3addiPfS_, %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 _Z18__device_stub__addiPfS_, .Lfunc_end0-_Z18__device_stub__addiPfS_ .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function main .LCPI1_0: .long 0xc0400000 # float -3 .section .rodata.cst16,"aM",@progbits,16 .p2align 4, 0x0 .LCPI1_1: .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .long 0x7fffffff # float NaN .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $144, %rsp .cfi_def_cfa_offset 192 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl $1073741824, %edi # imm = 0x40000000 callq malloc movq %rax, %rbx movl $1073741824, %edi # imm = 0x40000000 callq malloc movq %rax, %r14 leaq 16(%rsp), %rdi movl $1073741824, %esi # imm = 0x40000000 callq hipMalloc leaq 8(%rsp), %rdi movl $1073741824, %esi # imm = 0x40000000 callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%rbx,%rax,4) # imm = 0x3F800000 movl $1073741824, (%r14,%rax,4) # imm = 0x40000000 incq %rax cmpq $268435456, %rax # imm = 0x10000000 jne .LBB1_1 # %bb.2: movq 16(%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294968320, %rdx # imm = 0x100000400 leaq 261120(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 16(%rsp), %rax movq 8(%rsp), %rcx movl $268435456, 28(%rsp) # imm = 0x10000000 movq %rax, 88(%rsp) movq %rcx, 80(%rsp) leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 88(%rsp), %rax movq %rax, 104(%rsp) leaq 80(%rsp), %rax movq %rax, 112(%rsp) leaq 64(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z3addiPfS_, %edi pushq 32(%rsp) .cfi_adjust_cfa_offset 8 pushq 48(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: callq hipPeekAtLastError testl %eax, %eax jne .LBB1_12 # %bb.5: # %_Z9GPUAssert10hipError_tPKcib.exit movq 8(%rsp), %rsi movl $1073741824, %edx # imm = 0x40000000 movq %r14, %rdi movl $2, %ecx callq hipMemcpy xorps %xmm2, %xmm2 xorl %eax, %eax movss .LCPI1_0(%rip), %xmm0 # xmm0 = mem[0],zero,zero,zero movaps .LCPI1_1(%rip), %xmm1 # xmm1 = [NaN,NaN,NaN,NaN] movaps %xmm2, %xmm5 .p2align 4, 0x90 .LBB1_6: # =>This Inner Loop Header: Depth=1 movss (%r14,%rax,4), %xmm3 # xmm3 = mem[0],zero,zero,zero addss %xmm0, %xmm3 andps %xmm1, %xmm3 cmpunordss %xmm5, %xmm5 movaps %xmm5, %xmm4 andps %xmm3, %xmm4 maxss %xmm2, %xmm3 andnps %xmm3, %xmm5 orps %xmm4, %xmm5 incq %rax movaps %xmm5, %xmm2 cmpq $268435456, %rax # imm = 0x10000000 jne .LBB1_6 # %bb.7: movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $11, %edx movaps %xmm5, 128(%rsp) # 16-byte Spill callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movaps 128(%rsp), %xmm0 # 16-byte Reload cvtss2sd %xmm0, %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %r15 testq %r15, %r15 je .LBB1_13 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%r15) je .LBB1_10 # %bb.9: movzbl 67(%r15), %ecx jmp .LBB1_11 .LBB1_10: movq %r15, %rdi movq %rax, %r12 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r15), %rax movq %r15, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r12, %rax .LBB1_11: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $144, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .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_12: .cfi_def_cfa_offset 192 movq stderr(%rip), %rbx movl %eax, %edi movl %eax, %ebp callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $52, %r8d xorl %eax, %eax callq fprintf movl %ebp, %edi callq exit .LBB1_13: 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 $_Z3addiPfS_, %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 _Z3addiPfS_,@object # @_Z3addiPfS_ .section .rodata,"a",@progbits .globl _Z3addiPfS_ .p2align 3, 0x0 _Z3addiPfS_: .quad _Z18__device_stub__addiPfS_ .size _Z3addiPfS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/akhtyamovpavel/ParallelComputationExamples/master/CUDA/01-intro/04-memcpy/main.hip" .size .L.str, 140 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Max error: " .size .L.str.1, 12 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "GPUassert: %s %s %d\n" .size .L.str.2, 21 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z3addiPfS_" .size .L__unnamed_1, 12 .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__addiPfS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addiPfS_ .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 <math.h> #include <stdlib.h> #include <time.h> #include <cuda.h> #include <curand.h> #define CUDA_CALL(x) do { if((x)!=cudaSuccess) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) #define DEPTH 5 // no of layers #define BREADTH 1000 // nodes per layer #define RATE 0.0001 // learning rate #define NDIM(a,b) ((a)*BREADTH+(b)) #define WDIM(a,b,c) ((a)*BREADTH*BREADTH+(b)*BREADTH+(c)) float *W, *R; // [DEPTH-1][BREADTH][BREADTH] W_kij : weight of edge b/w jth and ith node from kth layer and k+1th layer resp. float *N; // [DEPTH][BREADTH] N_ki : jth node in kth layer float *Y, *pY; // [BREADTH] expected and predicted output vectors float *U, *V; // [BREADTH] vectors for backpropagation chain-rule float Yh[BREADTH], Dh[BREADTH]; // host arrays: expected and predicted output int count = 0; #define f(x) (tanhf((x))) // activation function #define df(y) (1 - (y) * (y)) // derivative of activation function in y terms __global__ void predict_N( float *N) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; // erase all nodes except first layer for ( i = tid + BREADTH ; i < BREADTH * DEPTH ; i += stride ) N[i] = 0; } __global__ void predict_WN( float *N, float *W, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; R[h] = W[wIndex + h] * N[nIndex + j]; } } __global__ void predict_Ri( float *R, int n, int j) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid % BREADTH ; i < BREADTH ; i += stride ) { int itid = tid / BREADTH; int istride = stride / BREADTH; istride += (istride == 0); int rIndex = i * BREADTH; for ( int h = 2 * j * itid; h < j * (n - 1); h += 2 * j * istride) R[rIndex + h] += R[rIndex + h + j]; } } __global__ void predict_Nk( float *N, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) N[NDIM(k,i)] = f( R[NDIM(i,0)]); } __global__ void predict_Yp( float *N, float *pY) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) pY[i] = N[NDIM(DEPTH-1,i)]; } __global__ void train_V( float *N, float *Y, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2; // caching chain-rule calculation in vector to aid in backpropagation for ( int i = tid; i < BREADTH; i += stride ) { V[i] = (N[NDIM(q+1,i)] - Y[i]) * df( N[NDIM(q+1,i)]); } } __global__ void train_Wf( float *N, float *W, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2, h; // update weights of outer layer for ( h = tid; h < BREADTH * BREADTH; h += stride ) { int i = h / BREADTH; int j = h % BREADTH; W[WDIM(q,i,j)] -= (float)RATE * V[i] * N[NDIM(q,j)]; } } __global__ void train_WN( float *N, float *W, float *A, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; R[NDIM(j,i)] = A[i] * W[wIndex + h] * df( N[nIndex + j]); } } __global__ void train_B( float *R, float *B) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) B[i] = R[NDIM(i,0)]; } __global__ void train_Wh( float *N, float *W, float *A, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; W[wIndex + h] -= (float)RATE * A[i] * N[nIndex + j]; } } void train() { train_V<<<128, 32>>>( N, Y, V); cudaDeviceSynchronize(); train_Wf<<<128, 32>>>( N, W, V); cudaDeviceSynchronize(); int q = DEPTH - 2; float *A = &V[0], *B = &U[0]; float *temp; // updating weights of hidden layers (backpropagation) for ( int k = q - 1; k >= 0; k-- ) { // chain-rule expansion by multiplying with next layer's weight matrix and // derivative of activation function on the same layer's nodes train_WN<<<128, 32>>>( N, W, A, R, k+1); cudaDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); cudaDeviceSynchronize(); j *= 2; n = (n+1)/2; } train_B<<<128, 32>>>( R, B); cudaDeviceSynchronize(); temp = B; B = A; A = temp; // adjust weights by gradient descent train_Wh<<<128, 32>>>( N, W, A, k); cudaDeviceSynchronize(); } } __global__ void normalize( float *W, float *N, float *Y) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i, w = (DEPTH-1) * BREADTH * BREADTH; for ( i = tid; i < w; i+= stride) W[i] = 1/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) N[i] = i/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) Y[i] = i/ (float)BREADTH; } void initialize() { int w = (DEPTH-1) * BREADTH * BREADTH; // device memory allocation of global arrays cudaMalloc( (void**)(&W), sizeof(float) * w); cudaMalloc( (void**)(&R), sizeof(float) * BREADTH * BREADTH); cudaMalloc( (void**)(&N), sizeof(float) * DEPTH * BREADTH); cudaMalloc( (void**)(&Y), sizeof(float) * BREADTH); cudaMalloc( (void**)(&pY), sizeof(float) * BREADTH); cudaMalloc( (void**)(&U), sizeof(float) * BREADTH); cudaMalloc( (void**)(&V), sizeof(float) * BREADTH); /* // random number generator curandGenerator_t gen; // Create pseudo-random number generator curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); // Set seed curandSetPseudoRandomGeneratorSeed(gen, time(0)); // Generate floats on device curandGenerateUniform(gen, W, w); curandGenerateUniform(gen, N, BREADTH * DEPTH); curandGenerateUniform(gen, Y, BREADTH); curandDestroyGenerator(gen); */ normalize<<<128, 32>>>(W, N, Y); cudaDeviceSynchronize(); } void calculateError() { float err = 0, dY; // dY : difference b/w expected and predicted output, and // err : sum of squares of dY for ( int i = 0 ; i < BREADTH ; i++ ) { dY = (Dh[i] - Yh[i]); err += dY * dY; } printf("ITERATION : %d\tLOSS : %f\t", count, err); } void freeResources() { cudaFree(Y); cudaFree(N); cudaFree(W); cudaFree(pY); cudaFree(U); cudaFree(V); cudaFree(R); } void predict() { predict_N<<<128, 32>>>(N); cudaDeviceSynchronize(); for ( int k = 1; k < DEPTH ; k++ ) { predict_WN<<<128, 32>>>( N, W, R, k-1); cudaDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); cudaDeviceSynchronize(); j *= 2; n = (n+1)/2; } predict_Nk<<<128, 32>>>( N, R, k); cudaDeviceSynchronize(); } predict_Yp<<<128, 32>>>( N, pY); cudaDeviceSynchronize(); } int main() { initialize(); for ( count = 0; count < 4; count++) { clock_t begin, end; begin = clock(); predict(); train(); end = clock(); float CpuTime = (float)(end - begin) / CLOCKS_PER_SEC; CUDA_CALL(cudaMemcpy(Yh, Y, BREADTH * sizeof(float), cudaMemcpyDeviceToHost)); CUDA_CALL(cudaMemcpy(Dh, pY, BREADTH * sizeof(float), cudaMemcpyDeviceToHost)); calculateError(); printf("TIME: %f\n", CpuTime); } freeResources(); return 0; }
.file "tmpxft_001405c6_00000000-6_deep.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2065: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2065: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "ITERATION : %d\tLOSS : %f\t" .text .globl _Z14calculateErrorv .type _Z14calculateErrorv, @function _Z14calculateErrorv: .LFB2059: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movl $0, %eax pxor %xmm1, %xmm1 leaq Dh(%rip), %rcx leaq Yh(%rip), %rdx .L4: movss (%rcx,%rax), %xmm0 subss (%rdx,%rax), %xmm0 mulss %xmm0, %xmm0 addss %xmm0, %xmm1 addq $4, %rax cmpq $4000, %rax jne .L4 pxor %xmm0, %xmm0 cvtss2sd %xmm1, %xmm0 movl count(%rip), %edx leaq .LC1(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _Z14calculateErrorv, .-_Z14calculateErrorv .globl _Z13freeResourcesv .type _Z13freeResourcesv, @function _Z13freeResourcesv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq Y(%rip), %rdi call cudaFree@PLT movq N(%rip), %rdi call cudaFree@PLT movq W(%rip), %rdi call cudaFree@PLT movq pY(%rip), %rdi call cudaFree@PLT movq U(%rip), %rdi call cudaFree@PLT movq V(%rip), %rdi call cudaFree@PLT movq R(%rip), %rdi call cudaFree@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _Z13freeResourcesv, .-_Z13freeResourcesv .globl _Z28__device_stub__Z9predict_NPfPf .type _Z28__device_stub__Z9predict_NPfPf, @function _Z28__device_stub__Z9predict_NPfPf: .LFB2087: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L13 .L9: movq 88(%rsp), %rax subq %fs:40, %rax jne .L14 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L13: .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 _Z9predict_NPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L9 .L14: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z28__device_stub__Z9predict_NPfPf, .-_Z28__device_stub__Z9predict_NPfPf .globl _Z9predict_NPf .type _Z9predict_NPf, @function _Z9predict_NPf: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z9predict_NPfPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z9predict_NPf, .-_Z9predict_NPf .globl _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i .type _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i, @function _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i: .LFB2089: .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 .L21 .L17: movq 136(%rsp), %rax subq %fs:40, %rax jne .L22 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L21: .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 _Z10predict_WNPfS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L17 .L22: call __stack_chk_fail@PLT .cfi_endproc .LFE2089: .size _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i, .-_Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i .globl _Z10predict_WNPfS_S_i .type _Z10predict_WNPfS_S_i, @function _Z10predict_WNPfS_S_i: .LFB2090: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2090: .size _Z10predict_WNPfS_S_i, .-_Z10predict_WNPfS_S_i .globl _Z32__device_stub__Z10predict_RiPfiiPfii .type _Z32__device_stub__Z10predict_RiPfiiPfii, @function _Z32__device_stub__Z10predict_RiPfiiPfii: .LFB2091: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L29 .L25: movq 104(%rsp), %rax subq %fs:40, %rax jne .L30 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L29: .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 _Z10predict_RiPfii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L25 .L30: call __stack_chk_fail@PLT .cfi_endproc .LFE2091: .size _Z32__device_stub__Z10predict_RiPfiiPfii, .-_Z32__device_stub__Z10predict_RiPfiiPfii .globl _Z10predict_RiPfii .type _Z10predict_RiPfii, @function _Z10predict_RiPfii: .LFB2092: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z10predict_RiPfiiPfii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2092: .size _Z10predict_RiPfii, .-_Z10predict_RiPfii .globl _Z33__device_stub__Z10predict_NkPfS_iPfS_i .type _Z33__device_stub__Z10predict_NkPfS_iPfS_i, @function _Z33__device_stub__Z10predict_NkPfS_iPfS_i: .LFB2093: .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 .L37 .L33: movq 120(%rsp), %rax subq %fs:40, %rax jne .L38 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L37: .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 _Z10predict_NkPfS_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L33 .L38: call __stack_chk_fail@PLT .cfi_endproc .LFE2093: .size _Z33__device_stub__Z10predict_NkPfS_iPfS_i, .-_Z33__device_stub__Z10predict_NkPfS_iPfS_i .globl _Z10predict_NkPfS_i .type _Z10predict_NkPfS_i, @function _Z10predict_NkPfS_i: .LFB2094: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z33__device_stub__Z10predict_NkPfS_iPfS_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2094: .size _Z10predict_NkPfS_i, .-_Z10predict_NkPfS_i .globl _Z32__device_stub__Z10predict_YpPfS_PfS_ .type _Z32__device_stub__Z10predict_YpPfS_PfS_, @function _Z32__device_stub__Z10predict_YpPfS_PfS_: .LFB2095: .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 .L45 .L41: movq 104(%rsp), %rax subq %fs:40, %rax jne .L46 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L45: .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 _Z10predict_YpPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L41 .L46: call __stack_chk_fail@PLT .cfi_endproc .LFE2095: .size _Z32__device_stub__Z10predict_YpPfS_PfS_, .-_Z32__device_stub__Z10predict_YpPfS_PfS_ .globl _Z10predict_YpPfS_ .type _Z10predict_YpPfS_, @function _Z10predict_YpPfS_: .LFB2096: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z10predict_YpPfS_PfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2096: .size _Z10predict_YpPfS_, .-_Z10predict_YpPfS_ .globl _Z7predictv .type _Z7predictv, @function _Z7predictv: .LFB2061: .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 $32, %rsp .cfi_def_cfa_offset 80 movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L60 .L50: call cudaDeviceSynchronize@PLT movl $1, %r13d movl $1000, %r14d jmp .L55 .L60: movq N(%rip), %rdi call _Z28__device_stub__Z9predict_NPfPf jmp .L50 .L64: leal -1(%r13), %ecx movq R(%rip), %rdx movq W(%rip), %rsi movq N(%rip), %rdi call _Z35__device_stub__Z10predict_WNPfS_S_iPfS_S_i jmp .L51 .L52: call cudaDeviceSynchronize@PLT addl %r12d, %r12d leal 1(%rbx), %eax movl %eax, %edx shrl $31, %edx addl %edx, %eax sarl %eax movl %eax, %ebx subl $1, %ebp je .L61 .L53: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 jne .L52 movl %r12d, %edx movl %ebx, %esi movq R(%rip), %rdi call _Z32__device_stub__Z10predict_RiPfiiPfii jmp .L52 .L61: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L62 .L54: call cudaDeviceSynchronize@PLT addl $1, %r13d cmpl $5, %r13d je .L63 .L55: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L64 .L51: call cudaDeviceSynchronize@PLT movl $10, %ebp movl %r14d, %ebx movl $1, %r12d jmp .L53 .L62: movl %r13d, %edx movq R(%rip), %rsi movq N(%rip), %rdi call _Z33__device_stub__Z10predict_NkPfS_iPfS_i jmp .L54 .L63: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L65 .L56: call cudaDeviceSynchronize@PLT addq $32, %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 .L65: .cfi_restore_state movq pY(%rip), %rsi movq N(%rip), %rdi call _Z32__device_stub__Z10predict_YpPfS_PfS_ jmp .L56 .cfi_endproc .LFE2061: .size _Z7predictv, .-_Z7predictv .globl _Z30__device_stub__Z7train_VPfS_S_PfS_S_ .type _Z30__device_stub__Z7train_VPfS_S_PfS_S_, @function _Z30__device_stub__Z7train_VPfS_S_PfS_S_: .LFB2097: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L70 .L66: movq 120(%rsp), %rax subq %fs:40, %rax jne .L71 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L70: .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 _Z7train_VPfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L66 .L71: call __stack_chk_fail@PLT .cfi_endproc .LFE2097: .size _Z30__device_stub__Z7train_VPfS_S_PfS_S_, .-_Z30__device_stub__Z7train_VPfS_S_PfS_S_ .globl _Z7train_VPfS_S_ .type _Z7train_VPfS_S_, @function _Z7train_VPfS_S_: .LFB2098: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z7train_VPfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2098: .size _Z7train_VPfS_S_, .-_Z7train_VPfS_S_ .globl _Z31__device_stub__Z8train_WfPfS_S_PfS_S_ .type _Z31__device_stub__Z8train_WfPfS_S_PfS_S_, @function _Z31__device_stub__Z8train_WfPfS_S_PfS_S_: .LFB2099: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L78 .L74: movq 120(%rsp), %rax subq %fs:40, %rax jne .L79 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L78: .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 _Z8train_WfPfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L74 .L79: call __stack_chk_fail@PLT .cfi_endproc .LFE2099: .size _Z31__device_stub__Z8train_WfPfS_S_PfS_S_, .-_Z31__device_stub__Z8train_WfPfS_S_PfS_S_ .globl _Z8train_WfPfS_S_ .type _Z8train_WfPfS_S_, @function _Z8train_WfPfS_S_: .LFB2100: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z8train_WfPfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2100: .size _Z8train_WfPfS_S_, .-_Z8train_WfPfS_S_ .globl _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i .type _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i, @function _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i: .LFB2101: .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 .L86 .L82: movq 152(%rsp), %rax subq %fs:40, %rax jne .L87 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L86: .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 _Z8train_WNPfS_S_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L82 .L87: call __stack_chk_fail@PLT .cfi_endproc .LFE2101: .size _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i, .-_Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i .globl _Z8train_WNPfS_S_S_i .type _Z8train_WNPfS_S_S_i, @function _Z8train_WNPfS_S_S_i: .LFB2102: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2102: .size _Z8train_WNPfS_S_S_i, .-_Z8train_WNPfS_S_S_i .globl _Z28__device_stub__Z7train_BPfS_PfS_ .type _Z28__device_stub__Z7train_BPfS_PfS_, @function _Z28__device_stub__Z7train_BPfS_PfS_: .LFB2103: .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 .L94 .L90: movq 104(%rsp), %rax subq %fs:40, %rax jne .L95 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L94: .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 _Z7train_BPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L90 .L95: call __stack_chk_fail@PLT .cfi_endproc .LFE2103: .size _Z28__device_stub__Z7train_BPfS_PfS_, .-_Z28__device_stub__Z7train_BPfS_PfS_ .globl _Z7train_BPfS_ .type _Z7train_BPfS_, @function _Z7train_BPfS_: .LFB2104: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z7train_BPfS_PfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2104: .size _Z7train_BPfS_, .-_Z7train_BPfS_ .globl _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i .type _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i, @function _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i: .LFB2105: .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 .L102 .L98: movq 136(%rsp), %rax subq %fs:40, %rax jne .L103 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L102: .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 _Z8train_WhPfS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L98 .L103: call __stack_chk_fail@PLT .cfi_endproc .LFE2105: .size _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i, .-_Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i .globl _Z8train_WhPfS_S_i .type _Z8train_WhPfS_S_i, @function _Z8train_WhPfS_S_i: .LFB2106: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2106: .size _Z8train_WhPfS_S_i, .-_Z8train_WhPfS_S_i .globl _Z5trainv .type _Z5trainv, @function _Z5trainv: .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 $40, %rsp .cfi_def_cfa_offset 96 movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L118 .L107: call cudaDeviceSynchronize@PLT movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L119 .L108: call cudaDeviceSynchronize@PLT movq V(%rip), %r15 movq U(%rip), %r14 movl $2, %r13d jmp .L114 .L118: movq V(%rip), %rdx movq Y(%rip), %rsi movq N(%rip), %rdi call _Z30__device_stub__Z7train_VPfS_S_PfS_S_ jmp .L107 .L119: movq V(%rip), %rdx movq W(%rip), %rsi movq N(%rip), %rdi call _Z31__device_stub__Z8train_WfPfS_S_PfS_S_ jmp .L108 .L124: leal 1(%r13), %r8d movq R(%rip), %rcx movq %r15, %rdx movq W(%rip), %rsi movq N(%rip), %rdi call _Z34__device_stub__Z8train_WNPfS_S_S_iPfS_S_S_i jmp .L109 .L110: call cudaDeviceSynchronize@PLT addl %r12d, %r12d leal 1(%rbx), %eax movl %eax, %ebx shrl $31, %ebx addl %eax, %ebx sarl %ebx subl $1, %ebp je .L120 .L111: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 jne .L110 movl %r12d, %edx movl %ebx, %esi movq R(%rip), %rdi call _Z32__device_stub__Z10predict_RiPfiiPfii jmp .L110 .L120: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L121 .L112: call cudaDeviceSynchronize@PLT movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L122 .L113: call cudaDeviceSynchronize@PLT subl $1, %r13d movq %r14, %rax movq %r15, %r14 cmpl $-1, %r13d je .L123 movq %rax, %r15 .L114: movl $32, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $128, 8(%rsp) movl $1, 12(%rsp) movl $1, 16(%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 .L124 .L109: call cudaDeviceSynchronize@PLT movl $10, %ebp movl $1000, %ebx movl $1, %r12d jmp .L111 .L121: movq %r14, %rsi movq R(%rip), %rdi call _Z28__device_stub__Z7train_BPfS_PfS_ jmp .L112 .L122: movl %r13d, %ecx movq %r14, %rdx movq W(%rip), %rsi movq N(%rip), %rdi call _Z32__device_stub__Z8train_WhPfS_S_iPfS_S_i jmp .L113 .L123: 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 .LFE2057: .size _Z5trainv, .-_Z5trainv .globl _Z32__device_stub__Z9normalizePfS_S_PfS_S_ .type _Z32__device_stub__Z9normalizePfS_S_PfS_S_, @function _Z32__device_stub__Z9normalizePfS_S_PfS_S_: .LFB2107: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L129 .L125: movq 120(%rsp), %rax subq %fs:40, %rax jne .L130 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L129: .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 _Z9normalizePfS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L125 .L130: call __stack_chk_fail@PLT .cfi_endproc .LFE2107: .size _Z32__device_stub__Z9normalizePfS_S_PfS_S_, .-_Z32__device_stub__Z9normalizePfS_S_PfS_S_ .globl _Z9normalizePfS_S_ .type _Z9normalizePfS_S_, @function _Z9normalizePfS_S_: .LFB2108: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z9normalizePfS_S_PfS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2108: .size _Z9normalizePfS_S_, .-_Z9normalizePfS_S_ .globl _Z10initializev .type _Z10initializev, @function _Z10initializev: .LFB2058: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $16000000, %esi leaq W(%rip), %rdi call cudaMalloc@PLT movl $4000000, %esi leaq R(%rip), %rdi call cudaMalloc@PLT movl $20000, %esi leaq N(%rip), %rdi call cudaMalloc@PLT movl $4000, %esi leaq Y(%rip), %rdi call cudaMalloc@PLT movl $4000, %esi leaq pY(%rip), %rdi call cudaMalloc@PLT movl $4000, %esi leaq U(%rip), %rdi call cudaMalloc@PLT movl $4000, %esi leaq V(%rip), %rdi call cudaMalloc@PLT movl $32, 20(%rsp) movl $1, 24(%rsp) movl $128, 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 .L136 .L134: call cudaDeviceSynchronize@PLT addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L136: .cfi_restore_state movq Y(%rip), %rdx movq N(%rip), %rsi movq W(%rip), %rdi call _Z32__device_stub__Z9normalizePfS_S_PfS_S_ jmp .L134 .cfi_endproc .LFE2058: .size _Z10initializev, .-_Z10initializev .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "/home/ubuntu/Datasets/stackv2/train-structured/ShamsArfeen/Deep-Learning-with-CUDA/main/deep.cu" .section .rodata.str1.1 .LC4: .string "Error at %s:%d\n" .LC5: .string "TIME: %f\n" .text .globl main .type main, @function main: .LFB2062: .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 $24, %rsp .cfi_def_cfa_offset 64 call _Z10initializev movl $0, count(%rip) leaq Yh(%rip), %rbx leaq Dh(%rip), %rbp leaq .LC5(%rip), %r12 .L141: call clock@PLT movq %rax, %r13 call _Z7predictv call _Z5trainv call clock@PLT subq %r13, %rax pxor %xmm0, %xmm0 cvtsi2ssq %rax, %xmm0 divss .LC2(%rip), %xmm0 movss %xmm0, 12(%rsp) movl $2, %ecx movl $4000, %edx movq Y(%rip), %rsi movq %rbx, %rdi call cudaMemcpy@PLT testl %eax, %eax jne .L144 movl $2, %ecx movl $4000, %edx movq pY(%rip), %rsi movq %rbp, %rdi call cudaMemcpy@PLT testl %eax, %eax jne .L145 call _Z14calculateErrorv pxor %xmm0, %xmm0 cvtss2sd 12(%rsp), %xmm0 movq %r12, %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl count(%rip), %eax addl $1, %eax movl %eax, count(%rip) cmpl $3, %eax jle .L141 call _Z13freeResourcesv movl $0, %eax .L137: addq $24, %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 .L144: .cfi_restore_state movl $318, %ecx leaq .LC3(%rip), %rdx leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %eax jmp .L137 .L145: movl $319, %ecx leaq .LC3(%rip), %rdx leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1, %eax jmp .L137 .cfi_endproc .LFE2062: .size main, .-main .section .rodata.str1.1 .LC6: .string "_Z9normalizePfS_S_" .LC7: .string "_Z8train_WhPfS_S_i" .LC8: .string "_Z7train_BPfS_" .LC9: .string "_Z8train_WNPfS_S_S_i" .LC10: .string "_Z8train_WfPfS_S_" .LC11: .string "_Z7train_VPfS_S_" .LC12: .string "_Z10predict_YpPfS_" .LC13: .string "_Z10predict_NkPfS_i" .LC14: .string "_Z10predict_RiPfii" .LC15: .string "_Z10predict_WNPfS_S_i" .LC16: .string "_Z9predict_NPf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2110: .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 .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z9normalizePfS_S_(%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 .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z8train_WhPfS_S_i(%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 .LC8(%rip), %rdx movq %rdx, %rcx leaq _Z7train_BPfS_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _Z8train_WNPfS_S_S_i(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _Z8train_WfPfS_S_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _Z7train_VPfS_S_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _Z10predict_YpPfS_(%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 .LC13(%rip), %rdx movq %rdx, %rcx leaq _Z10predict_NkPfS_i(%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 .LC14(%rip), %rdx movq %rdx, %rcx leaq _Z10predict_RiPfii(%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 .LC15(%rip), %rdx movq %rdx, %rcx leaq _Z10predict_WNPfS_S_i(%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 .LC16(%rip), %rdx movq %rdx, %rcx leaq _Z9predict_NPf(%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 .LFE2110: .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 count .bss .align 4 .type count, @object .size count, 4 count: .zero 4 .globl Dh .align 32 .type Dh, @object .size Dh, 4000 Dh: .zero 4000 .globl Yh .align 32 .type Yh, @object .size Yh, 4000 Yh: .zero 4000 .globl V .align 8 .type V, @object .size V, 8 V: .zero 8 .globl U .align 8 .type U, @object .size U, 8 U: .zero 8 .globl pY .align 8 .type pY, @object .size pY, 8 pY: .zero 8 .globl Y .align 8 .type Y, @object .size Y, 8 Y: .zero 8 .globl N .align 8 .type N, @object .size N, 8 N: .zero 8 .globl R .align 8 .type R, @object .size R, 8 R: .zero 8 .globl W .align 8 .type W, @object .size W, 8 W: .zero 8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC2: .long 1232348160 .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 <math.h> #include <stdlib.h> #include <time.h> #include <cuda.h> #include <curand.h> #define CUDA_CALL(x) do { if((x)!=cudaSuccess) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) #define DEPTH 5 // no of layers #define BREADTH 1000 // nodes per layer #define RATE 0.0001 // learning rate #define NDIM(a,b) ((a)*BREADTH+(b)) #define WDIM(a,b,c) ((a)*BREADTH*BREADTH+(b)*BREADTH+(c)) float *W, *R; // [DEPTH-1][BREADTH][BREADTH] W_kij : weight of edge b/w jth and ith node from kth layer and k+1th layer resp. float *N; // [DEPTH][BREADTH] N_ki : jth node in kth layer float *Y, *pY; // [BREADTH] expected and predicted output vectors float *U, *V; // [BREADTH] vectors for backpropagation chain-rule float Yh[BREADTH], Dh[BREADTH]; // host arrays: expected and predicted output int count = 0; #define f(x) (tanhf((x))) // activation function #define df(y) (1 - (y) * (y)) // derivative of activation function in y terms __global__ void predict_N( float *N) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; // erase all nodes except first layer for ( i = tid + BREADTH ; i < BREADTH * DEPTH ; i += stride ) N[i] = 0; } __global__ void predict_WN( float *N, float *W, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; R[h] = W[wIndex + h] * N[nIndex + j]; } } __global__ void predict_Ri( float *R, int n, int j) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid % BREADTH ; i < BREADTH ; i += stride ) { int itid = tid / BREADTH; int istride = stride / BREADTH; istride += (istride == 0); int rIndex = i * BREADTH; for ( int h = 2 * j * itid; h < j * (n - 1); h += 2 * j * istride) R[rIndex + h] += R[rIndex + h + j]; } } __global__ void predict_Nk( float *N, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) N[NDIM(k,i)] = f( R[NDIM(i,0)]); } __global__ void predict_Yp( float *N, float *pY) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) pY[i] = N[NDIM(DEPTH-1,i)]; } __global__ void train_V( float *N, float *Y, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2; // caching chain-rule calculation in vector to aid in backpropagation for ( int i = tid; i < BREADTH; i += stride ) { V[i] = (N[NDIM(q+1,i)] - Y[i]) * df( N[NDIM(q+1,i)]); } } __global__ void train_Wf( float *N, float *W, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2, h; // update weights of outer layer for ( h = tid; h < BREADTH * BREADTH; h += stride ) { int i = h / BREADTH; int j = h % BREADTH; W[WDIM(q,i,j)] -= (float)RATE * V[i] * N[NDIM(q,j)]; } } __global__ void train_WN( float *N, float *W, float *A, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; R[NDIM(j,i)] = A[i] * W[wIndex + h] * df( N[nIndex + j]); } } __global__ void train_B( float *R, float *B) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) B[i] = R[NDIM(i,0)]; } __global__ void train_Wh( float *N, float *W, float *A, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; W[wIndex + h] -= (float)RATE * A[i] * N[nIndex + j]; } } void train() { train_V<<<128, 32>>>( N, Y, V); cudaDeviceSynchronize(); train_Wf<<<128, 32>>>( N, W, V); cudaDeviceSynchronize(); int q = DEPTH - 2; float *A = &V[0], *B = &U[0]; float *temp; // updating weights of hidden layers (backpropagation) for ( int k = q - 1; k >= 0; k-- ) { // chain-rule expansion by multiplying with next layer's weight matrix and // derivative of activation function on the same layer's nodes train_WN<<<128, 32>>>( N, W, A, R, k+1); cudaDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); cudaDeviceSynchronize(); j *= 2; n = (n+1)/2; } train_B<<<128, 32>>>( R, B); cudaDeviceSynchronize(); temp = B; B = A; A = temp; // adjust weights by gradient descent train_Wh<<<128, 32>>>( N, W, A, k); cudaDeviceSynchronize(); } } __global__ void normalize( float *W, float *N, float *Y) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i, w = (DEPTH-1) * BREADTH * BREADTH; for ( i = tid; i < w; i+= stride) W[i] = 1/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) N[i] = i/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) Y[i] = i/ (float)BREADTH; } void initialize() { int w = (DEPTH-1) * BREADTH * BREADTH; // device memory allocation of global arrays cudaMalloc( (void**)(&W), sizeof(float) * w); cudaMalloc( (void**)(&R), sizeof(float) * BREADTH * BREADTH); cudaMalloc( (void**)(&N), sizeof(float) * DEPTH * BREADTH); cudaMalloc( (void**)(&Y), sizeof(float) * BREADTH); cudaMalloc( (void**)(&pY), sizeof(float) * BREADTH); cudaMalloc( (void**)(&U), sizeof(float) * BREADTH); cudaMalloc( (void**)(&V), sizeof(float) * BREADTH); /* // random number generator curandGenerator_t gen; // Create pseudo-random number generator curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); // Set seed curandSetPseudoRandomGeneratorSeed(gen, time(0)); // Generate floats on device curandGenerateUniform(gen, W, w); curandGenerateUniform(gen, N, BREADTH * DEPTH); curandGenerateUniform(gen, Y, BREADTH); curandDestroyGenerator(gen); */ normalize<<<128, 32>>>(W, N, Y); cudaDeviceSynchronize(); } void calculateError() { float err = 0, dY; // dY : difference b/w expected and predicted output, and // err : sum of squares of dY for ( int i = 0 ; i < BREADTH ; i++ ) { dY = (Dh[i] - Yh[i]); err += dY * dY; } printf("ITERATION : %d\tLOSS : %f\t", count, err); } void freeResources() { cudaFree(Y); cudaFree(N); cudaFree(W); cudaFree(pY); cudaFree(U); cudaFree(V); cudaFree(R); } void predict() { predict_N<<<128, 32>>>(N); cudaDeviceSynchronize(); for ( int k = 1; k < DEPTH ; k++ ) { predict_WN<<<128, 32>>>( N, W, R, k-1); cudaDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); cudaDeviceSynchronize(); j *= 2; n = (n+1)/2; } predict_Nk<<<128, 32>>>( N, R, k); cudaDeviceSynchronize(); } predict_Yp<<<128, 32>>>( N, pY); cudaDeviceSynchronize(); } int main() { initialize(); for ( count = 0; count < 4; count++) { clock_t begin, end; begin = clock(); predict(); train(); end = clock(); float CpuTime = (float)(end - begin) / CLOCKS_PER_SEC; CUDA_CALL(cudaMemcpy(Yh, Y, BREADTH * sizeof(float), cudaMemcpyDeviceToHost)); CUDA_CALL(cudaMemcpy(Dh, pY, BREADTH * sizeof(float), cudaMemcpyDeviceToHost)); calculateError(); printf("TIME: %f\n", CpuTime); } freeResources(); return 0; }
#include <stdio.h> #include <math.h> #include <stdlib.h> #include <time.h> #include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #define CUDA_CALL(x) do { if((x)!=hipSuccess) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) #define DEPTH 5 // no of layers #define BREADTH 1000 // nodes per layer #define RATE 0.0001 // learning rate #define NDIM(a,b) ((a)*BREADTH+(b)) #define WDIM(a,b,c) ((a)*BREADTH*BREADTH+(b)*BREADTH+(c)) float *W, *R; // [DEPTH-1][BREADTH][BREADTH] W_kij : weight of edge b/w jth and ith node from kth layer and k+1th layer resp. float *N; // [DEPTH][BREADTH] N_ki : jth node in kth layer float *Y, *pY; // [BREADTH] expected and predicted output vectors float *U, *V; // [BREADTH] vectors for backpropagation chain-rule float Yh[BREADTH], Dh[BREADTH]; // host arrays: expected and predicted output int count = 0; #define f(x) (tanhf((x))) // activation function #define df(y) (1 - (y) * (y)) // derivative of activation function in y terms __global__ void predict_N( float *N) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; // erase all nodes except first layer for ( i = tid + BREADTH ; i < BREADTH * DEPTH ; i += stride ) N[i] = 0; } __global__ void predict_WN( float *N, float *W, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; R[h] = W[wIndex + h] * N[nIndex + j]; } } __global__ void predict_Ri( float *R, int n, int j) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid % BREADTH ; i < BREADTH ; i += stride ) { int itid = tid / BREADTH; int istride = stride / BREADTH; istride += (istride == 0); int rIndex = i * BREADTH; for ( int h = 2 * j * itid; h < j * (n - 1); h += 2 * j * istride) R[rIndex + h] += R[rIndex + h + j]; } } __global__ void predict_Nk( float *N, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) N[NDIM(k,i)] = f( R[NDIM(i,0)]); } __global__ void predict_Yp( float *N, float *pY) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) pY[i] = N[NDIM(DEPTH-1,i)]; } __global__ void train_V( float *N, float *Y, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2; // caching chain-rule calculation in vector to aid in backpropagation for ( int i = tid; i < BREADTH; i += stride ) { V[i] = (N[NDIM(q+1,i)] - Y[i]) * df( N[NDIM(q+1,i)]); } } __global__ void train_Wf( float *N, float *W, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2, h; // update weights of outer layer for ( h = tid; h < BREADTH * BREADTH; h += stride ) { int i = h / BREADTH; int j = h % BREADTH; W[WDIM(q,i,j)] -= (float)RATE * V[i] * N[NDIM(q,j)]; } } __global__ void train_WN( float *N, float *W, float *A, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; R[NDIM(j,i)] = A[i] * W[wIndex + h] * df( N[nIndex + j]); } } __global__ void train_B( float *R, float *B) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) B[i] = R[NDIM(i,0)]; } __global__ void train_Wh( float *N, float *W, float *A, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; W[wIndex + h] -= (float)RATE * A[i] * N[nIndex + j]; } } void train() { train_V<<<128, 32>>>( N, Y, V); hipDeviceSynchronize(); train_Wf<<<128, 32>>>( N, W, V); hipDeviceSynchronize(); int q = DEPTH - 2; float *A = &V[0], *B = &U[0]; float *temp; // updating weights of hidden layers (backpropagation) for ( int k = q - 1; k >= 0; k-- ) { // chain-rule expansion by multiplying with next layer's weight matrix and // derivative of activation function on the same layer's nodes train_WN<<<128, 32>>>( N, W, A, R, k+1); hipDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); hipDeviceSynchronize(); j *= 2; n = (n+1)/2; } train_B<<<128, 32>>>( R, B); hipDeviceSynchronize(); temp = B; B = A; A = temp; // adjust weights by gradient descent train_Wh<<<128, 32>>>( N, W, A, k); hipDeviceSynchronize(); } } __global__ void normalize( float *W, float *N, float *Y) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i, w = (DEPTH-1) * BREADTH * BREADTH; for ( i = tid; i < w; i+= stride) W[i] = 1/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) N[i] = i/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) Y[i] = i/ (float)BREADTH; } void initialize() { int w = (DEPTH-1) * BREADTH * BREADTH; // device memory allocation of global arrays hipMalloc( (void**)(&W), sizeof(float) * w); hipMalloc( (void**)(&R), sizeof(float) * BREADTH * BREADTH); hipMalloc( (void**)(&N), sizeof(float) * DEPTH * BREADTH); hipMalloc( (void**)(&Y), sizeof(float) * BREADTH); hipMalloc( (void**)(&pY), sizeof(float) * BREADTH); hipMalloc( (void**)(&U), sizeof(float) * BREADTH); hipMalloc( (void**)(&V), sizeof(float) * BREADTH); /* // random number generator curandGenerator_t gen; // Create pseudo-random number generator curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); // Set seed curandSetPseudoRandomGeneratorSeed(gen, time(0)); // Generate floats on device curandGenerateUniform(gen, W, w); curandGenerateUniform(gen, N, BREADTH * DEPTH); curandGenerateUniform(gen, Y, BREADTH); curandDestroyGenerator(gen); */ normalize<<<128, 32>>>(W, N, Y); hipDeviceSynchronize(); } void calculateError() { float err = 0, dY; // dY : difference b/w expected and predicted output, and // err : sum of squares of dY for ( int i = 0 ; i < BREADTH ; i++ ) { dY = (Dh[i] - Yh[i]); err += dY * dY; } printf("ITERATION : %d\tLOSS : %f\t", count, err); } void freeResources() { hipFree(Y); hipFree(N); hipFree(W); hipFree(pY); hipFree(U); hipFree(V); hipFree(R); } void predict() { predict_N<<<128, 32>>>(N); hipDeviceSynchronize(); for ( int k = 1; k < DEPTH ; k++ ) { predict_WN<<<128, 32>>>( N, W, R, k-1); hipDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); hipDeviceSynchronize(); j *= 2; n = (n+1)/2; } predict_Nk<<<128, 32>>>( N, R, k); hipDeviceSynchronize(); } predict_Yp<<<128, 32>>>( N, pY); hipDeviceSynchronize(); } int main() { initialize(); for ( count = 0; count < 4; count++) { clock_t begin, end; begin = clock(); predict(); train(); end = clock(); float CpuTime = (float)(end - begin) / CLOCKS_PER_SEC; CUDA_CALL(hipMemcpy(Yh, Y, BREADTH * sizeof(float), hipMemcpyDeviceToHost)); CUDA_CALL(hipMemcpy(Dh, pY, BREADTH * sizeof(float), hipMemcpyDeviceToHost)); calculateError(); printf("TIME: %f\n", CpuTime); } freeResources(); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <stdio.h> #include <math.h> #include <stdlib.h> #include <time.h> #include <hip/hip_runtime.h> #include <hiprand/hiprand.h> #define CUDA_CALL(x) do { if((x)!=hipSuccess) { \ printf("Error at %s:%d\n",__FILE__,__LINE__);\ return EXIT_FAILURE;}} while(0) #define DEPTH 5 // no of layers #define BREADTH 1000 // nodes per layer #define RATE 0.0001 // learning rate #define NDIM(a,b) ((a)*BREADTH+(b)) #define WDIM(a,b,c) ((a)*BREADTH*BREADTH+(b)*BREADTH+(c)) float *W, *R; // [DEPTH-1][BREADTH][BREADTH] W_kij : weight of edge b/w jth and ith node from kth layer and k+1th layer resp. float *N; // [DEPTH][BREADTH] N_ki : jth node in kth layer float *Y, *pY; // [BREADTH] expected and predicted output vectors float *U, *V; // [BREADTH] vectors for backpropagation chain-rule float Yh[BREADTH], Dh[BREADTH]; // host arrays: expected and predicted output int count = 0; #define f(x) (tanhf((x))) // activation function #define df(y) (1 - (y) * (y)) // derivative of activation function in y terms __global__ void predict_N( float *N) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; // erase all nodes except first layer for ( i = tid + BREADTH ; i < BREADTH * DEPTH ; i += stride ) N[i] = 0; } __global__ void predict_WN( float *N, float *W, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; R[h] = W[wIndex + h] * N[nIndex + j]; } } __global__ void predict_Ri( float *R, int n, int j) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid % BREADTH ; i < BREADTH ; i += stride ) { int itid = tid / BREADTH; int istride = stride / BREADTH; istride += (istride == 0); int rIndex = i * BREADTH; for ( int h = 2 * j * itid; h < j * (n - 1); h += 2 * j * istride) R[rIndex + h] += R[rIndex + h + j]; } } __global__ void predict_Nk( float *N, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) N[NDIM(k,i)] = f( R[NDIM(i,0)]); } __global__ void predict_Yp( float *N, float *pY) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) pY[i] = N[NDIM(DEPTH-1,i)]; } __global__ void train_V( float *N, float *Y, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2; // caching chain-rule calculation in vector to aid in backpropagation for ( int i = tid; i < BREADTH; i += stride ) { V[i] = (N[NDIM(q+1,i)] - Y[i]) * df( N[NDIM(q+1,i)]); } } __global__ void train_Wf( float *N, float *W, float *V) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int q = DEPTH - 2, h; // update weights of outer layer for ( h = tid; h < BREADTH * BREADTH; h += stride ) { int i = h / BREADTH; int j = h % BREADTH; W[WDIM(q,i,j)] -= (float)RATE * V[i] * N[NDIM(q,j)]; } } __global__ void train_WN( float *N, float *W, float *A, float *R, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; R[NDIM(j,i)] = A[i] * W[wIndex + h] * df( N[nIndex + j]); } } __global__ void train_B( float *R, float *B) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i; for ( i = tid; i < BREADTH; i += stride ) B[i] = R[NDIM(i,0)]; } __global__ void train_Wh( float *N, float *W, float *A, int k) { // feedforward pass algorithm int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int h; // matrix-vector multiplication W*N b/w k-1th and kth layer int wIndex = k * BREADTH * BREADTH; int nIndex = k * BREADTH; for ( h = tid ; h < BREADTH * BREADTH ; h += stride ) { int j = h % BREADTH; int i = h / BREADTH; W[wIndex + h] -= (float)RATE * A[i] * N[nIndex + j]; } } void train() { train_V<<<128, 32>>>( N, Y, V); hipDeviceSynchronize(); train_Wf<<<128, 32>>>( N, W, V); hipDeviceSynchronize(); int q = DEPTH - 2; float *A = &V[0], *B = &U[0]; float *temp; // updating weights of hidden layers (backpropagation) for ( int k = q - 1; k >= 0; k-- ) { // chain-rule expansion by multiplying with next layer's weight matrix and // derivative of activation function on the same layer's nodes train_WN<<<128, 32>>>( N, W, A, R, k+1); hipDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); hipDeviceSynchronize(); j *= 2; n = (n+1)/2; } train_B<<<128, 32>>>( R, B); hipDeviceSynchronize(); temp = B; B = A; A = temp; // adjust weights by gradient descent train_Wh<<<128, 32>>>( N, W, A, k); hipDeviceSynchronize(); } } __global__ void normalize( float *W, float *N, float *Y) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; int i, w = (DEPTH-1) * BREADTH * BREADTH; for ( i = tid; i < w; i+= stride) W[i] = 1/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) N[i] = i/ (float)BREADTH; for ( i = tid; i < BREADTH; i+= stride) Y[i] = i/ (float)BREADTH; } void initialize() { int w = (DEPTH-1) * BREADTH * BREADTH; // device memory allocation of global arrays hipMalloc( (void**)(&W), sizeof(float) * w); hipMalloc( (void**)(&R), sizeof(float) * BREADTH * BREADTH); hipMalloc( (void**)(&N), sizeof(float) * DEPTH * BREADTH); hipMalloc( (void**)(&Y), sizeof(float) * BREADTH); hipMalloc( (void**)(&pY), sizeof(float) * BREADTH); hipMalloc( (void**)(&U), sizeof(float) * BREADTH); hipMalloc( (void**)(&V), sizeof(float) * BREADTH); /* // random number generator curandGenerator_t gen; // Create pseudo-random number generator curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); // Set seed curandSetPseudoRandomGeneratorSeed(gen, time(0)); // Generate floats on device curandGenerateUniform(gen, W, w); curandGenerateUniform(gen, N, BREADTH * DEPTH); curandGenerateUniform(gen, Y, BREADTH); curandDestroyGenerator(gen); */ normalize<<<128, 32>>>(W, N, Y); hipDeviceSynchronize(); } void calculateError() { float err = 0, dY; // dY : difference b/w expected and predicted output, and // err : sum of squares of dY for ( int i = 0 ; i < BREADTH ; i++ ) { dY = (Dh[i] - Yh[i]); err += dY * dY; } printf("ITERATION : %d\tLOSS : %f\t", count, err); } void freeResources() { hipFree(Y); hipFree(N); hipFree(W); hipFree(pY); hipFree(U); hipFree(V); hipFree(R); } void predict() { predict_N<<<128, 32>>>(N); hipDeviceSynchronize(); for ( int k = 1; k < DEPTH ; k++ ) { predict_WN<<<128, 32>>>( N, W, R, k-1); hipDeviceSynchronize(); int j = 1; int n = BREADTH; while ( n > 1) { predict_Ri<<<128, 32>>>( R, n, j); hipDeviceSynchronize(); j *= 2; n = (n+1)/2; } predict_Nk<<<128, 32>>>( N, R, k); hipDeviceSynchronize(); } predict_Yp<<<128, 32>>>( N, pY); hipDeviceSynchronize(); } int main() { initialize(); for ( count = 0; count < 4; count++) { clock_t begin, end; begin = clock(); predict(); train(); end = clock(); float CpuTime = (float)(end - begin) / CLOCKS_PER_SEC; CUDA_CALL(hipMemcpy(Yh, Y, BREADTH * sizeof(float), hipMemcpyDeviceToHost)); CUDA_CALL(hipMemcpy(Dh, pY, BREADTH * sizeof(float), hipMemcpyDeviceToHost)); calculateError(); printf("TIME: %f\n", CpuTime); } freeResources(); return 0; }
.text .file "deep.hip" .globl _Z24__device_stub__predict_NPf # -- Begin function _Z24__device_stub__predict_NPf .p2align 4, 0x90 .type _Z24__device_stub__predict_NPf,@function _Z24__device_stub__predict_NPf: # @_Z24__device_stub__predict_NPf .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movq %rdi, 64(%rsp) leaq 64(%rsp), %rax movq %rax, (%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d movq %rsp, %r9 movl $_Z9predict_NPf, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size _Z24__device_stub__predict_NPf, .Lfunc_end0-_Z24__device_stub__predict_NPf .cfi_endproc # -- End function .globl _Z25__device_stub__predict_WNPfS_S_i # -- Begin function _Z25__device_stub__predict_WNPfS_S_i .p2align 4, 0x90 .type _Z25__device_stub__predict_WNPfS_S_i,@function _Z25__device_stub__predict_WNPfS_S_i: # @_Z25__device_stub__predict_WNPfS_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 $_Z10predict_WNPfS_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__predict_WNPfS_S_i, .Lfunc_end1-_Z25__device_stub__predict_WNPfS_S_i .cfi_endproc # -- End function .globl _Z25__device_stub__predict_RiPfii # -- Begin function _Z25__device_stub__predict_RiPfii .p2align 4, 0x90 .type _Z25__device_stub__predict_RiPfii,@function _Z25__device_stub__predict_RiPfii: # @_Z25__device_stub__predict_RiPfii .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 4(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10predict_RiPfii, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end2: .size _Z25__device_stub__predict_RiPfii, .Lfunc_end2-_Z25__device_stub__predict_RiPfii .cfi_endproc # -- End function .globl _Z25__device_stub__predict_NkPfS_i # -- Begin function _Z25__device_stub__predict_NkPfS_i .p2align 4, 0x90 .type _Z25__device_stub__predict_NkPfS_i,@function _Z25__device_stub__predict_NkPfS_i: # @_Z25__device_stub__predict_NkPfS_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 $_Z10predict_NkPfS_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_end3: .size _Z25__device_stub__predict_NkPfS_i, .Lfunc_end3-_Z25__device_stub__predict_NkPfS_i .cfi_endproc # -- End function .globl _Z25__device_stub__predict_YpPfS_ # -- Begin function _Z25__device_stub__predict_YpPfS_ .p2align 4, 0x90 .type _Z25__device_stub__predict_YpPfS_,@function _Z25__device_stub__predict_YpPfS_: # @_Z25__device_stub__predict_YpPfS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10predict_YpPfS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end4: .size _Z25__device_stub__predict_YpPfS_, .Lfunc_end4-_Z25__device_stub__predict_YpPfS_ .cfi_endproc # -- End function .globl _Z22__device_stub__train_VPfS_S_ # -- Begin function _Z22__device_stub__train_VPfS_S_ .p2align 4, 0x90 .type _Z22__device_stub__train_VPfS_S_,@function _Z22__device_stub__train_VPfS_S_: # @_Z22__device_stub__train_VPfS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7train_VPfS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end5: .size _Z22__device_stub__train_VPfS_S_, .Lfunc_end5-_Z22__device_stub__train_VPfS_S_ .cfi_endproc # -- End function .globl _Z23__device_stub__train_WfPfS_S_ # -- Begin function _Z23__device_stub__train_WfPfS_S_ .p2align 4, 0x90 .type _Z23__device_stub__train_WfPfS_S_,@function _Z23__device_stub__train_WfPfS_S_: # @_Z23__device_stub__train_WfPfS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8train_WfPfS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end6: .size _Z23__device_stub__train_WfPfS_S_, .Lfunc_end6-_Z23__device_stub__train_WfPfS_S_ .cfi_endproc # -- End function .globl _Z23__device_stub__train_WNPfS_S_S_i # -- Begin function _Z23__device_stub__train_WNPfS_S_S_i .p2align 4, 0x90 .type _Z23__device_stub__train_WNPfS_S_S_i,@function _Z23__device_stub__train_WNPfS_S_S_i: # @_Z23__device_stub__train_WNPfS_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 $_Z8train_WNPfS_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_end7: .size _Z23__device_stub__train_WNPfS_S_S_i, .Lfunc_end7-_Z23__device_stub__train_WNPfS_S_S_i .cfi_endproc # -- End function .globl _Z22__device_stub__train_BPfS_ # -- Begin function _Z22__device_stub__train_BPfS_ .p2align 4, 0x90 .type _Z22__device_stub__train_BPfS_,@function _Z22__device_stub__train_BPfS_: # @_Z22__device_stub__train_BPfS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z7train_BPfS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end8: .size _Z22__device_stub__train_BPfS_, .Lfunc_end8-_Z22__device_stub__train_BPfS_ .cfi_endproc # -- End function .globl _Z23__device_stub__train_WhPfS_S_i # -- Begin function _Z23__device_stub__train_WhPfS_S_i .p2align 4, 0x90 .type _Z23__device_stub__train_WhPfS_S_i,@function _Z23__device_stub__train_WhPfS_S_i: # @_Z23__device_stub__train_WhPfS_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 $_Z8train_WhPfS_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_end9: .size _Z23__device_stub__train_WhPfS_S_i, .Lfunc_end9-_Z23__device_stub__train_WhPfS_S_i .cfi_endproc # -- End function .globl _Z5trainv # -- Begin function _Z5trainv .p2align 4, 0x90 .type _Z5trainv,@function _Z5trainv: # @_Z5trainv .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 $168, %rsp .cfi_def_cfa_offset 224 .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 $4294967328, %rbx # imm = 0x100000020 leaq 96(%rbx), %r14 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB10_2 # %bb.1: movq N(%rip), %rax movq Y(%rip), %rcx movq V(%rip), %rdx movq %rax, 48(%rsp) movq %rcx, 40(%rsp) movq %rdx, 32(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rax movq %rax, 88(%rsp) leaq 32(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rdi leaq 64(%rsp), %rsi leaq 8(%rsp), %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z7train_VPfS_S_, %edi pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB10_2: callq hipDeviceSynchronize movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB10_4 # %bb.3: movq N(%rip), %rax movq W(%rip), %rcx movq V(%rip), %rdx movq %rax, 48(%rsp) movq %rcx, 40(%rsp) movq %rdx, 32(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rax movq %rax, 88(%rsp) leaq 32(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rdi leaq 64(%rsp), %rsi leaq 8(%rsp), %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8train_WfPfS_S_, %edi pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB10_4: callq hipDeviceSynchronize movq V(%rip), %rax movl $2, %r13d movq U(%rip), %rcx leaq 64(%rsp), %r15 leaq 80(%rsp), %r12 jmp .LBB10_5 .p2align 4, 0x90 .LBB10_15: # in Loop: Header=BB10_5 Depth=1 callq hipDeviceSynchronize movq %rbp, %rax movq 144(%rsp), %rcx # 8-byte Reload subl $1, %r13d jb .LBB10_16 .LBB10_5: # =>This Loop Header: Depth=1 # Child Loop BB10_8 Depth 2 movq %rcx, 152(%rsp) # 8-byte Spill movq %rax, %rbp movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB10_7 # %bb.6: # in Loop: Header=BB10_5 Depth=1 movq N(%rip), %rax movq W(%rip), %rcx movq R(%rip), %rdx leal 1(%r13), %esi movq %rax, 48(%rsp) movq %rcx, 40(%rsp) movq %rbp, 32(%rsp) movq %rdx, 8(%rsp) movl %esi, 140(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rax movq %rax, 88(%rsp) leaq 32(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 140(%rsp), %rax movq %rax, 112(%rsp) leaq 16(%rsp), %rdi movq %r15, %rsi leaq 56(%rsp), %rdx leaq 128(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z8train_WNPfS_S_S_i, %edi movq %r12, %r9 pushq 128(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB10_7: # in Loop: Header=BB10_5 Depth=1 movq %rbp, 144(%rsp) # 8-byte Spill movq %r13, 160(%rsp) # 8-byte Spill callq hipDeviceSynchronize movl $1000, %ebp # imm = 0x3E8 movl $1, %r13d jmp .LBB10_8 .p2align 4, 0x90 .LBB10_10: # in Loop: Header=BB10_8 Depth=2 callq hipDeviceSynchronize addl %r13d, %r13d leal 1(%rbp), %eax shrl %eax cmpl $2, %ebp movl %eax, %ebp jbe .LBB10_11 .LBB10_8: # Parent Loop BB10_5 Depth=1 # => This Inner Loop Header: Depth=2 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB10_10 # %bb.9: # in Loop: Header=BB10_8 Depth=2 movq R(%rip), %rax movq %rax, 48(%rsp) movl %ebp, 8(%rsp) movl %r13d, 56(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rdi movq %r15, %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z10predict_RiPfii, %edi movq %r12, %r9 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 pushq 48(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB10_10 .p2align 4, 0x90 .LBB10_11: # in Loop: Header=BB10_5 Depth=1 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax movq 152(%rsp), %rbp # 8-byte Reload jne .LBB10_13 # %bb.12: # in Loop: Header=BB10_5 Depth=1 movq R(%rip), %rax movq %rax, 48(%rsp) movq %rbp, 40(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rax movq %rax, 88(%rsp) leaq 16(%rsp), %rdi movq %r15, %rsi leaq 32(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z7train_BPfS_, %edi movq %r12, %r9 pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB10_13: # in Loop: Header=BB10_5 Depth=1 callq hipDeviceSynchronize movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax movq 160(%rsp), %r13 # 8-byte Reload jne .LBB10_15 # %bb.14: # in Loop: Header=BB10_5 Depth=1 movq N(%rip), %rax movq W(%rip), %rcx movq %rax, 48(%rsp) movq %rcx, 40(%rsp) movq %rbp, 32(%rsp) movl %r13d, 128(%rsp) leaq 48(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rax movq %rax, 88(%rsp) leaq 32(%rsp), %rax movq %rax, 96(%rsp) leaq 128(%rsp), %rax movq %rax, 104(%rsp) leaq 16(%rsp), %rdi movq %r15, %rsi leaq 8(%rsp), %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 64(%rsp), %rcx movl 72(%rsp), %r8d movl $_Z8train_WhPfS_S_i, %edi movq %r12, %r9 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB10_15 .LBB10_16: addq $168, %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_end10: .size _Z5trainv, .Lfunc_end10-_Z5trainv .cfi_endproc # -- End function .globl _Z24__device_stub__normalizePfS_S_ # -- Begin function _Z24__device_stub__normalizePfS_S_ .p2align 4, 0x90 .type _Z24__device_stub__normalizePfS_S_,@function _Z24__device_stub__normalizePfS_S_: # @_Z24__device_stub__normalizePfS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z9normalizePfS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end11: .size _Z24__device_stub__normalizePfS_S_, .Lfunc_end11-_Z24__device_stub__normalizePfS_S_ .cfi_endproc # -- End function .globl _Z10initializev # -- Begin function _Z10initializev .p2align 4, 0x90 .type _Z10initializev,@function _Z10initializev: # @_Z10initializev .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movl $W, %edi movl $16000000, %esi # imm = 0xF42400 callq hipMalloc movl $R, %edi movl $4000000, %esi # imm = 0x3D0900 callq hipMalloc movl $N, %edi movl $20000, %esi # imm = 0x4E20 callq hipMalloc movl $Y, %edi movl $4000, %esi # imm = 0xFA0 callq hipMalloc movl $pY, %edi movl $4000, %esi # imm = 0xFA0 callq hipMalloc movl $U, %edi movl $4000, %esi # imm = 0xFA0 callq hipMalloc movl $V, %edi movl $4000, %esi # imm = 0xFA0 callq hipMalloc movabsq $4294967328, %rdx # imm = 0x100000020 leaq 96(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB12_2 # %bb.1: movq W(%rip), %rax movq N(%rip), %rcx movq Y(%rip), %rdx movq %rax, 72(%rsp) movq %rcx, 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 $_Z9normalizePfS_S_, %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 .LBB12_2: callq hipDeviceSynchronize addq $104, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end12: .size _Z10initializev, .Lfunc_end12-_Z10initializev .cfi_endproc # -- End function .globl _Z14calculateErrorv # -- Begin function _Z14calculateErrorv .p2align 4, 0x90 .type _Z14calculateErrorv,@function _Z14calculateErrorv: # @_Z14calculateErrorv .cfi_startproc # %bb.0: xorps %xmm0, %xmm0 movq $-4000, %rax # imm = 0xF060 .p2align 4, 0x90 .LBB13_1: # =>This Inner Loop Header: Depth=1 movss Dh+4000(%rax), %xmm1 # xmm1 = mem[0],zero,zero,zero subss Yh+4000(%rax), %xmm1 mulss %xmm1, %xmm1 addss %xmm1, %xmm0 addq $4, %rax jne .LBB13_1 # %bb.2: movl count(%rip), %esi cvtss2sd %xmm0, %xmm0 movl $.L.str, %edi movb $1, %al jmp printf # TAILCALL .Lfunc_end13: .size _Z14calculateErrorv, .Lfunc_end13-_Z14calculateErrorv .cfi_endproc # -- End function .globl _Z13freeResourcesv # -- Begin function _Z13freeResourcesv .p2align 4, 0x90 .type _Z13freeResourcesv,@function _Z13freeResourcesv: # @_Z13freeResourcesv .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 movq Y(%rip), %rdi callq hipFree movq N(%rip), %rdi callq hipFree movq W(%rip), %rdi callq hipFree movq pY(%rip), %rdi callq hipFree movq U(%rip), %rdi callq hipFree movq V(%rip), %rdi callq hipFree movq R(%rip), %rdi popq %rax .cfi_def_cfa_offset 8 jmp hipFree # TAILCALL .Lfunc_end14: .size _Z13freeResourcesv, .Lfunc_end14-_Z13freeResourcesv .cfi_endproc # -- End function .globl _Z7predictv # -- Begin function _Z7predictv .p2align 4, 0x90 .type _Z7predictv,@function _Z7predictv: # @_Z7predictv .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $120, %rsp .cfi_def_cfa_offset 176 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movabsq $4294967328, %rbx # imm = 0x100000020 leaq 96(%rbx), %r14 movl $1, %r15d movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB15_2 # %bb.1: leaq 40(%rsp), %rax movq N(%rip), %rcx movq %rcx, 40(%rsp) movq %rax, (%rsp) leaq 64(%rsp), %rdi leaq 8(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq 8(%rsp), %rcx movl 16(%rsp), %r8d movq %rsp, %r9 movl $_Z9predict_NPf, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB15_2: callq hipDeviceSynchronize leaq 96(%rsp), %r12 leaq 64(%rsp), %r13 jmp .LBB15_3 .p2align 4, 0x90 .LBB15_9: # in Loop: Header=BB15_3 Depth=1 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax movq 112(%rsp), %r15 # 8-byte Reload je .LBB15_10 .LBB15_11: # in Loop: Header=BB15_3 Depth=1 callq hipDeviceSynchronize incl %r15d cmpl $5, %r15d je .LBB15_12 .LBB15_3: # =>This Loop Header: Depth=1 # Child Loop BB15_6 Depth 2 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB15_5 # %bb.4: # in Loop: Header=BB15_3 Depth=1 movq N(%rip), %rax movq W(%rip), %rcx movq R(%rip), %rdx leal -1(%r15), %esi movq %rax, 32(%rsp) movq %rcx, 24(%rsp) movq %rdx, (%rsp) movl %esi, 108(%rsp) leaq 32(%rsp), %rax movq %rax, 64(%rsp) leaq 24(%rsp), %rax movq %rax, 72(%rsp) movq %rsp, %rax movq %rax, 80(%rsp) leaq 108(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rdi leaq 40(%rsp), %rsi leaq 56(%rsp), %rdx movq %r12, %rcx callq __hipPopCallConfiguration movq 8(%rsp), %rsi movl 16(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z10predict_WNPfS_S_i, %edi movq %r13, %r9 pushq 96(%rsp) .cfi_adjust_cfa_offset 8 pushq 64(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB15_5: # in Loop: Header=BB15_3 Depth=1 movq %r15, 112(%rsp) # 8-byte Spill callq hipDeviceSynchronize movl $1000, %r15d # imm = 0x3E8 movl $1, %ebp jmp .LBB15_6 .p2align 4, 0x90 .LBB15_8: # in Loop: Header=BB15_6 Depth=2 callq hipDeviceSynchronize addl %ebp, %ebp leal 1(%r15), %eax shrl %eax cmpl $2, %r15d movl %eax, %r15d jbe .LBB15_9 .LBB15_6: # Parent Loop BB15_3 Depth=1 # => This Inner Loop Header: Depth=2 movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB15_8 # %bb.7: # in Loop: Header=BB15_6 Depth=2 movq R(%rip), %rax movq %rax, 32(%rsp) movl %r15d, 56(%rsp) movl %ebp, 96(%rsp) leaq 32(%rsp), %rax movq %rax, 64(%rsp) leaq 56(%rsp), %rax movq %rax, 72(%rsp) movq %r12, 80(%rsp) leaq 8(%rsp), %rdi leaq 40(%rsp), %rsi leaq 24(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 8(%rsp), %rsi movl 16(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z10predict_RiPfii, %edi movq %r13, %r9 pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB15_8 .p2align 4, 0x90 .LBB15_10: # in Loop: Header=BB15_3 Depth=1 movq N(%rip), %rax movq R(%rip), %rcx movq %rax, 32(%rsp) movq %rcx, 24(%rsp) movl %r15d, 96(%rsp) leaq 32(%rsp), %rax movq %rax, 64(%rsp) leaq 24(%rsp), %rax movq %rax, 72(%rsp) movq %r12, 80(%rsp) leaq 8(%rsp), %rdi leaq 40(%rsp), %rsi movq %rsp, %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 8(%rsp), %rsi movl 16(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d movl $_Z10predict_NkPfS_i, %edi movq %r13, %r9 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 8(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 jmp .LBB15_11 .LBB15_12: movq %r14, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB15_14 # %bb.13: movq N(%rip), %rax movq pY(%rip), %rcx movq %rax, 32(%rsp) movq %rcx, 24(%rsp) leaq 32(%rsp), %rax movq %rax, 64(%rsp) leaq 24(%rsp), %rax movq %rax, 72(%rsp) leaq 8(%rsp), %rdi leaq 40(%rsp), %rsi movq %rsp, %rdx leaq 56(%rsp), %rcx callq __hipPopCallConfiguration movq 8(%rsp), %rsi movl 16(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z10predict_YpPfS_, %edi pushq 56(%rsp) .cfi_adjust_cfa_offset 8 pushq 8(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB15_14: callq hipDeviceSynchronize addq $120, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end15: .size _Z7predictv, .Lfunc_end15-_Z7predictv .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function main .LCPI16_0: .long 0x49742400 # float 1.0E+6 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $16, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 callq _Z10initializev movl $0, count(%rip) xorl %ebx, %ebx .p2align 4, 0x90 .LBB16_1: # =>This Loop Header: Depth=1 # Child Loop BB16_6 Depth 2 callq clock movq %rax, %r14 callq _Z7predictv callq _Z5trainv callq clock subq %r14, %rax xorps %xmm0, %xmm0 cvtsi2ss %rax, %xmm0 divss .LCPI16_0(%rip), %xmm0 movss %xmm0, 12(%rsp) # 4-byte Spill movq Y(%rip), %rsi movl $Yh, %edi movl $4000, %edx # imm = 0xFA0 movl $2, %ecx callq hipMemcpy testl %eax, %eax je .LBB16_4 # %bb.2: # in Loop: Header=BB16_1 Depth=1 xorl %ebp, %ebp movl $.L.str.1, %edi movl $.L.str.2, %esi movl $318, %edx # imm = 0x13E jmp .LBB16_3 .p2align 4, 0x90 .LBB16_4: # in Loop: Header=BB16_1 Depth=1 movq pY(%rip), %rsi movl $Dh, %edi movl $4000, %edx # imm = 0xFA0 movl $2, %ecx callq hipMemcpy testl %eax, %eax je .LBB16_5 # %bb.12: # in Loop: Header=BB16_1 Depth=1 xorl %ebp, %ebp movl $.L.str.1, %edi movl $.L.str.2, %esi movl $319, %edx # imm = 0x13F .LBB16_3: # in Loop: Header=BB16_1 Depth=1 xorl %eax, %eax callq printf movl $1, %ebx testb %bpl, %bpl je .LBB16_11 .LBB16_9: # in Loop: Header=BB16_1 Depth=1 movl count(%rip), %eax leal 1(%rax), %ecx movl %ecx, count(%rip) cmpl $3, %eax jl .LBB16_1 jmp .LBB16_10 .LBB16_5: # %.preheader.preheader # in Loop: Header=BB16_1 Depth=1 xorps %xmm0, %xmm0 movq $-4000, %rax # imm = 0xF060 .p2align 4, 0x90 .LBB16_6: # %.preheader # Parent Loop BB16_1 Depth=1 # => This Inner Loop Header: Depth=2 movss Dh+4000(%rax), %xmm1 # xmm1 = mem[0],zero,zero,zero subss Yh+4000(%rax), %xmm1 mulss %xmm1, %xmm1 addss %xmm1, %xmm0 addq $4, %rax jne .LBB16_6 # %bb.7: # %_Z14calculateErrorv.exit # in Loop: Header=BB16_1 Depth=1 movl count(%rip), %esi cvtss2sd %xmm0, %xmm0 movb $1, %bpl movl $.L.str, %edi movb $1, %al callq printf movss 12(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str.3, %edi movb $1, %al callq printf testb %bpl, %bpl jne .LBB16_9 jmp .LBB16_11 .LBB16_10: movq Y(%rip), %rdi callq hipFree movq N(%rip), %rdi callq hipFree movq W(%rip), %rdi callq hipFree movq pY(%rip), %rdi callq hipFree movq U(%rip), %rdi callq hipFree movq V(%rip), %rdi callq hipFree movq R(%rip), %rdi callq hipFree xorl %ebx, %ebx .LBB16_11: # %.loopexit movl %ebx, %eax addq $16, %rsp .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end16: .size main, .Lfunc_end16-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 .LBB17_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB17_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z9predict_NPf, %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 $_Z10predict_WNPfS_S_i, %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 $_Z10predict_RiPfii, %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 $_Z10predict_NkPfS_i, %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 $_Z10predict_YpPfS_, %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 $_Z7train_VPfS_S_, %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 $_Z8train_WfPfS_S_, %esi movl $.L__unnamed_7, %edx movl $.L__unnamed_7, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8train_WNPfS_S_S_i, %esi movl $.L__unnamed_8, %edx movl $.L__unnamed_8, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7train_BPfS_, %esi movl $.L__unnamed_9, %edx movl $.L__unnamed_9, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8train_WhPfS_S_i, %esi movl $.L__unnamed_10, %edx movl $.L__unnamed_10, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z9normalizePfS_S_, %esi movl $.L__unnamed_11, %edx movl $.L__unnamed_11, %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_end17: .size __hip_module_ctor, .Lfunc_end17-__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 .LBB18_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 .LBB18_2: retq .Lfunc_end18: .size __hip_module_dtor, .Lfunc_end18-__hip_module_dtor .cfi_endproc # -- End function .type W,@object # @W .bss .globl W .p2align 3, 0x0 W: .quad 0 .size W, 8 .type R,@object # @R .globl R .p2align 3, 0x0 R: .quad 0 .size R, 8 .type N,@object # @N .globl N .p2align 3, 0x0 N: .quad 0 .size N, 8 .type Y,@object # @Y .globl Y .p2align 3, 0x0 Y: .quad 0 .size Y, 8 .type pY,@object # @pY .globl pY .p2align 3, 0x0 pY: .quad 0 .size pY, 8 .type U,@object # @U .globl U .p2align 3, 0x0 U: .quad 0 .size U, 8 .type V,@object # @V .globl V .p2align 3, 0x0 V: .quad 0 .size V, 8 .type Yh,@object # @Yh .globl Yh .p2align 4, 0x0 Yh: .zero 4000 .size Yh, 4000 .type Dh,@object # @Dh .globl Dh .p2align 4, 0x0 Dh: .zero 4000 .size Dh, 4000 .type count,@object # @count .globl count .p2align 2, 0x0 count: .long 0 # 0x0 .size count, 4 .type _Z9predict_NPf,@object # @_Z9predict_NPf .section .rodata,"a",@progbits .globl _Z9predict_NPf .p2align 3, 0x0 _Z9predict_NPf: .quad _Z24__device_stub__predict_NPf .size _Z9predict_NPf, 8 .type _Z10predict_WNPfS_S_i,@object # @_Z10predict_WNPfS_S_i .globl _Z10predict_WNPfS_S_i .p2align 3, 0x0 _Z10predict_WNPfS_S_i: .quad _Z25__device_stub__predict_WNPfS_S_i .size _Z10predict_WNPfS_S_i, 8 .type _Z10predict_RiPfii,@object # @_Z10predict_RiPfii .globl _Z10predict_RiPfii .p2align 3, 0x0 _Z10predict_RiPfii: .quad _Z25__device_stub__predict_RiPfii .size _Z10predict_RiPfii, 8 .type _Z10predict_NkPfS_i,@object # @_Z10predict_NkPfS_i .globl _Z10predict_NkPfS_i .p2align 3, 0x0 _Z10predict_NkPfS_i: .quad _Z25__device_stub__predict_NkPfS_i .size _Z10predict_NkPfS_i, 8 .type _Z10predict_YpPfS_,@object # @_Z10predict_YpPfS_ .globl _Z10predict_YpPfS_ .p2align 3, 0x0 _Z10predict_YpPfS_: .quad _Z25__device_stub__predict_YpPfS_ .size _Z10predict_YpPfS_, 8 .type _Z7train_VPfS_S_,@object # @_Z7train_VPfS_S_ .globl _Z7train_VPfS_S_ .p2align 3, 0x0 _Z7train_VPfS_S_: .quad _Z22__device_stub__train_VPfS_S_ .size _Z7train_VPfS_S_, 8 .type _Z8train_WfPfS_S_,@object # @_Z8train_WfPfS_S_ .globl _Z8train_WfPfS_S_ .p2align 3, 0x0 _Z8train_WfPfS_S_: .quad _Z23__device_stub__train_WfPfS_S_ .size _Z8train_WfPfS_S_, 8 .type _Z8train_WNPfS_S_S_i,@object # @_Z8train_WNPfS_S_S_i .globl _Z8train_WNPfS_S_S_i .p2align 3, 0x0 _Z8train_WNPfS_S_S_i: .quad _Z23__device_stub__train_WNPfS_S_S_i .size _Z8train_WNPfS_S_S_i, 8 .type _Z7train_BPfS_,@object # @_Z7train_BPfS_ .globl _Z7train_BPfS_ .p2align 3, 0x0 _Z7train_BPfS_: .quad _Z22__device_stub__train_BPfS_ .size _Z7train_BPfS_, 8 .type _Z8train_WhPfS_S_i,@object # @_Z8train_WhPfS_S_i .globl _Z8train_WhPfS_S_i .p2align 3, 0x0 _Z8train_WhPfS_S_i: .quad _Z23__device_stub__train_WhPfS_S_i .size _Z8train_WhPfS_S_i, 8 .type _Z9normalizePfS_S_,@object # @_Z9normalizePfS_S_ .globl _Z9normalizePfS_S_ .p2align 3, 0x0 _Z9normalizePfS_S_: .quad _Z24__device_stub__normalizePfS_S_ .size _Z9normalizePfS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "ITERATION : %d\tLOSS : %f\t" .size .L.str, 26 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Error at %s:%d\n" .size .L.str.1, 16 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/ShamsArfeen/Deep-Learning-with-CUDA/main/deep.hip" .size .L.str.2, 107 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "TIME: %f\n" .size .L.str.3, 10 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z9predict_NPf" .size .L__unnamed_1, 15 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z10predict_WNPfS_S_i" .size .L__unnamed_2, 22 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z10predict_RiPfii" .size .L__unnamed_3, 19 .type .L__unnamed_4,@object # @3 .L__unnamed_4: .asciz "_Z10predict_NkPfS_i" .size .L__unnamed_4, 20 .type .L__unnamed_5,@object # @4 .L__unnamed_5: .asciz "_Z10predict_YpPfS_" .size .L__unnamed_5, 19 .type .L__unnamed_6,@object # @5 .L__unnamed_6: .asciz "_Z7train_VPfS_S_" .size .L__unnamed_6, 17 .type .L__unnamed_7,@object # @6 .L__unnamed_7: .asciz "_Z8train_WfPfS_S_" .size .L__unnamed_7, 18 .type .L__unnamed_8,@object # @7 .L__unnamed_8: .asciz "_Z8train_WNPfS_S_S_i" .size .L__unnamed_8, 21 .type .L__unnamed_9,@object # @8 .L__unnamed_9: .asciz "_Z7train_BPfS_" .size .L__unnamed_9, 15 .type .L__unnamed_10,@object # @9 .L__unnamed_10: .asciz "_Z8train_WhPfS_S_i" .size .L__unnamed_10, 19 .type .L__unnamed_11,@object # @10 .L__unnamed_11: .asciz "_Z9normalizePfS_S_" .size .L__unnamed_11, 19 .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 _Z24__device_stub__predict_NPf .addrsig_sym _Z25__device_stub__predict_WNPfS_S_i .addrsig_sym _Z25__device_stub__predict_RiPfii .addrsig_sym _Z25__device_stub__predict_NkPfS_i .addrsig_sym _Z25__device_stub__predict_YpPfS_ .addrsig_sym _Z22__device_stub__train_VPfS_S_ .addrsig_sym _Z23__device_stub__train_WfPfS_S_ .addrsig_sym _Z23__device_stub__train_WNPfS_S_S_i .addrsig_sym _Z22__device_stub__train_BPfS_ .addrsig_sym _Z23__device_stub__train_WhPfS_S_i .addrsig_sym _Z24__device_stub__normalizePfS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym W .addrsig_sym R .addrsig_sym N .addrsig_sym Y .addrsig_sym pY .addrsig_sym U .addrsig_sym V .addrsig_sym Yh .addrsig_sym Dh .addrsig_sym _Z9predict_NPf .addrsig_sym _Z10predict_WNPfS_S_i .addrsig_sym _Z10predict_RiPfii .addrsig_sym _Z10predict_NkPfS_i .addrsig_sym _Z10predict_YpPfS_ .addrsig_sym _Z7train_VPfS_S_ .addrsig_sym _Z8train_WfPfS_S_ .addrsig_sym _Z8train_WNPfS_S_S_i .addrsig_sym _Z7train_BPfS_ .addrsig_sym _Z8train_WhPfS_S_i .addrsig_sym _Z9normalizePfS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <cufft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { cufftHandle plan; cufftPlan1d(&plan, n, CUFFT_C2C, 1); cufftSetStream(plan, (cudaStream_t)stream); cufftExecC2C(plan, (cufftComplex*)d_data, (cufftComplex*)d_data,CUFFT_FORWARD); cufftDestroy(plan); }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cufft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { cufftHandle plan; cufftPlan1d(&plan, n, CUFFT_C2C, 1); cufftSetStream(plan, (cudaStream_t)stream); cufftExecC2C(plan, (cufftComplex*)d_data, (cufftComplex*)d_data,CUFFT_FORWARD); cufftDestroy(plan); }
.file "tmpxft_000e27f5_00000000-6_cufft.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2053: .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 .LFE2053: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl launchCUFFT .type launchCUFFT, @function launchCUFFT: .LFB2050: .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 $24, %rsp .cfi_def_cfa_offset 48 movq %rdi, %rbx movq %rdx, %rbp movq %fs:40, %rax movq %rax, 8(%rsp) xorl %eax, %eax leaq 4(%rsp), %rdi movl $1, %ecx movl $41, %edx call cufftPlan1d@PLT movq %rbp, %rsi movl 4(%rsp), %edi call cufftSetStream@PLT movl $-1, %ecx movq %rbx, %rdx movq %rbx, %rsi movl 4(%rsp), %edi call cufftExecC2C@PLT movl 4(%rsp), %edi call cufftDestroy@PLT movq 8(%rsp), %rax subq %fs:40, %rax jne .L6 addq $24, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2050: .size launchCUFFT, .-launchCUFFT .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2076: .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 .LFE2076: .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 <cufft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { cufftHandle plan; cufftPlan1d(&plan, n, CUFFT_C2C, 1); cufftSetStream(plan, (cudaStream_t)stream); cufftExecC2C(plan, (cufftComplex*)d_data, (cufftComplex*)d_data,CUFFT_FORWARD); cufftDestroy(plan); }
#include <hip/hip_runtime.h> #include <hipfft/hipfft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { hipfftHandle plan; hipfftPlan1d(&plan, n, HIPFFT_C2C, 1); hipfftSetStream(plan, (hipStream_t)stream); hipfftExecC2C(plan, (hipfftComplex*)d_data, (hipfftComplex*)d_data,HIPFFT_FORWARD); hipfftDestroy(plan); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <hipfft/hipfft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { hipfftHandle plan; hipfftPlan1d(&plan, n, HIPFFT_C2C, 1); hipfftSetStream(plan, (hipStream_t)stream); hipfftExecC2C(plan, (hipfftComplex*)d_data, (hipfftComplex*)d_data,HIPFFT_FORWARD); hipfftDestroy(plan); }
.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 <hipfft/hipfft.h> // Declared extern "C" to disable C++ name mangling extern "C" void launchCUFFT(float *d_data, int n, void *stream) { hipfftHandle plan; hipfftPlan1d(&plan, n, HIPFFT_C2C, 1); hipfftSetStream(plan, (hipStream_t)stream); hipfftExecC2C(plan, (hipfftComplex*)d_data, (hipfftComplex*)d_data,HIPFFT_FORWARD); hipfftDestroy(plan); }
.text .file "cufft.hip" .globl launchCUFFT # -- Begin function launchCUFFT .p2align 4, 0x90 .type launchCUFFT,@function launchCUFFT: # @launchCUFFT .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movq %rdx, %rbx movq %rdi, %r14 movq %rsp, %rdi movl $41, %edx movl $1, %ecx callq hipfftPlan1d movq (%rsp), %rdi movq %rbx, %rsi callq hipfftSetStream movq (%rsp), %rdi movq %r14, %rsi movq %r14, %rdx movl $-1, %ecx callq hipfftExecC2C movq (%rsp), %rdi callq hipfftDestroy addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size launchCUFFT, .Lfunc_end0-launchCUFFT .cfi_endproc # -- End function .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_000e27f5_00000000-6_cufft.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2053: .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 .LFE2053: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl launchCUFFT .type launchCUFFT, @function launchCUFFT: .LFB2050: .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 $24, %rsp .cfi_def_cfa_offset 48 movq %rdi, %rbx movq %rdx, %rbp movq %fs:40, %rax movq %rax, 8(%rsp) xorl %eax, %eax leaq 4(%rsp), %rdi movl $1, %ecx movl $41, %edx call cufftPlan1d@PLT movq %rbp, %rsi movl 4(%rsp), %edi call cufftSetStream@PLT movl $-1, %ecx movq %rbx, %rdx movq %rbx, %rsi movl 4(%rsp), %edi call cufftExecC2C@PLT movl 4(%rsp), %edi call cufftDestroy@PLT movq 8(%rsp), %rax subq %fs:40, %rax jne .L6 addq $24, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2050: .size launchCUFFT, .-launchCUFFT .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2076: .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 .LFE2076: .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 "cufft.hip" .globl launchCUFFT # -- Begin function launchCUFFT .p2align 4, 0x90 .type launchCUFFT,@function launchCUFFT: # @launchCUFFT .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 pushq %rax .cfi_def_cfa_offset 32 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 movq %rdx, %rbx movq %rdi, %r14 movq %rsp, %rdi movl $41, %edx movl $1, %ecx callq hipfftPlan1d movq (%rsp), %rdi movq %rbx, %rsi callq hipfftSetStream movq (%rsp), %rdi movq %r14, %rsi movq %r14, %rdx movl $-1, %ecx callq hipfftExecC2C movq (%rsp), %rdi callq hipfftDestroy addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size launchCUFFT, .Lfunc_end0-launchCUFFT .cfi_endproc # -- End function .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 <cuda.h> int main(void) { cudaDeviceProp prop;//cudaDevieProp is a structure, prop is a structure variable. int count; cudaGetDeviceCount (&count); for (int i=0; i< count; i++){ cudaGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } return 0; }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> //#include <cuda.h> int main(void) { cudaDeviceProp prop;//cudaDevieProp is a structure, prop is a structure variable. int count; cudaGetDeviceCount (&count); for (int i=0; i< count; i++){ cudaGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } return 0; }
.file "tmpxft_0018eaaa_00000000-6_page32_gpuinfo.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "\n****************************Device %d*************************************\n" .align 8 .LC1: .string "----------General Information----------\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC2: .string "Name: %s\n" .LC3: .string "Compute capability: %d.%d\n" .LC4: .string "Clock rate: %d MHz\n" .LC5: .string "GPU type: " .LC6: .string "Integrated GPU\n" .LC7: .string "Discreted GPU\n" .LC8: .string "Device copy overlap: " .LC9: .string "Enabled\n" .LC10: .string "Disabled\n" .LC11: .string "Kernel execution timeout: " .section .rodata.str1.8 .align 8 .LC12: .string "Mapping Host Memory into CUDA Device Address Space: " .align 8 .LC13: .string "Executing Multiple Kernels Support: " .section .rodata.str1.1 .LC14: .string "Device Computing Mode: " .LC15: .string "Default\n" .LC16: .string "Exclusive\n" .LC17: .string "Prohibited\n" .section .rodata.str1.8 .align 8 .LC18: .string "\n----------Memory Information for device----------\n" .section .rodata.str1.1 .LC19: .string "Total Global Memory: %ld MB\n" .section .rodata.str1.8 .align 8 .LC20: .string "Total Constant Memory: %ld kB\n" .align 8 .LC21: .string "Max Memory Pitch Allowed for Memory Copies: %ld MB\n" .section .rodata.str1.1 .LC22: .string "Texture Alignment: %ld B\n" .section .rodata.str1.8 .align 8 .LC23: .string "Max Size Supported for 1D Textures: %ld\n" .align 8 .LC24: .string "Max Dimensions Supported for 2D Texture: (%d, %d)\n" .align 8 .LC25: .string "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n" .align 8 .LC26: .string "\n----------Thread Information for device----------\n" .align 8 .LC27: .string "Streaming Multiprocessor count: %d\n" .align 8 .LC28: .string "Shared Memory per Block: %ld KB\n" .section .rodata.str1.1 .LC29: .string "Registers per Block: %ld\n" .LC30: .string "Threads in warp: %ld\n" .LC31: .string "Max threads per block: %d\n" .section .rodata.str1.8 .align 8 .LC32: .string "Max thread dimensions: (%d, %d, %d)\n" .align 8 .LC33: .string "Max grid dimensions: (%d, %d, %d)\n" .align 8 .LC34: .string "*****************************************************************\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $1064, %rsp .cfi_def_cfa_offset 1120 movq %fs:40, %rax movq %rax, 1048(%rsp) xorl %eax, %eax leaq 12(%rsp), %rdi call cudaGetDeviceCount@PLT cmpl $0, 12(%rsp) jle .L4 movl $0, %ebx leaq .LC0(%rip), %r14 leaq .LC1(%rip), %r13 leaq .LC2(%rip), %r12 leaq .LC3(%rip), %rbp jmp .L18 .L5: leaq .LC7(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L6 .L7: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L8 .L9: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L10 .L11: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L12 .L13: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L14: leaq .LC14(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 420(%rsp) je .L22 .L15: cmpl $1, 420(%rsp) je .L23 .L16: cmpl $2, 420(%rsp) je .L24 .L17: leaq .LC18(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 304(%rsp), %rdx shrq $20, %rdx leaq .LC19(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 368(%rsp), %rdx shrq $10, %rdx leaq .LC20(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 328(%rsp), %rdx shrq $20, %rdx leaq .LC21(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 384(%rsp), %rdx leaq .LC22(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 424(%rsp), %edx leaq .LC23(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 440(%rsp), %ecx movl 436(%rsp), %edx leaq .LC24(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 480(%rsp), %r8d movl 476(%rsp), %ecx movl 472(%rsp), %edx leaq .LC25(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq .LC26(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 404(%rsp), %edx leaq .LC27(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 312(%rsp), %rdx shrq $10, %rdx leaq .LC28(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 320(%rsp), %edx leaq .LC29(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 324(%rsp), %edx leaq .LC30(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 336(%rsp), %edx leaq .LC31(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 348(%rsp), %r8d movl 344(%rsp), %ecx movl 340(%rsp), %edx leaq .LC32(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 360(%rsp), %r8d movl 356(%rsp), %ecx movl 352(%rsp), %edx leaq .LC33(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq .LC34(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %ebx cmpl %ebx, 12(%rsp) jle .L4 .L18: leaq 16(%rsp), %r15 movl %ebx, %esi movq %r15, %rdi call cudaGetDeviceProperties_v2@PLT movl %ebx, %edx movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r15, %rdx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 380(%rsp), %ecx movl 376(%rsp), %edx movq %rbp, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 364(%rsp), %eax movslq %eax, %rdx imulq $274877907, %rdx, %rdx sarq $38, %rdx sarl $31, %eax subl %eax, %edx leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 412(%rsp) je .L5 leaq .LC6(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L6: leaq .LC8(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 400(%rsp) je .L7 leaq .LC9(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L8: leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 408(%rsp) je .L9 leaq .LC9(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L10: leaq .LC12(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 416(%rsp) je .L11 leaq .LC9(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L12: leaq .LC13(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT cmpl $0, 592(%rsp) je .L13 leaq .LC9(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L14 .L22: leaq .LC15(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L15 .L23: leaq .LC16(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L16 .L24: leaq .LC17(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L17 .L4: movq 1048(%rsp), %rax subq %fs:40, %rax jne .L25 movl $0, %eax addq $1064, %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 call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> //#include <cuda.h> int main(void) { cudaDeviceProp prop;//cudaDevieProp is a structure, prop is a structure variable. int count; cudaGetDeviceCount (&count); for (int i=0; i< count; i++){ cudaGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> //#include <cuda.h> int main(void) { hipDeviceProp_t prop;//cudaDevieProp is a structure, prop is a structure variable. int count; hipGetDeviceCount (&count); for (int i=0; i< count; i++){ hipGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } 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 <cuda.h> int main(void) { hipDeviceProp_t prop;//cudaDevieProp is a structure, prop is a structure variable. int count; hipGetDeviceCount (&count); for (int i=0; i< count; i++){ hipGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } return 0; }
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> //#include <cuda.h> int main(void) { hipDeviceProp_t prop;//cudaDevieProp is a structure, prop is a structure variable. int count; hipGetDeviceCount (&count); for (int i=0; i< count; i++){ hipGetDeviceProperties (&prop, i); printf("\n****************************Device %d*************************************\n",i); printf("----------General Information----------\n"); printf("Name: %s\n", prop.name); printf("Compute capability: %d.%d\n", prop.major, prop.minor); // printf("Clock rate: %d\n", prop.clockRate); printf("Clock rate: %d MHz\n", (prop.clockRate/1000)); printf("GPU type: "); if (prop.integrated) printf ("Integrated GPU\n"); else printf ("Discreted GPU\n"); printf("Device copy overlap: "); if (prop.deviceOverlap) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Kernel execution timeout: "); if (prop.kernelExecTimeoutEnabled) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Mapping Host Memory into CUDA Device Address Space: "); if (prop.canMapHostMemory) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Executing Multiple Kernels Support: "); if (prop.concurrentKernels) printf ("Enabled\n"); else printf ("Disabled\n"); printf( "Device Computing Mode: "); if (prop.computeMode == 0) printf("Default\n"); if (prop.computeMode == 1) printf("Exclusive\n"); if (prop.computeMode == 2) printf("Prohibited\n"); printf( "\n----------Memory Information for device----------\n"); // printf( "Total Global Memory: %ld\n", prop.totalGlobalMem); printf( "Total Global Memory: %ld MB\n", (prop.totalGlobalMem/1024/1024)); // printf( "Total Constant Memory: %ld\n", prop.totalConstMem ); printf( "Total Constant Memory: %ld kB\n", (prop.totalConstMem/1024)); // printf( "Total Constant Memory: %ld MB\n", (prop.totalConstMem/1024/1024)); // printf( "Max Memory Pitch Allowed for Memory Copies: %ld\n", prop.memPitch ); printf( "Max Memory Pitch Allowed for Memory Copies: %ld MB\n", (prop.memPitch/1024/1024)); printf( "Texture Alignment: %ld B\n", prop.textureAlignment ); printf( "Max Size Supported for 1D Textures: %ld\n", prop.maxTexture1D); printf( "Max Dimensions Supported for 2D Texture: (%d, %d)\n", prop.maxTexture2D[0], prop.maxTexture2D[1]); printf( "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n", prop.maxTexture3D[0], prop.maxTexture3D[1], prop.maxTexture3D[2]); /* printf( "Max Dimensions Supported for 2D Texture Array: (%d, %d, %d)\n", prop.maxTexture2DArray[0], prop.maxTexture2DArray[1], prop.maxTexture2DArray[2]); */ printf( "\n----------Thread Information for device----------\n"); printf( "Streaming Multiprocessor count: %d\n", prop.multiProcessorCount); // printf( "Shared Memory per Block: %ld\n", prop.sharedMemPerBlock); printf( "Shared Memory per Block: %ld KB\n", (prop.sharedMemPerBlock/1024)); printf( "Registers per Block: %ld\n", prop.regsPerBlock); printf( "Threads in warp: %ld\n", prop.warpSize); printf( "Max threads per block: %d\n", prop.maxThreadsPerBlock); printf( "Max thread dimensions: (%d, %d, %d)\n", prop.maxThreadsDim[0], prop.maxThreadsDim[1], prop.maxThreadsDim[2]); printf( "Max grid dimensions: (%d, %d, %d)\n", prop.maxGridSize[0], prop.maxGridSize[1], prop.maxGridSize[2]); printf("*****************************************************************\n"); } return 0; }
.text .file "page32_gpuinfo.hip" .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 %rbx .cfi_def_cfa_offset 40 subq $1480, %rsp # imm = 0x5C8 .cfi_def_cfa_offset 1520 .cfi_offset %rbx, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 leaq 4(%rsp), %rdi callq hipGetDeviceCount cmpl $0, 4(%rsp) jle .LBB0_9 # %bb.1: # %.lr.ph leaq 8(%rsp), %rbx movl $.Lstr.1, %r14d movl $.Lstr.9, %r15d xorl %ebp, %ebp jmp .LBB0_2 .p2align 4, 0x90 .LBB0_8: # in Loop: Header=BB0_2 Depth=1 movl $.Lstr.14, %edi callq puts@PLT movq 296(%rsp), %rsi shrq $20, %rsi movl $.L.str.19, %edi xorl %eax, %eax callq printf movq 360(%rsp), %rsi shrq $10, %rsi movl $.L.str.20, %edi xorl %eax, %eax callq printf movq 320(%rsp), %rsi shrq $20, %rsi movl $.L.str.21, %edi xorl %eax, %eax callq printf movq 376(%rsp), %rsi movl $.L.str.22, %edi xorl %eax, %eax callq printf movl 416(%rsp), %esi movl $.L.str.23, %edi xorl %eax, %eax callq printf movl 428(%rsp), %esi movl 432(%rsp), %edx movl $.L.str.24, %edi xorl %eax, %eax callq printf movl 464(%rsp), %esi movl 468(%rsp), %edx movl 472(%rsp), %ecx movl $.L.str.25, %edi xorl %eax, %eax callq printf movl $.Lstr.15, %edi callq puts@PLT movl 396(%rsp), %esi movl $.L.str.27, %edi xorl %eax, %eax callq printf movq 304(%rsp), %rsi shrq $10, %rsi movl $.L.str.28, %edi xorl %eax, %eax callq printf movl 312(%rsp), %esi movl $.L.str.29, %edi xorl %eax, %eax callq printf movl 316(%rsp), %esi movl $.L.str.30, %edi xorl %eax, %eax callq printf movl 328(%rsp), %esi movl $.L.str.31, %edi xorl %eax, %eax callq printf movl 332(%rsp), %esi movl 336(%rsp), %edx movl 340(%rsp), %ecx movl $.L.str.32, %edi xorl %eax, %eax callq printf movl 344(%rsp), %esi movl 348(%rsp), %edx movl 352(%rsp), %ecx movl $.L.str.33, %edi xorl %eax, %eax callq printf movl $.Lstr.16, %edi callq puts@PLT incl %ebp cmpl 4(%rsp), %ebp jge .LBB0_9 .LBB0_2: # =>This Inner Loop Header: Depth=1 movq %rbx, %rdi movl %ebp, %esi callq hipGetDevicePropertiesR0600 movl $.L.str, %edi movl %ebp, %esi xorl %eax, %eax callq printf movl $.Lstr, %edi callq puts@PLT movl $.L.str.2, %edi movq %rbx, %rsi xorl %eax, %eax callq printf movl 368(%rsp), %esi movl 372(%rsp), %edx movl $.L.str.3, %edi xorl %eax, %eax callq printf movslq 356(%rsp), %rax imulq $274877907, %rax, %rsi # imm = 0x10624DD3 movq %rsi, %rax shrq $63, %rax sarq $38, %rsi addl %eax, %esi movl $.L.str.4, %edi # kill: def $esi killed $esi killed $rsi xorl %eax, %eax callq printf movl $.L.str.5, %edi xorl %eax, %eax callq printf cmpl $0, 404(%rsp) movl $.Lstr.2, %edi cmoveq %r14, %rdi callq puts@PLT movl $.L.str.8, %edi xorl %eax, %eax callq printf cmpl $0, 392(%rsp) movl $.Lstr.10, %edi cmoveq %r15, %rdi callq puts@PLT movl $.L.str.11, %edi xorl %eax, %eax callq printf cmpl $0, 400(%rsp) movl $.Lstr.10, %edi cmoveq %r15, %rdi callq puts@PLT movl $.L.str.12, %edi xorl %eax, %eax callq printf cmpl $0, 408(%rsp) movl $.Lstr.10, %edi cmoveq %r15, %rdi callq puts@PLT movl $.L.str.13, %edi xorl %eax, %eax callq printf cmpl $0, 584(%rsp) movl $.Lstr.10, %edi cmoveq %r15, %rdi callq puts@PLT movl $.L.str.14, %edi xorl %eax, %eax callq printf cmpl $0, 412(%rsp) je .LBB0_3 # %bb.4: # in Loop: Header=BB0_2 Depth=1 cmpl $1, 412(%rsp) je .LBB0_5 .LBB0_6: # in Loop: Header=BB0_2 Depth=1 cmpl $2, 412(%rsp) jne .LBB0_8 jmp .LBB0_7 .p2align 4, 0x90 .LBB0_3: # in Loop: Header=BB0_2 Depth=1 movl $.Lstr.11, %edi callq puts@PLT cmpl $1, 412(%rsp) jne .LBB0_6 .LBB0_5: # in Loop: Header=BB0_2 Depth=1 movl $.Lstr.12, %edi callq puts@PLT cmpl $2, 412(%rsp) jne .LBB0_8 .LBB0_7: # in Loop: Header=BB0_2 Depth=1 movl $.Lstr.13, %edi callq puts@PLT jmp .LBB0_8 .LBB0_9: # %._crit_edge xorl %eax, %eax addq $1480, %rsp # imm = 0x5C8 .cfi_def_cfa_offset 40 popq %rbx .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 main, .Lfunc_end0-main .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "\n****************************Device %d*************************************\n" .size .L.str, 77 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Name: %s\n" .size .L.str.2, 10 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Compute capability: %d.%d\n" .size .L.str.3, 27 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "Clock rate: %d MHz\n" .size .L.str.4, 20 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "GPU type: " .size .L.str.5, 11 .type .L.str.8,@object # @.str.8 .L.str.8: .asciz "Device copy overlap: " .size .L.str.8, 22 .type .L.str.11,@object # @.str.11 .L.str.11: .asciz "Kernel execution timeout: " .size .L.str.11, 27 .type .L.str.12,@object # @.str.12 .L.str.12: .asciz "Mapping Host Memory into CUDA Device Address Space: " .size .L.str.12, 53 .type .L.str.13,@object # @.str.13 .L.str.13: .asciz "Executing Multiple Kernels Support: " .size .L.str.13, 37 .type .L.str.14,@object # @.str.14 .L.str.14: .asciz "Device Computing Mode: " .size .L.str.14, 24 .type .L.str.19,@object # @.str.19 .L.str.19: .asciz "Total Global Memory: %ld MB\n" .size .L.str.19, 29 .type .L.str.20,@object # @.str.20 .L.str.20: .asciz "Total Constant Memory: %ld kB\n" .size .L.str.20, 31 .type .L.str.21,@object # @.str.21 .L.str.21: .asciz "Max Memory Pitch Allowed for Memory Copies: %ld MB\n" .size .L.str.21, 52 .type .L.str.22,@object # @.str.22 .L.str.22: .asciz "Texture Alignment: %ld B\n" .size .L.str.22, 26 .type .L.str.23,@object # @.str.23 .L.str.23: .asciz "Max Size Supported for 1D Textures: %ld\n" .size .L.str.23, 41 .type .L.str.24,@object # @.str.24 .L.str.24: .asciz "Max Dimensions Supported for 2D Texture: (%d, %d)\n" .size .L.str.24, 51 .type .L.str.25,@object # @.str.25 .L.str.25: .asciz "Max Dimensions Supported for 3D Texture: (%d, %d, %d)\n" .size .L.str.25, 55 .type .L.str.27,@object # @.str.27 .L.str.27: .asciz "Streaming Multiprocessor count: %d\n" .size .L.str.27, 36 .type .L.str.28,@object # @.str.28 .L.str.28: .asciz "Shared Memory per Block: %ld KB\n" .size .L.str.28, 33 .type .L.str.29,@object # @.str.29 .L.str.29: .asciz "Registers per Block: %ld\n" .size .L.str.29, 26 .type .L.str.30,@object # @.str.30 .L.str.30: .asciz "Threads in warp: %ld\n" .size .L.str.30, 22 .type .L.str.31,@object # @.str.31 .L.str.31: .asciz "Max threads per block: %d\n" .size .L.str.31, 27 .type .L.str.32,@object # @.str.32 .L.str.32: .asciz "Max thread dimensions: (%d, %d, %d)\n" .size .L.str.32, 37 .type .L.str.33,@object # @.str.33 .L.str.33: .asciz "Max grid dimensions: (%d, %d, %d)\n" .size .L.str.33, 35 .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 "----------General Information----------" .size .Lstr, 40 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Discreted GPU" .size .Lstr.1, 14 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Integrated GPU" .size .Lstr.2, 15 .type .Lstr.9,@object # @str.9 .Lstr.9: .asciz "Disabled" .size .Lstr.9, 9 .type .Lstr.10,@object # @str.10 .Lstr.10: .asciz "Enabled" .size .Lstr.10, 8 .type .Lstr.11,@object # @str.11 .Lstr.11: .asciz "Default" .size .Lstr.11, 8 .type .Lstr.12,@object # @str.12 .Lstr.12: .asciz "Exclusive" .size .Lstr.12, 10 .type .Lstr.13,@object # @str.13 .Lstr.13: .asciz "Prohibited" .size .Lstr.13, 11 .type .Lstr.14,@object # @str.14 .Lstr.14: .asciz "\n----------Memory Information for device----------" .size .Lstr.14, 51 .type .Lstr.15,@object # @str.15 .Lstr.15: .asciz "\n----------Thread Information for device----------" .size .Lstr.15, 51 .type .Lstr.16,@object # @str.16 .Lstr.16: .asciz "*****************************************************************" .size .Lstr.16, 66 .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