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 "includes.h" /* This file is copied from https://github.com/jzbonter/mc-cnn */ extern "C" { } #define TB 128 #define DISP_MAX 256 __global__ void copy_fill(float *in, float *out, int size, int in_size2, int in_size3, int out_size2, int out_size3) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int out_x = id % out_size3; int out_y = id / out_size3; int in_x = out_x - (out_size3 - in_size3) / 2; int in_y = out_y - (out_size2 - in_size2) / 2; int x = min(in_size3 - 1, max(0, in_x)); int y = min(in_size2 - 1, max(0, in_y)); out[id] = in[y * in_size3 + x]; } }
.file "tmpxft_000a1ade_00000000-6_copy_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 _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii .type _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii, @function _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii: .LFB2051: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%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 152(%rsp), %rax subq %fs:40, %rax jne .L8 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 184 pushq 40(%rsp) .cfi_def_cfa_offset 192 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z9copy_fillPfS_iiiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii, .-_Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii .globl _Z9copy_fillPfS_iiiii .type _Z9copy_fillPfS_iiiii, @function _Z9copy_fillPfS_iiiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z9copy_fillPfS_iiiii, .-_Z9copy_fillPfS_iiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z9copy_fillPfS_iiiii" .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 _Z9copy_fillPfS_iiiii(%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" /* This file is copied from https://github.com/jzbonter/mc-cnn */ extern "C" { } #define TB 128 #define DISP_MAX 256 __global__ void copy_fill(float *in, float *out, int size, int in_size2, int in_size3, int out_size2, int out_size3) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int out_x = id % out_size3; int out_y = id / out_size3; int in_x = out_x - (out_size3 - in_size3) / 2; int in_y = out_y - (out_size2 - in_size2) / 2; int x = min(in_size3 - 1, max(0, in_x)); int y = min(in_size2 - 1, max(0, in_y)); out[id] = in[y * in_size3 + x]; } }
#include <hip/hip_runtime.h> #include "includes.h" /* This file is copied from https://github.com/jzbonter/mc-cnn */ extern "C" { } #define TB 128 #define DISP_MAX 256 __global__ void copy_fill(float *in, float *out, int size, int in_size2, int in_size3, int out_size2, int out_size3) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int out_x = id % out_size3; int out_y = id / out_size3; int in_x = out_x - (out_size3 - in_size3) / 2; int in_y = out_y - (out_size2 - in_size2) / 2; int x = min(in_size3 - 1, max(0, in_x)); int y = min(in_size2 - 1, max(0, in_y)); out[id] = in[y * in_size3 + x]; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" /* This file is copied from https://github.com/jzbonter/mc-cnn */ extern "C" { } #define TB 128 #define DISP_MAX 256 __global__ void copy_fill(float *in, float *out, int size, int in_size2, int in_size3, int out_size2, int out_size3) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int out_x = id % out_size3; int out_y = id / out_size3; int in_x = out_x - (out_size3 - in_size3) / 2; int in_y = out_y - (out_size2 - in_size2) / 2; int x = min(in_size3 - 1, max(0, in_x)); int y = min(in_size2 - 1, max(0, in_y)); out[id] = in[y * in_size3 + x]; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z9copy_fillPfS_iiiii .globl _Z9copy_fillPfS_iiiii .p2align 8 .type _Z9copy_fillPfS_iiiii,@function _Z9copy_fillPfS_iiiii: s_clause 0x1 s_load_b32 s2, s[0:1], 0x34 s_load_b32 s3, s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_i32_e64 s3, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x14 v_ashrrev_i32_e32 v3, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v4, v1, v3 v_xor_b32_e32 v4, v4, v3 s_waitcnt lgkmcnt(0) s_ashr_i32 s2, s7, 31 s_delay_alu instid0(SALU_CYCLE_1) s_add_i32 s3, s7, s2 v_xor_b32_e32 v3, s2, v3 s_xor_b32 s3, s3, s2 s_sub_i32 s2, s6, s4 v_cvt_f32_u32_e32 v0, s3 s_sub_i32 s8, 0, s3 s_lshr_b32 s6, s2, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) s_add_i32 s2, s2, s6 v_rcp_iflag_f32_e32 v0, v0 s_ashr_i32 s2, s2, 1 s_waitcnt_depctr 0xfff v_mul_f32_e32 v0, 0x4f7ffffe, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_u32_f32_e32 v0, v0 v_mul_lo_u32 v2, s8, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v2, v0, v2 v_add_nc_u32_e32 v0, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v0, v4, v0 v_mul_lo_u32 v2, v0, s3 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v2, v4, v2 v_add_nc_u32_e32 v4, 1, v0 v_subrev_nc_u32_e32 v5, s3, v2 v_cmp_le_u32_e32 vcc_lo, s3, v2 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v0, v0, v4, vcc_lo v_cndmask_b32_e32 v2, v2, v5, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, 1, v0 v_cmp_le_u32_e32 vcc_lo, s3, v2 s_sub_i32 s3, s7, s5 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) 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, s7 s_lshr_b32 s7, s3, 31 v_subrev_nc_u32_e32 v0, s2, v0 s_add_i32 s3, s3, s7 s_add_i32 s2, s4, -1 s_ashr_i32 s3, s3, 1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_maxmin_i32 v4, v0, 0, s2 v_sub_nc_u32_e32 v2, v1, v2 s_delay_alu instid0(VALU_DEP_1) v_subrev_nc_u32_e32 v2, s3, v2 s_add_i32 s3, s5, -1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_maxmin_i32 v0, v2, 0, s3 s_load_b128 s[0:3], s[0:1], 0x0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, v4, s5, v[0:1] v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[2:3], 2, v[2:3] s_waitcnt lgkmcnt(0) 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, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, 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 _Z9copy_fillPfS_iiiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z9copy_fillPfS_iiiii, .Lfunc_end0-_Z9copy_fillPfS_iiiii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: by_value - .offset: 20 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z9copy_fillPfS_iiiii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z9copy_fillPfS_iiiii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" /* This file is copied from https://github.com/jzbonter/mc-cnn */ extern "C" { } #define TB 128 #define DISP_MAX 256 __global__ void copy_fill(float *in, float *out, int size, int in_size2, int in_size3, int out_size2, int out_size3) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int out_x = id % out_size3; int out_y = id / out_size3; int in_x = out_x - (out_size3 - in_size3) / 2; int in_y = out_y - (out_size2 - in_size2) / 2; int x = min(in_size3 - 1, max(0, in_x)); int y = min(in_size2 - 1, max(0, in_y)); out[id] = in[y * in_size3 + x]; } }
.text .file "copy_fill.hip" .globl _Z24__device_stub__copy_fillPfS_iiiii # -- Begin function _Z24__device_stub__copy_fillPfS_iiiii .p2align 4, 0x90 .type _Z24__device_stub__copy_fillPfS_iiiii,@function _Z24__device_stub__copy_fillPfS_iiiii: # @_Z24__device_stub__copy_fillPfS_iiiii .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 144(%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 80(%rsp), %r9 movl $_Z9copy_fillPfS_iiiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end0: .size _Z24__device_stub__copy_fillPfS_iiiii, .Lfunc_end0-_Z24__device_stub__copy_fillPfS_iiiii .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 $_Z9copy_fillPfS_iiiii, %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 _Z9copy_fillPfS_iiiii,@object # @_Z9copy_fillPfS_iiiii .section .rodata,"a",@progbits .globl _Z9copy_fillPfS_iiiii .p2align 3, 0x0 _Z9copy_fillPfS_iiiii: .quad _Z24__device_stub__copy_fillPfS_iiiii .size _Z9copy_fillPfS_iiiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z9copy_fillPfS_iiiii" .size .L__unnamed_1, 22 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z24__device_stub__copy_fillPfS_iiiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z9copy_fillPfS_iiiii .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 : _Z9copy_fillPfS_iiiii .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ IABS R5, c[0x0][0x180] ; /* 0x0000600000057a13 */ /* 0x000fe20000000000 */ /*0070*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe40000000800 */ /*0080*/ ULDC UR6, c[0x0][0x174] ; /* 0x00005d0000067ab9 */ /* 0x000fe20000000800 */ /*0090*/ I2F.RP R4, R5 ; /* 0x0000000500047306 */ /* 0x000e220000209400 */ /*00a0*/ ULDC.64 UR8, c[0x0][0x178] ; /* 0x00005e0000087ab9 */ /* 0x000fe40000000a00 */ /*00b0*/ UIADD3 UR4, UR4, -UR8, URZ ; /* 0x8000000804047290 */ /* 0x000fe4000fffe03f */ /*00c0*/ UIADD3 UR5, UR9, -UR6, URZ ; /* 0x8000000609057290 */ /* 0x000fe4000fffe03f */ /*00d0*/ ULEA.HI UR4, UR4, UR4, URZ, 0x1 ; /* 0x0000000404047291 */ /* 0x000fc4000f8f083f */ /*00e0*/ ULEA.HI UR5, UR5, UR5, URZ, 0x1 ; /* 0x0000000505057291 */ /* 0x000fe4000f8f083f */ /*00f0*/ USHF.R.S32.HI UR4, URZ, 0x1, UR4 ; /* 0x000000013f047899 */ /* 0x000fe40008011404 */ /*0100*/ USHF.R.S32.HI UR5, URZ, 0x1, UR5 ; /* 0x000000013f057899 */ /* 0x000fe40008011405 */ /*0110*/ UMOV UR7, 0x1 ; /* 0x0000000100077882 */ /* 0x000fe20000000000 */ /*0120*/ MUFU.RCP R4, R4 ; /* 0x0000000400047308 */ /* 0x001e220000001000 */ /*0130*/ UIADD3 UR6, -UR7, UR6, URZ ; /* 0x0000000607067290 */ /* 0x000fe2000fffe13f */ /*0140*/ IADD3 R2, R4, 0xffffffe, RZ ; /* 0x0ffffffe04027810 */ /* 0x001fcc0007ffe0ff */ /*0150*/ F2I.FTZ.U32.TRUNC.NTZ R3, R2 ; /* 0x0000000200037305 */ /* 0x000064000021f000 */ /*0160*/ IMAD.MOV.U32 R2, RZ, RZ, RZ ; /* 0x000000ffff027224 */ /* 0x001fe400078e00ff */ /*0170*/ IMAD.MOV R6, RZ, RZ, -R3 ; /* 0x000000ffff067224 */ /* 0x002fc800078e0a03 */ /*0180*/ IMAD R7, R6, R5, RZ ; /* 0x0000000506077224 */ /* 0x000fe200078e02ff */ /*0190*/ IABS R6, R0 ; /* 0x0000000000067213 */ /* 0x000fc60000000000 */ /*01a0*/ IMAD.HI.U32 R3, R3, R7, R2 ; /* 0x0000000703037227 */ /* 0x000fe200078e0002 */ /*01b0*/ LOP3.LUT R2, R0, c[0x0][0x180], RZ, 0x3c, !PT ; /* 0x0000600000027a12 */ /* 0x000fc800078e3cff */ /*01c0*/ ISETP.GE.AND P1, PT, R2, RZ, PT ; /* 0x000000ff0200720c */ /* 0x000fe20003f26270 */ /*01d0*/ IMAD.HI.U32 R3, R3, R6, RZ ; /* 0x0000000603037227 */ /* 0x000fc800078e00ff */ /*01e0*/ IMAD.MOV R4, RZ, RZ, -R3 ; /* 0x000000ffff047224 */ /* 0x000fc800078e0a03 */ /*01f0*/ IMAD R4, R5, R4, R6 ; /* 0x0000000405047224 */ /* 0x000fca00078e0206 */ /*0200*/ ISETP.GT.U32.AND P2, PT, R5, R4, PT ; /* 0x000000040500720c */ /* 0x000fda0003f44070 */ /*0210*/ @!P2 IMAD.IADD R4, R4, 0x1, -R5 ; /* 0x000000010404a824 */ /* 0x000fe200078e0a05 */ /*0220*/ @!P2 IADD3 R3, R3, 0x1, RZ ; /* 0x000000010303a810 */ /* 0x000fe40007ffe0ff */ /*0230*/ ISETP.NE.AND P2, PT, RZ, c[0x0][0x180], PT ; /* 0x00006000ff007a0c */ /* 0x000fe40003f45270 */ /*0240*/ ISETP.GE.U32.AND P0, PT, R4, R5, PT ; /* 0x000000050400720c */ /* 0x000fda0003f06070 */ /*0250*/ @P0 IADD3 R3, R3, 0x1, RZ ; /* 0x0000000103030810 */ /* 0x000fca0007ffe0ff */ /*0260*/ @!P1 IMAD.MOV R3, RZ, RZ, -R3 ; /* 0x000000ffff039224 */ /* 0x000fe200078e0a03 */ /*0270*/ @!P2 LOP3.LUT R3, RZ, c[0x0][0x180], RZ, 0x33, !PT ; /* 0x00006000ff03aa12 */ /* 0x000fca00078e33ff */ /*0280*/ IMAD.MOV R5, RZ, RZ, -R3 ; /* 0x000000ffff057224 */ /* 0x000fe200078e0a03 */ /*0290*/ IADD3 R3, R3, -UR5, RZ ; /* 0x8000000503037c10 */ /* 0x000fc6000fffe0ff */ /*02a0*/ IMAD R2, R5, c[0x0][0x180], R0 ; /* 0x0000600005027a24 */ /* 0x000fe200078e0200 */ /*02b0*/ IMNMX R3, RZ, R3, !PT ; /* 0x00000003ff037217 */ /* 0x000fe20007800200 */ /*02c0*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fc600078e00ff */ /*02d0*/ IADD3 R2, R2, -UR4, RZ ; /* 0x8000000402027c10 */ /* 0x000fe2000fffe0ff */ /*02e0*/ UIADD3 UR4, -UR7, UR8, URZ ; /* 0x0000000807047290 */ /* 0x000fe2000fffe13f */ /*02f0*/ IMNMX R3, R3, UR6, PT ; /* 0x0000000603037c17 */ /* 0x000fe4000b800200 */ /*0300*/ IMNMX R2, RZ, R2, !PT ; /* 0x00000002ff027217 */ /* 0x000fc80007800200 */ /*0310*/ IMNMX R2, R2, UR4, PT ; /* 0x0000000402027c17 */ /* 0x000fe2000b800200 */ /*0320*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc80000000a00 */ /*0330*/ IMAD R2, R3, c[0x0][0x178], R2 ; /* 0x00005e0003027a24 */ /* 0x000fc800078e0202 */ /*0340*/ IMAD.WIDE R2, R2, R5, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fcc00078e0205 */ /*0350*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0360*/ IMAD.WIDE R4, R0, R5, c[0x0][0x168] ; /* 0x00005a0000047625 */ /* 0x000fca00078e0205 */ /*0370*/ STG.E [R4.64], R3 ; /* 0x0000000304007986 */ /* 0x004fe2000c101904 */ /*0380*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0390*/ BRA 0x390; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*03a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*03f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0400*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0410*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0420*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0430*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0440*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0450*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0460*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0470*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z9copy_fillPfS_iiiii .globl _Z9copy_fillPfS_iiiii .p2align 8 .type _Z9copy_fillPfS_iiiii,@function _Z9copy_fillPfS_iiiii: s_clause 0x1 s_load_b32 s2, s[0:1], 0x34 s_load_b32 s3, s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_i32_e64 s3, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x14 v_ashrrev_i32_e32 v3, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v4, v1, v3 v_xor_b32_e32 v4, v4, v3 s_waitcnt lgkmcnt(0) s_ashr_i32 s2, s7, 31 s_delay_alu instid0(SALU_CYCLE_1) s_add_i32 s3, s7, s2 v_xor_b32_e32 v3, s2, v3 s_xor_b32 s3, s3, s2 s_sub_i32 s2, s6, s4 v_cvt_f32_u32_e32 v0, s3 s_sub_i32 s8, 0, s3 s_lshr_b32 s6, s2, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) s_add_i32 s2, s2, s6 v_rcp_iflag_f32_e32 v0, v0 s_ashr_i32 s2, s2, 1 s_waitcnt_depctr 0xfff v_mul_f32_e32 v0, 0x4f7ffffe, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_u32_f32_e32 v0, v0 v_mul_lo_u32 v2, s8, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v2, v0, v2 v_add_nc_u32_e32 v0, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v0, v4, v0 v_mul_lo_u32 v2, v0, s3 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v2, v4, v2 v_add_nc_u32_e32 v4, 1, v0 v_subrev_nc_u32_e32 v5, s3, v2 v_cmp_le_u32_e32 vcc_lo, s3, v2 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v0, v0, v4, vcc_lo v_cndmask_b32_e32 v2, v2, v5, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, 1, v0 v_cmp_le_u32_e32 vcc_lo, s3, v2 s_sub_i32 s3, s7, s5 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) 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, s7 s_lshr_b32 s7, s3, 31 v_subrev_nc_u32_e32 v0, s2, v0 s_add_i32 s3, s3, s7 s_add_i32 s2, s4, -1 s_ashr_i32 s3, s3, 1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_maxmin_i32 v4, v0, 0, s2 v_sub_nc_u32_e32 v2, v1, v2 s_delay_alu instid0(VALU_DEP_1) v_subrev_nc_u32_e32 v2, s3, v2 s_add_i32 s3, s5, -1 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_maxmin_i32 v0, v2, 0, s3 s_load_b128 s[0:3], s[0:1], 0x0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[2:3], null, v4, s5, v[0:1] v_ashrrev_i32_e32 v3, 31, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[2:3], 2, v[2:3] s_waitcnt lgkmcnt(0) 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, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, 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 _Z9copy_fillPfS_iiiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z9copy_fillPfS_iiiii, .Lfunc_end0-_Z9copy_fillPfS_iiiii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: by_value - .offset: 20 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z9copy_fillPfS_iiiii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z9copy_fillPfS_iiiii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000a1ade_00000000-6_copy_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 _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii .type _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii, @function _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii: .LFB2051: .cfi_startproc endbr64 subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) movq %fs:40, %rax movq %rax, 152(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 176(%rsp), %rax movq %rax, 144(%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 152(%rsp), %rax subq %fs:40, %rax jne .L8 addq $168, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 184 pushq 40(%rsp) .cfi_def_cfa_offset 192 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z9copy_fillPfS_iiiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 176 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii, .-_Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii .globl _Z9copy_fillPfS_iiiii .type _Z9copy_fillPfS_iiiii, @function _Z9copy_fillPfS_iiiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z35__device_stub__Z9copy_fillPfS_iiiiiPfS_iiiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z9copy_fillPfS_iiiii, .-_Z9copy_fillPfS_iiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z9copy_fillPfS_iiiii" .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 _Z9copy_fillPfS_iiiii(%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 "copy_fill.hip" .globl _Z24__device_stub__copy_fillPfS_iiiii # -- Begin function _Z24__device_stub__copy_fillPfS_iiiii .p2align 4, 0x90 .type _Z24__device_stub__copy_fillPfS_iiiii,@function _Z24__device_stub__copy_fillPfS_iiiii: # @_Z24__device_stub__copy_fillPfS_iiiii .cfi_startproc # %bb.0: subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 144(%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 80(%rsp), %r9 movl $_Z9copy_fillPfS_iiiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $152, %rsp .cfi_adjust_cfa_offset -152 retq .Lfunc_end0: .size _Z24__device_stub__copy_fillPfS_iiiii, .Lfunc_end0-_Z24__device_stub__copy_fillPfS_iiiii .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 $_Z9copy_fillPfS_iiiii, %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 _Z9copy_fillPfS_iiiii,@object # @_Z9copy_fillPfS_iiiii .section .rodata,"a",@progbits .globl _Z9copy_fillPfS_iiiii .p2align 3, 0x0 _Z9copy_fillPfS_iiiii: .quad _Z24__device_stub__copy_fillPfS_iiiii .size _Z9copy_fillPfS_iiiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z9copy_fillPfS_iiiii" .size .L__unnamed_1, 22 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z24__device_stub__copy_fillPfS_iiiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z9copy_fillPfS_iiiii .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.
int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); return 0; }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); return 0; }
.file "tmpxft_001296a1_00000000-6_t.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl main .type main, @function main: .LFB2027: .cfi_startproc endbr64 movl $0, %eax ret .cfi_endproc .LFE2027: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); return 0; }
#include <hip/hip_runtime.h> int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); 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> int main() { int a = 10; #pragma omp parallel for for (int i = 0; i < 10; ++i); return 0; }
.text .file "t.hip" .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: xorl %eax, %eax retq .Lfunc_end0: .size main, .Lfunc_end0-main .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_001296a1_00000000-6_t.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl main .type main, @function main: .LFB2027: .cfi_startproc endbr64 movl $0, %eax ret .cfi_endproc .LFE2027: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "t.hip" .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: xorl %eax, %eax retq .Lfunc_end0: .size main, .Lfunc_end0-main .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 "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <stdio.h> __device__ float devData; __global__ void checkGlobalVariable() { printf("Device: The value of global variable is %f\n", devData); devData += 2.0; } int main() { float value = 3.14f; cudaMemcpyToSymbol(&devData, &value, sizeof(value)); printf("Host: copied %f to the global variable\n", value); checkGlobalVariable<<<1, 1 >>> (); cudaMemcpyFromSymbol(&value, &devData, sizeof(value)); printf("Host: Value changed by kernel to %f\n", value); cudaDeviceReset(); system("Pause"); return 0; }
code for sm_80 Function : _Z19checkGlobalVariablev .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ IMAD.MOV.U32 R16, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff107624 */ /* 0x000fe200078e00ff */ /*0020*/ ULDC.64 UR36, c[0x0][0x118] ; /* 0x0000460000247ab9 */ /* 0x000fe20000000a00 */ /*0030*/ IMAD.MOV.U32 R17, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff117624 */ /* 0x000fe200078e00ff */ /*0040*/ IADD3 R1, R1, -0x8, RZ ; /* 0xfffffff801017810 */ /* 0x000fc80007ffe0ff */ /*0050*/ LDG.E R0, [R16.64] ; /* 0x0000002410007981 */ /* 0x000ea2000c1e1900 */ /*0060*/ MOV R8, 0x0 ; /* 0x0000000000087802 */ /* 0x000fe20000000f00 */ /*0070*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x10] ; /* 0x01000400ff047624 */ /* 0x000fe200078e00ff */ /*0080*/ IADD3 R6, P0, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fe20007f1e0ff */ /*0090*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0x14] ; /* 0x01000500ff057624 */ /* 0x000fc600078e00ff */ /*00a0*/ LDC.64 R8, c[0x4][R8] ; /* 0x0100000008087b82 */ /* 0x000e220000000a00 */ /*00b0*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P0 ; /* 0x00000900ff077624 */ /* 0x000fe200000e06ff */ /*00c0*/ F2F.F64.F32 R2, R0 ; /* 0x0000000000027310 */ /* 0x004e640000201800 */ /*00d0*/ STL.64 [R1], R2 ; /* 0x0000000201007387 */ /* 0x0023ea0000100a00 */ /*00e0*/ LEPC R2 ; /* 0x000000000002734e */ /* 0x003fe40000000000 */ /*00f0*/ MOV R11, 0x160 ; /* 0x00000160000b7802 */ /* 0x000fe40000000f00 */ /*0100*/ MOV R20, 0xe0 ; /* 0x000000e000147802 */ /* 0x000fe40000000f00 */ /*0110*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fe40000000f00 */ /*0120*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fc40000000f00 */ /*0130*/ IADD3 R20, P0, P1, -R20, R11, R2 ; /* 0x0000000b14147210 */ /* 0x000fc8000791e102 */ /*0140*/ IADD3.X R21, ~R0, R21, R3, P0, P1 ; /* 0x0000001500157210 */ /* 0x000fc800007e2503 */ /*0150*/ CALL.ABS.NOINC R8 ; /* 0x0000000008007343 */ /* 0x000fea0003c00000 */ /*0160*/ LDG.E R0, [R16.64] ; /* 0x0000002410007981 */ /* 0x000ea4000c1e1900 */ /*0170*/ FADD R3, R0, 2 ; /* 0x4000000000037421 */ /* 0x004fca0000000000 */ /*0180*/ STG.E [R16.64], R3 ; /* 0x0000000310007986 */ /* 0x000fe2000c101924 */ /*0190*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*01a0*/ BRA 0x1a0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <stdio.h> __device__ float devData; __global__ void checkGlobalVariable() { printf("Device: The value of global variable is %f\n", devData); devData += 2.0; } int main() { float value = 3.14f; cudaMemcpyToSymbol(&devData, &value, sizeof(value)); printf("Host: copied %f to the global variable\n", value); checkGlobalVariable<<<1, 1 >>> (); cudaMemcpyFromSymbol(&value, &devData, sizeof(value)); printf("Host: Value changed by kernel to %f\n", value); cudaDeviceReset(); system("Pause"); return 0; }
.file "tmpxft_000030c1_00000000-6_SaticGlobalMemory.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z38__device_stub__Z19checkGlobalVariablevv .type _Z38__device_stub__Z19checkGlobalVariablevv, @function _Z38__device_stub__Z19checkGlobalVariablevv: .LFB3694: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z19checkGlobalVariablev(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z38__device_stub__Z19checkGlobalVariablevv, .-_Z38__device_stub__Z19checkGlobalVariablevv .globl _Z19checkGlobalVariablev .type _Z19checkGlobalVariablev, @function _Z19checkGlobalVariablev: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z19checkGlobalVariablevv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z19checkGlobalVariablev, .-_Z19checkGlobalVariablev .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC1: .string "Host: copied %f to the global variable\n" .align 8 .LC2: .string "Host: Value changed by kernel to %f\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC3: .string "Pause" .text .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 subq $56, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0x4048f5c3, 8(%rsp) leaq _ZL7devData(%rip), %rax movq %rax, 24(%rsp) leaq 8(%rsp), %rsi leaq 24(%rsp), %rdi movl $1, %r8d movl $0, %ecx movl $4, %edx call cudaMemcpyToSymbol@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 leaq .LC1(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 12(%rsp) movl $1, 16(%rsp) movl $0, %r9d movl $0, %r8d movq 24(%rsp), %rdx movl $1, %ecx movq 12(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L15 .L12: leaq _ZL7devData(%rip), %rax movq %rax, 24(%rsp) leaq 24(%rsp), %rsi leaq 8(%rsp), %rdi movl $2, %r8d movl $0, %ecx movl $4, %edx call cudaMemcpyFromSymbol@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 leaq .LC2(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT call cudaDeviceReset@PLT leaq .LC3(%rip), %rdi call system@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L16 movl $0, %eax addq $56, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state call _Z38__device_stub__Z19checkGlobalVariablevv jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z19checkGlobalVariablev" .LC5: .string "devData" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z19checkGlobalVariablev(%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 movl $4, %r9d movl $0, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _ZL7devData(%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 .LFE3697: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL7devData .comm _ZL7devData,4,4 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include <iostream> #include <stdio.h> __device__ float devData; __global__ void checkGlobalVariable() { printf("Device: The value of global variable is %f\n", devData); devData += 2.0; } int main() { float value = 3.14f; cudaMemcpyToSymbol(&devData, &value, sizeof(value)); printf("Host: copied %f to the global variable\n", value); checkGlobalVariable<<<1, 1 >>> (); cudaMemcpyFromSymbol(&value, &devData, sizeof(value)); printf("Host: Value changed by kernel to %f\n", value); cudaDeviceReset(); system("Pause"); return 0; }
#include "hip/hip_runtime.h" #include <iostream> #include <stdio.h> __device__ float devData; __global__ void checkGlobalVariable() { printf("Device: The value of global variable is %f\n", devData); devData += 2.0; } int main() { float value = 3.14f; hipMemcpyToSymbol(HIP_SYMBOL(&devData), &value, sizeof(value)); printf("Host: copied %f to the global variable\n", value); checkGlobalVariable<<<1, 1 >>> (); hipMemcpyFromSymbol(&value, HIP_SYMBOL(&devData), sizeof(value)); printf("Host: Value changed by kernel to %f\n", value); hipDeviceReset(); system("Pause"); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include "hip/hip_runtime.h" #include <iostream> #include <stdio.h> __device__ float devData; __global__ void checkGlobalVariable() { printf("Device: The value of global variable is %f\n", devData); devData += 2.0; } int main() { float value = 3.14f; hipMemcpyToSymbol(HIP_SYMBOL(&devData), &value, sizeof(value)); printf("Host: copied %f to the global variable\n", value); checkGlobalVariable<<<1, 1 >>> (); hipMemcpyFromSymbol(&value, HIP_SYMBOL(&devData), sizeof(value)); printf("Host: Value changed by kernel to %f\n", value); hipDeviceReset(); system("Pause"); return 0; }
.text .file "SaticGlobalMemory.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z34__device_stub__checkGlobalVariablev # -- Begin function _Z34__device_stub__checkGlobalVariablev .p2align 4, 0x90 .type _Z34__device_stub__checkGlobalVariablev,@function _Z34__device_stub__checkGlobalVariablev: # @_Z34__device_stub__checkGlobalVariablev .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 $_Z19checkGlobalVariablev, %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 _Z34__device_stub__checkGlobalVariablev, .Lfunc_end0-_Z34__device_stub__checkGlobalVariablev .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movl $1078523331, 12(%rsp) # imm = 0x4048F5C3 movq $devData, 16(%rsp) leaq 16(%rsp), %rdi leaq 12(%rsp), %rsi movl $4, %edx xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol movss 12(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str, %edi movb $1, %al callq printf movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 16(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z19checkGlobalVariablev, %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_2: movq $devData, 16(%rsp) leaq 12(%rsp), %rdi leaq 16(%rsp), %rsi movl $4, %edx xorl %ecx, %ecx movl $2, %r8d callq hipMemcpyFromSymbol movss 12(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str.1, %edi movb $1, %al callq printf callq hipDeviceReset movl $.L.str.2, %edi callq system xorl %eax, %eax addq $72, %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: 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 .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19checkGlobalVariablev, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $0, 8(%rsp) movl $0, (%rsp) movl $devData, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $4, %r9d movq %rbx, %rdi xorl %r8d, %r8d callq __hipRegisterVar movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .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 devData,@object # @devData .local devData .comm devData,4,4 .type _Z19checkGlobalVariablev,@object # @_Z19checkGlobalVariablev .section .rodata,"a",@progbits .globl _Z19checkGlobalVariablev .p2align 3, 0x0 _Z19checkGlobalVariablev: .quad _Z34__device_stub__checkGlobalVariablev .size _Z19checkGlobalVariablev, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Host: copied %f to the global variable\n" .size .L.str, 40 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Host: Value changed by kernel to %f\n" .size .L.str.1, 37 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Pause" .size .L.str.2, 6 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19checkGlobalVariablev" .size .L__unnamed_1, 25 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "devData" .size .L__unnamed_2, 8 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z34__device_stub__checkGlobalVariablev .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym devData .addrsig_sym _Z19checkGlobalVariablev .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_000030c1_00000000-6_SaticGlobalMemory.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z38__device_stub__Z19checkGlobalVariablevv .type _Z38__device_stub__Z19checkGlobalVariablevv, @function _Z38__device_stub__Z19checkGlobalVariablevv: .LFB3694: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z19checkGlobalVariablev(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z38__device_stub__Z19checkGlobalVariablevv, .-_Z38__device_stub__Z19checkGlobalVariablevv .globl _Z19checkGlobalVariablev .type _Z19checkGlobalVariablev, @function _Z19checkGlobalVariablev: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z19checkGlobalVariablevv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z19checkGlobalVariablev, .-_Z19checkGlobalVariablev .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC1: .string "Host: copied %f to the global variable\n" .align 8 .LC2: .string "Host: Value changed by kernel to %f\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC3: .string "Pause" .text .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 subq $56, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0x4048f5c3, 8(%rsp) leaq _ZL7devData(%rip), %rax movq %rax, 24(%rsp) leaq 8(%rsp), %rsi leaq 24(%rsp), %rdi movl $1, %r8d movl $0, %ecx movl $4, %edx call cudaMemcpyToSymbol@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 leaq .LC1(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 12(%rsp) movl $1, 16(%rsp) movl $0, %r9d movl $0, %r8d movq 24(%rsp), %rdx movl $1, %ecx movq 12(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L15 .L12: leaq _ZL7devData(%rip), %rax movq %rax, 24(%rsp) leaq 24(%rsp), %rsi leaq 8(%rsp), %rdi movl $2, %r8d movl $0, %ecx movl $4, %edx call cudaMemcpyFromSymbol@PLT pxor %xmm0, %xmm0 cvtss2sd 8(%rsp), %xmm0 leaq .LC2(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT call cudaDeviceReset@PLT leaq .LC3(%rip), %rdi call system@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L16 movl $0, %eax addq $56, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state call _Z38__device_stub__Z19checkGlobalVariablevv jmp .L12 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z19checkGlobalVariablev" .LC5: .string "devData" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z19checkGlobalVariablev(%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 movl $4, %r9d movl $0, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _ZL7devData(%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 .LFE3697: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL7devData .comm _ZL7devData,4,4 .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 "SaticGlobalMemory.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z34__device_stub__checkGlobalVariablev # -- Begin function _Z34__device_stub__checkGlobalVariablev .p2align 4, 0x90 .type _Z34__device_stub__checkGlobalVariablev,@function _Z34__device_stub__checkGlobalVariablev: # @_Z34__device_stub__checkGlobalVariablev .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 $_Z19checkGlobalVariablev, %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 _Z34__device_stub__checkGlobalVariablev, .Lfunc_end0-_Z34__device_stub__checkGlobalVariablev .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movl $1078523331, 12(%rsp) # imm = 0x4048F5C3 movq $devData, 16(%rsp) leaq 16(%rsp), %rdi leaq 12(%rsp), %rsi movl $4, %edx xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol movss 12(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str, %edi movb $1, %al callq printf movabsq $4294967297, %rdi # imm = 0x100000001 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 16(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 16(%rsp), %rsi movl 24(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z19checkGlobalVariablev, %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_2: movq $devData, 16(%rsp) leaq 12(%rsp), %rdi leaq 16(%rsp), %rsi movl $4, %edx xorl %ecx, %ecx movl $2, %r8d callq hipMemcpyFromSymbol movss 12(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero cvtss2sd %xmm0, %xmm0 movl $.L.str.1, %edi movb $1, %al callq printf callq hipDeviceReset movl $.L.str.2, %edi callq system xorl %eax, %eax addq $72, %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: 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 .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z19checkGlobalVariablev, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $0, 8(%rsp) movl $0, (%rsp) movl $devData, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $4, %r9d movq %rbx, %rdi xorl %r8d, %r8d callq __hipRegisterVar movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .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 devData,@object # @devData .local devData .comm devData,4,4 .type _Z19checkGlobalVariablev,@object # @_Z19checkGlobalVariablev .section .rodata,"a",@progbits .globl _Z19checkGlobalVariablev .p2align 3, 0x0 _Z19checkGlobalVariablev: .quad _Z34__device_stub__checkGlobalVariablev .size _Z19checkGlobalVariablev, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Host: copied %f to the global variable\n" .size .L.str, 40 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Host: Value changed by kernel to %f\n" .size .L.str.1, 37 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Pause" .size .L.str.2, 6 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z19checkGlobalVariablev" .size .L__unnamed_1, 25 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "devData" .size .L__unnamed_2, 8 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z34__device_stub__checkGlobalVariablev .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym devData .addrsig_sym _Z19checkGlobalVariablev .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" __global__ void cuda_graph_avgpool_bprop(float* gradInput, const float *gradOutput, const float* clusters, const int nClusters, const int poolsize, const int dim, const int nClustersPerThread) { extern __shared__ float shared_mem[]; float* gradOutput_data = (float*)shared_mem; const int tidx = threadIdx.x; gradInput += blockIdx.x * dim; gradOutput += blockIdx.x * nClusters; __syncthreads(); for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { gradOutput_data[idx] = gradOutput[idx]; } } __syncthreads(); if (tidx < poolsize) { for (int i = 0; i < nClusters; ++i) { gradInput[(int)(clusters[i*poolsize+tidx]-1)] += gradOutput[i]/poolsize; } } /* for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[tidx*poolsize+j]-1)] += gradOutput[tidx]/poolsize; __syncthreads(); } */ __syncthreads(); /* //ouch... if (tidx == 1) { for (int i = 0; i < nClusters; ++i) { // int idx = tidx + i*blockDim.x; for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[i*poolsize+j]-1)] += gradOutput[i]/poolsize; } } } */ /* for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { for (int j = 0; j < poolsize; ++j) { gradInput[(int)clusters[idx*poolsize+j]] += gradOutput_data[idx]/poolsize; } } } */ }
.file "tmpxft_00172c58_00000000-6_cuda_graph_avgpool_bprop.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii .type _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii, @function _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii: .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) 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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii(%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 _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii, .-_Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii .globl _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .type _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, @function _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, .-_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii" .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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii(%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 cuda_graph_avgpool_bprop(float* gradInput, const float *gradOutput, const float* clusters, const int nClusters, const int poolsize, const int dim, const int nClustersPerThread) { extern __shared__ float shared_mem[]; float* gradOutput_data = (float*)shared_mem; const int tidx = threadIdx.x; gradInput += blockIdx.x * dim; gradOutput += blockIdx.x * nClusters; __syncthreads(); for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { gradOutput_data[idx] = gradOutput[idx]; } } __syncthreads(); if (tidx < poolsize) { for (int i = 0; i < nClusters; ++i) { gradInput[(int)(clusters[i*poolsize+tidx]-1)] += gradOutput[i]/poolsize; } } /* for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[tidx*poolsize+j]-1)] += gradOutput[tidx]/poolsize; __syncthreads(); } */ __syncthreads(); /* //ouch... if (tidx == 1) { for (int i = 0; i < nClusters; ++i) { // int idx = tidx + i*blockDim.x; for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[i*poolsize+j]-1)] += gradOutput[i]/poolsize; } } } */ /* for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { for (int j = 0; j < poolsize; ++j) { gradInput[(int)clusters[idx*poolsize+j]] += gradOutput_data[idx]/poolsize; } } } */ }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void cuda_graph_avgpool_bprop(float* gradInput, const float *gradOutput, const float* clusters, const int nClusters, const int poolsize, const int dim, const int nClustersPerThread) { extern __shared__ float shared_mem[]; float* gradOutput_data = (float*)shared_mem; const int tidx = threadIdx.x; gradInput += blockIdx.x * dim; gradOutput += blockIdx.x * nClusters; __syncthreads(); for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { gradOutput_data[idx] = gradOutput[idx]; } } __syncthreads(); if (tidx < poolsize) { for (int i = 0; i < nClusters; ++i) { gradInput[(int)(clusters[i*poolsize+tidx]-1)] += gradOutput[i]/poolsize; } } /* for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[tidx*poolsize+j]-1)] += gradOutput[tidx]/poolsize; __syncthreads(); } */ __syncthreads(); /* //ouch... if (tidx == 1) { for (int i = 0; i < nClusters; ++i) { // int idx = tidx + i*blockDim.x; for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[i*poolsize+j]-1)] += gradOutput[i]/poolsize; } } } */ /* for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { for (int j = 0; j < poolsize; ++j) { gradInput[(int)clusters[idx*poolsize+j]] += gradOutput_data[idx]/poolsize; } } } */ }
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 cuda_graph_avgpool_bprop(float* gradInput, const float *gradOutput, const float* clusters, const int nClusters, const int poolsize, const int dim, const int nClustersPerThread) { extern __shared__ float shared_mem[]; float* gradOutput_data = (float*)shared_mem; const int tidx = threadIdx.x; gradInput += blockIdx.x * dim; gradOutput += blockIdx.x * nClusters; __syncthreads(); for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { gradOutput_data[idx] = gradOutput[idx]; } } __syncthreads(); if (tidx < poolsize) { for (int i = 0; i < nClusters; ++i) { gradInput[(int)(clusters[i*poolsize+tidx]-1)] += gradOutput[i]/poolsize; } } /* for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[tidx*poolsize+j]-1)] += gradOutput[tidx]/poolsize; __syncthreads(); } */ __syncthreads(); /* //ouch... if (tidx == 1) { for (int i = 0; i < nClusters; ++i) { // int idx = tidx + i*blockDim.x; for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[i*poolsize+j]-1)] += gradOutput[i]/poolsize; } } } */ /* for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { for (int j = 0; j < poolsize; ++j) { gradInput[(int)clusters[idx*poolsize+j]] += gradOutput_data[idx]/poolsize; } } } */ }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .globl _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .p2align 8 .type _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii,@function _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii: s_clause 0x2 s_load_b32 s6, s[0:1], 0x18 s_load_b64 s[2:3], s[0:1], 0x8 s_load_b32 s4, s[0:1], 0x24 s_mov_b32 s9, 0 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_mul_i32 s8, s15, s6 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_lshl_b64 s[8:9], s[8:9], 2 s_add_u32 s2, s2, s8 s_addc_u32 s3, s3, s9 s_cmp_lt_i32 s4, 1 s_cbranch_scc1 .LBB0_5 s_load_b32 s5, s[0:1], 0x34 v_lshl_add_u32 v3, v0, 2, 0 v_mov_b32_e32 v1, v0 s_waitcnt lgkmcnt(0) s_and_b32 s5, s5, 0xffff s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b32 s7, s5, 2 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_or_b32 exec_lo, exec_lo, s8 v_add_nc_u32_e32 v3, s7, v3 v_add_nc_u32_e32 v1, s5, v1 s_add_i32 s4, s4, -1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s4, 0 s_cbranch_scc1 .LBB0_5 .LBB0_3: s_mov_b32 s8, exec_lo v_cmpx_gt_i32_e64 s6, v1 s_cbranch_execz .LBB0_2 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[1:2] v_add_co_u32 v4, vcc_lo, s2, v4 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v5, vcc_lo, s3, v5, vcc_lo global_load_b32 v2, v[4:5], off s_waitcnt vmcnt(0) ds_store_b32 v3, v2 s_branch .LBB0_2 .LBB0_5: s_load_b32 s4, s[0:1], 0x1c s_cmp_gt_i32 s6, 0 s_waitcnt lgkmcnt(0) s_cselect_b32 s5, -1, 0 s_barrier buffer_gl0_inv v_cmp_gt_i32_e32 vcc_lo, s4, v0 s_and_b32 s5, vcc_lo, s5 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s7, s5 s_cbranch_execz .LBB0_8 s_clause 0x2 s_load_b32 s5, s[0:1], 0x20 s_load_b64 s[8:9], s[0:1], 0x10 s_load_b64 s[0:1], s[0:1], 0x0 v_dual_mov_b32 v3, 0 :: v_dual_lshlrev_b32 v0, 2, v0 s_mov_b32 s11, 0 v_cvt_f32_i32_e32 v2, s4 s_waitcnt lgkmcnt(0) s_mul_i32 s10, s15, s5 v_add_co_u32 v0, s5, s8, v0 s_lshl_b64 s[10:11], s[10:11], 2 v_add_co_ci_u32_e64 v1, null, s9, 0, s5 s_add_u32 s8, s0, s10 s_addc_u32 s9, s1, s11 s_ashr_i32 s5, s4, 31 s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b64 s[0:1], s[4:5], 2 s_set_inst_prefetch_distance 0x1 .p2align 6 .LBB0_7: global_load_b32 v4, v[0:1], off global_load_b32 v6, v3, s[2:3] s_add_i32 s6, s6, -1 s_add_u32 s2, s2, 4 s_addc_u32 s3, s3, 0 s_cmp_lg_u32 s6, 0 s_waitcnt vmcnt(0) v_div_scale_f32 v8, null, v2, v2, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_rcp_f32_e32 v9, v8 s_waitcnt_depctr 0xfff v_fma_f32 v10, -v8, v9, 1.0 v_dual_add_f32 v4, -1.0, v4 :: v_dual_fmac_f32 v9, v10, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_i32_f32_e32 v4, v4 v_ashrrev_i32_e32 v5, 31, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[4:5], 2, v[4:5] v_add_co_u32 v4, vcc_lo, s8, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_co_ci_u32_e32 v5, vcc_lo, s9, v5, vcc_lo v_div_scale_f32 v11, vcc_lo, v6, v2, v6 global_load_b32 v7, v[4:5], off v_mul_f32_e32 v10, v11, v9 v_fma_f32 v12, -v8, v10, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v10, v12, v9 v_fma_f32 v8, -v8, v10, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_div_fmas_f32 v8, v8, v9, v10 v_add_co_u32 v0, vcc_lo, v0, s0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo v_div_fixup_f32 v6, v8, v2, v6 s_waitcnt vmcnt(0) s_delay_alu instid0(VALU_DEP_1) v_add_f32_e32 v6, v6, v7 global_store_b32 v[4:5], v6, off s_cbranch_scc1 .LBB0_7 .LBB0_8: s_set_inst_prefetch_distance 0x2 s_or_b32 exec_lo, exec_lo, s7 s_waitcnt_vscnt null, 0x0 s_barrier buffer_gl0_inv s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 13 .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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, .Lfunc_end0-_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims - .offset: 160 .size: 4 .value_kind: hidden_dynamic_lds_size .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 13 .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 cuda_graph_avgpool_bprop(float* gradInput, const float *gradOutput, const float* clusters, const int nClusters, const int poolsize, const int dim, const int nClustersPerThread) { extern __shared__ float shared_mem[]; float* gradOutput_data = (float*)shared_mem; const int tidx = threadIdx.x; gradInput += blockIdx.x * dim; gradOutput += blockIdx.x * nClusters; __syncthreads(); for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { gradOutput_data[idx] = gradOutput[idx]; } } __syncthreads(); if (tidx < poolsize) { for (int i = 0; i < nClusters; ++i) { gradInput[(int)(clusters[i*poolsize+tidx]-1)] += gradOutput[i]/poolsize; } } /* for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[tidx*poolsize+j]-1)] += gradOutput[tidx]/poolsize; __syncthreads(); } */ __syncthreads(); /* //ouch... if (tidx == 1) { for (int i = 0; i < nClusters; ++i) { // int idx = tidx + i*blockDim.x; for (int j = 0; j < poolsize; ++j) { gradInput[(int)(clusters[i*poolsize+j]-1)] += gradOutput[i]/poolsize; } } } */ /* for (int i = 0; i < nClustersPerThread; ++i) { int idx = tidx + i*blockDim.x; if (idx < nClusters) { for (int j = 0; j < poolsize; ++j) { gradInput[(int)clusters[idx*poolsize+j]] += gradOutput_data[idx]/poolsize; } } } */ }
.text .file "cuda_graph_avgpool_bprop.hip" .globl _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii # -- Begin function _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .p2align 4, 0x90 .type _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii,@function _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii: # @_Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .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 160(%rsp), %rax movq %rax, 144(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, %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 _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii, .Lfunc_end0-_Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .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 $_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, %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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii,@object # @_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .section .rodata,"a",@progbits .globl _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .p2align 3, 0x0 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii: .quad _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .size _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii" .size .L__unnamed_1, 41 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .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_00172c58_00000000-6_cuda_graph_avgpool_bprop.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii .type _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii, @function _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii: .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) 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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii(%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 _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii, .-_Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii .globl _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .type _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, @function _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z54__device_stub__Z24cuda_graph_avgpool_bpropPfPKfS1_iiiiPfPKfS1_iiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, .-_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii" .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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii(%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 "cuda_graph_avgpool_bprop.hip" .globl _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii # -- Begin function _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .p2align 4, 0x90 .type _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii,@function _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii: # @_Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .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 160(%rsp), %rax movq %rax, 144(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, %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 _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii, .Lfunc_end0-_Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .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 $_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, %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 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii,@object # @_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .section .rodata,"a",@progbits .globl _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .p2align 3, 0x0 _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii: .quad _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .size _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z24cuda_graph_avgpool_bpropPfPKfS1_iiii" .size .L__unnamed_1, 41 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z39__device_stub__cuda_graph_avgpool_bpropPfPKfS1_iiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z24cuda_graph_avgpool_bpropPfPKfS1_iiii .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> #define CUDA_CALL(...) \ do { \ cudaError_t err = __VA_ARGS__; \ if (err != cudaSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ cudaGetErrorName(err), cudaGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(cudaGetDevice(&selected_device)); cudaDeviceProp device_prop; CUDA_CALL(cudaGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #define CUDA_CALL(...) \ do { \ cudaError_t err = __VA_ARGS__; \ if (err != cudaSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ cudaGetErrorName(err), cudaGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(cudaGetDevice(&selected_device)); cudaDeviceProp device_prop; CUDA_CALL(cudaGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
.file "tmpxft_00013089_00000000-6_main.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "CUDA ERROR: %s: %s\n" .LC1: .string "w" .LC2: .string "sm" .LC3: .string "%d%d\n" .text .globl main .type main, @function main: .LFB2057: .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 $1064, %rsp .cfi_def_cfa_offset 1088 movq %fs:40, %rax movq %rax, 1048(%rsp) xorl %eax, %eax leaq 12(%rsp), %rdi call cudaGetDevice@PLT testl %eax, %eax jne .L9 leaq 16(%rsp), %rdi movl 12(%rsp), %esi call cudaGetDeviceProperties_v2@PLT movl %eax, %ebx testl %eax, %eax jne .L10 leaq .LC1(%rip), %rsi leaq .LC2(%rip), %rdi call fopen@PLT movq %rax, %rbx movl 380(%rsp), %r8d movl 376(%rsp), %ecx leaq .LC3(%rip), %rdx movl $2, %esi movq %rax, %rdi movl $0, %eax call __fprintf_chk@PLT movq %rbx, %rdi call fclose@PLT movl $0, %eax .L3: movq 1048(%rsp), %rdx subq %fs:40, %rdx jne .L11 addq $1064, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L9: .cfi_restore_state movl %eax, %ebx movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rbp movl %ebx, %edi call cudaGetErrorName@PLT movq %rax, %rdx movq %rbp, %rcx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %eax jmp .L3 .L10: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rbp movl %ebx, %edi call cudaGetErrorName@PLT movq %rax, %rdx movq %rbp, %rcx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %eax jmp .L3 .L11: 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> #define CUDA_CALL(...) \ do { \ cudaError_t err = __VA_ARGS__; \ if (err != cudaSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ cudaGetErrorName(err), cudaGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(cudaGetDevice(&selected_device)); cudaDeviceProp device_prop; CUDA_CALL(cudaGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
#include <hip/hip_runtime.h> #include <stdio.h> #define CUDA_CALL(...) \ do { \ hipError_t err = __VA_ARGS__; \ if (err != hipSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ hipGetErrorName(err), hipGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(hipGetDevice(&selected_device)); hipDeviceProp_t device_prop; CUDA_CALL(hipGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #define CUDA_CALL(...) \ do { \ hipError_t err = __VA_ARGS__; \ if (err != hipSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ hipGetErrorName(err), hipGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(hipGetDevice(&selected_device)); hipDeviceProp_t device_prop; CUDA_CALL(hipGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
.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> #define CUDA_CALL(...) \ do { \ hipError_t err = __VA_ARGS__; \ if (err != hipSuccess) \ { \ printf("CUDA ERROR: %s: %s\n", \ hipGetErrorName(err), hipGetErrorString(err)); \ return err; \ } \ } while (false) int main() { int selected_device; CUDA_CALL(hipGetDevice(&selected_device)); hipDeviceProp_t device_prop; CUDA_CALL(hipGetDeviceProperties(&device_prop, selected_device)); FILE * output = fopen("sm", "w"); fprintf(output, "%d%d\n", device_prop.major, device_prop.minor); fclose(output); }
.text .file "main.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 %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 1520 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 leaq 12(%rsp), %rdi callq hipGetDevice movl %eax, %ebp testl %eax, %eax je .LBB0_1 # %bb.2: movl %ebp, %edi callq hipGetErrorName movq %rax, %rbx movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rbx, %rsi movq %rax, %rdx xorl %eax, %eax callq printf movl %ebp, %ebx testl %ebp, %ebp jne .LBB0_8 jmp .LBB0_4 .LBB0_1: xorl %ebx, %ebx testl %ebp, %ebp jne .LBB0_8 .LBB0_4: movl 12(%rsp), %esi leaq 16(%rsp), %rdi callq hipGetDevicePropertiesR0600 movl %eax, %ebp testl %eax, %eax je .LBB0_6 # %bb.5: movl %ebp, %edi callq hipGetErrorName movq %rax, %rbx movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rbx, %rsi movq %rax, %rdx xorl %eax, %eax callq printf movl %ebp, %ebx .LBB0_6: testl %ebp, %ebp jne .LBB0_8 # %bb.7: movl $.L.str.1, %edi movl $.L.str.2, %esi callq fopen movq %rax, %r14 movl 376(%rsp), %edx movl 380(%rsp), %ecx movl $.L.str.3, %esi movq %rax, %rdi xorl %eax, %eax callq fprintf movq %r14, %rdi callq fclose .LBB0_8: movl %ebx, %eax addq $1488, %rsp # imm = 0x5D0 .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_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 "CUDA ERROR: %s: %s\n" .size .L.str, 20 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "sm" .size .L.str.1, 3 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "w" .size .L.str.2, 2 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "%d%d\n" .size .L.str.3, 6 .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_00013089_00000000-6_main.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "CUDA ERROR: %s: %s\n" .LC1: .string "w" .LC2: .string "sm" .LC3: .string "%d%d\n" .text .globl main .type main, @function main: .LFB2057: .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 $1064, %rsp .cfi_def_cfa_offset 1088 movq %fs:40, %rax movq %rax, 1048(%rsp) xorl %eax, %eax leaq 12(%rsp), %rdi call cudaGetDevice@PLT testl %eax, %eax jne .L9 leaq 16(%rsp), %rdi movl 12(%rsp), %esi call cudaGetDeviceProperties_v2@PLT movl %eax, %ebx testl %eax, %eax jne .L10 leaq .LC1(%rip), %rsi leaq .LC2(%rip), %rdi call fopen@PLT movq %rax, %rbx movl 380(%rsp), %r8d movl 376(%rsp), %ecx leaq .LC3(%rip), %rdx movl $2, %esi movq %rax, %rdi movl $0, %eax call __fprintf_chk@PLT movq %rbx, %rdi call fclose@PLT movl $0, %eax .L3: movq 1048(%rsp), %rdx subq %fs:40, %rdx jne .L11 addq $1064, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L9: .cfi_restore_state movl %eax, %ebx movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rbp movl %ebx, %edi call cudaGetErrorName@PLT movq %rax, %rdx movq %rbp, %rcx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %eax jmp .L3 .L10: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rbp movl %ebx, %edi call cudaGetErrorName@PLT movq %rax, %rdx movq %rbp, %rcx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %eax jmp .L3 .L11: 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:
.text .file "main.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 %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 1520 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 leaq 12(%rsp), %rdi callq hipGetDevice movl %eax, %ebp testl %eax, %eax je .LBB0_1 # %bb.2: movl %ebp, %edi callq hipGetErrorName movq %rax, %rbx movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rbx, %rsi movq %rax, %rdx xorl %eax, %eax callq printf movl %ebp, %ebx testl %ebp, %ebp jne .LBB0_8 jmp .LBB0_4 .LBB0_1: xorl %ebx, %ebx testl %ebp, %ebp jne .LBB0_8 .LBB0_4: movl 12(%rsp), %esi leaq 16(%rsp), %rdi callq hipGetDevicePropertiesR0600 movl %eax, %ebp testl %eax, %eax je .LBB0_6 # %bb.5: movl %ebp, %edi callq hipGetErrorName movq %rax, %rbx movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rbx, %rsi movq %rax, %rdx xorl %eax, %eax callq printf movl %ebp, %ebx .LBB0_6: testl %ebp, %ebp jne .LBB0_8 # %bb.7: movl $.L.str.1, %edi movl $.L.str.2, %esi callq fopen movq %rax, %r14 movl 376(%rsp), %edx movl 380(%rsp), %ecx movl $.L.str.3, %esi movq %rax, %rdi xorl %eax, %eax callq fprintf movq %r14, %rdi callq fclose .LBB0_8: movl %ebx, %eax addq $1488, %rsp # imm = 0x5D0 .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_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 "CUDA ERROR: %s: %s\n" .size .L.str, 20 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "sm" .size .L.str.1, 3 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "w" .size .L.str.2, 2 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "%d%d\n" .size .L.str.3, 6 .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 "includes.h" __global__ void gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
code for sm_80 Function : _Z4gemmPfS_S_iii .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 R13, SR_CTAID.X ; /* 0x00000000000d7919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff047624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC UR5, c[0x0][0x180] ; /* 0x0000600000057ab9 */ /* 0x000fe20000000800 */ /*0040*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0050*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e620000002100 */ /*0060*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe20000000a00 */ /*0070*/ ISETP.GE.AND P1, PT, R4, 0x10, PT ; /* 0x000000100400780c */ /* 0x000fe20003f26270 */ /*0080*/ USHF.L.U32 UR5, UR5, 0x4, URZ ; /* 0x0000000405057899 */ /* 0x000fe2000800063f */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000ea80000002200 */ /*00a0*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000ea20000002600 */ /*00b0*/ IMAD.SHL.U32 R13, R13, 0x10, RZ ; /* 0x000000100d0d7824 */ /* 0x001fca00078e00ff */ /*00c0*/ IADD3 R5, R13, R0, RZ ; /* 0x000000000d057210 */ /* 0x002fc80007ffe0ff */ /*00d0*/ ISETP.GE.AND P0, PT, R5, c[0x0][0x180], PT ; /* 0x0000600005007a0c */ /* 0x000fe40003f06270 */ /*00e0*/ LEA R6, R2, R3, 0x4 ; /* 0x0000000302067211 */ /* 0x004fc800078e20ff */ /*00f0*/ ISETP.GE.OR P0, PT, R6, c[0x0][0x178], P0 ; /* 0x00005e0006007a0c */ /* 0x000fe20000706670 */ /*0100*/ @!P1 BRA 0x560 ; /* 0x0000045000009947 */ /* 0x000ff60003800000 */ /*0110*/ SHF.R.S32.HI R4, RZ, 0x1f, R4 ; /* 0x0000001fff047819 */ /* 0x000fe20000011404 */ /*0120*/ IMAD R19, R2, c[0x0][0x17c], RZ ; /* 0x00005f0002137a24 */ /* 0x000fe200078e02ff */ /*0130*/ SHF.L.U32 R15, R3.reuse, 0x6, RZ ; /* 0x00000006030f7819 */ /* 0x040fe200000006ff */ /*0140*/ IMAD R14, R3.reuse, c[0x0][0x17c], R0.reuse ; /* 0x00005f00030e7a24 */ /* 0x140fe200078e0200 */ /*0150*/ LEA.HI R4, R4, c[0x0][0x17c], RZ, 0x4 ; /* 0x00005f0004047a11 */ /* 0x000fe200078f20ff */ /*0160*/ IMAD R17, R3, c[0x0][0x180], R0 ; /* 0x0000600003117a24 */ /* 0x000fe200078e0200 */ /*0170*/ MOV R12, R13 ; /* 0x0000000d000c7202 */ /* 0x000fe20000000f00 */ /*0180*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0190*/ LEA R18, R0, R15, 0x2 ; /* 0x0000000f00127211 */ /* 0x000fe200078e10ff */ /*01a0*/ IMAD.SHL.U32 R19, R19, 0x10, RZ ; /* 0x0000001013137824 */ /* 0x000fe200078e00ff */ /*01b0*/ SHF.R.S32.HI R16, RZ, 0x4, R4 ; /* 0x00000004ff107819 */ /* 0x000fe20000011404 */ /*01c0*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fc40008000000 */ /*01d0*/ SHF.R.S32.HI R4, RZ, 0x1f, R12 ; /* 0x0000001fff047819 */ /* 0x000fe4000001140c */ /*01e0*/ IADD3 R6, P2, R17, R12, RZ ; /* 0x0000000c11067210 */ /* 0x000fc40007f5e0ff */ /*01f0*/ SHF.R.S32.HI R5, RZ, 0x1f, R19 ; /* 0x0000001fff057819 */ /* 0x000fe40000011413 */ /*0200*/ IADD3 R8, P1, R14.reuse, R19, RZ ; /* 0x000000130e087210 */ /* 0x040fe40007f3e0ff */ /*0210*/ LEA.HI.X.SX32 R7, R17, R4, 0x1, P2 ; /* 0x0000000411077211 */ /* 0x000fe400010f0eff */ /*0220*/ LEA.HI.X.SX32 R5, R14, R5, 0x1, P1 ; /* 0x000000050e057211 */ /* 0x000fe400008f0eff */ /*0230*/ LEA R24, P2, R6, c[0x0][0x168], 0x2 ; /* 0x00005a0006187a11 */ /* 0x000fe400078410ff */ /*0240*/ LEA R4, P1, R8, c[0x0][0x160], 0x2 ; /* 0x0000580008047a11 */ /* 0x000fc400078210ff */ /*0250*/ LEA.HI.X R25, R6, c[0x0][0x16c], R7, 0x2, P2 ; /* 0x00005b0006197a11 */ /* 0x000fe400010f1407 */ /*0260*/ LEA.HI.X R5, R8, c[0x0][0x164], R5, 0x2, P1 ; /* 0x0000590008057a11 */ /* 0x000fc800008f1405 */ /*0270*/ LDG.E R25, [R24.64] ; /* 0x0000000618197981 */ /* 0x000ea8000c1e1900 */ /*0280*/ LDG.E R29, [R4.64] ; /* 0x00000006041d7981 */ /* 0x000ee2000c1e1900 */ /*0290*/ UIADD3 UR4, UR4, 0x1, URZ ; /* 0x0000000104047890 */ /* 0x000fcc000fffe03f */ /*02a0*/ ISETP.LE.AND P1, PT, R16, UR4, PT ; /* 0x0000000410007c0c */ /* 0x000fe4000bf23270 */ /*02b0*/ IADD3 R12, R12, UR5, RZ ; /* 0x000000050c0c7c10 */ /* 0x000fe4000fffe0ff */ /*02c0*/ IADD3 R19, R19, 0x10, RZ ; /* 0x0000001013137810 */ /* 0x000fe20007ffe0ff */ /*02d0*/ STS [R18+0x400], R25 ; /* 0x0004001912007388 */ /* 0x004fe80000000800 */ /*02e0*/ STS [R18], R29 ; /* 0x0000001d12007388 */ /* 0x008fe80000000800 */ /*02f0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0300*/ LDS R22, [R0.X4+0x400] ; /* 0x0004000000167984 */ /* 0x000fe80000004800 */ /*0310*/ LDS.128 R8, [R15] ; /* 0x000000000f087984 */ /* 0x000e280000000c00 */ /*0320*/ LDS R28, [R0.X4+0x440] ; /* 0x00044000001c7984 */ /* 0x000e680000004800 */ /*0330*/ LDS R27, [R0.X4+0x480] ; /* 0x00048000001b7984 */ /* 0x000ea80000004800 */ /*0340*/ LDS R26, [R0.X4+0x4c0] ; /* 0x0004c000001a7984 */ /* 0x000ee80000004800 */ /*0350*/ LDS R23, [R0.X4+0x500] ; /* 0x0005000000177984 */ /* 0x000fe80000004800 */ /*0360*/ LDS.128 R4, [R15+0x10] ; /* 0x000010000f047984 */ /* 0x000f280000000c00 */ /*0370*/ LDS R20, [R0.X4+0x540] ; /* 0x0005400000147984 */ /* 0x000f680000004800 */ /*0380*/ LDS R25, [R0.X4+0x580] ; /* 0x0005800000197984 */ /* 0x000f680000004800 */ /*0390*/ LDS R24, [R0.X4+0x640] ; /* 0x0006400000187984 */ /* 0x000fe20000004800 */ /*03a0*/ FFMA R8, R22, R8, R21 ; /* 0x0000000816087223 */ /* 0x001fc60000000015 */ /*03b0*/ LDS R22, [R0.X4+0x5c0] ; /* 0x0005c00000167984 */ /* 0x000e220000004800 */ /*03c0*/ FFMA R8, R28, R9, R8 ; /* 0x000000091c087223 */ /* 0x002fc60000000008 */ /*03d0*/ LDS R21, [R0.X4+0x600] ; /* 0x0006000000157984 */ /* 0x000fe20000004800 */ /*03e0*/ FFMA R8, R27, R10, R8 ; /* 0x0000000a1b087223 */ /* 0x004fc80000000008 */ /*03f0*/ FFMA R26, R26, R11, R8 ; /* 0x0000000b1a1a7223 */ /* 0x008fe40000000008 */ /*0400*/ LDS.128 R8, [R15+0x20] ; /* 0x000020000f087984 */ /* 0x000e640000000c00 */ /*0410*/ FFMA R4, R23, R4, R26 ; /* 0x0000000417047223 */ /* 0x010fe4000000001a */ /*0420*/ LDS R23, [R0.X4+0x680] ; /* 0x0006800000177984 */ /* 0x000ea40000004800 */ /*0430*/ FFMA R4, R20, R5, R4 ; /* 0x0000000514047223 */ /* 0x020fe40000000004 */ /*0440*/ LDS R20, [R0.X4+0x6c0] ; /* 0x0006c00000147984 */ /* 0x000ee40000004800 */ /*0450*/ FFMA R4, R25, R6, R4 ; /* 0x0000000619047223 */ /* 0x000fc40000000004 */ /*0460*/ LDS R25, [R0.X4+0x700] ; /* 0x0007000000197984 */ /* 0x000fe40000004800 */ /*0470*/ FFMA R26, R22, R7, R4 ; /* 0x00000007161a7223 */ /* 0x001fe40000000004 */ /*0480*/ LDS.128 R4, [R15+0x30] ; /* 0x000030000f047984 */ /* 0x000e280000000c00 */ /*0490*/ LDS R22, [R0.X4+0x740] ; /* 0x0007400000167984 */ /* 0x000f220000004800 */ /*04a0*/ FFMA R26, R21, R8, R26 ; /* 0x00000008151a7223 */ /* 0x002fc6000000001a */ /*04b0*/ LDS R21, [R0.X4+0x780] ; /* 0x0007800000157984 */ /* 0x000e680000004800 */ /*04c0*/ LDS R8, [R0.X4+0x7c0] ; /* 0x0007c00000087984 */ /* 0x000f620000004800 */ /*04d0*/ FFMA R9, R24, R9, R26 ; /* 0x0000000918097223 */ /* 0x000fc8000000001a */ /*04e0*/ FFMA R9, R23, R10, R9 ; /* 0x0000000a17097223 */ /* 0x004fc80000000009 */ /*04f0*/ FFMA R9, R20, R11, R9 ; /* 0x0000000b14097223 */ /* 0x008fc80000000009 */ /*0500*/ FFMA R4, R25, R4, R9 ; /* 0x0000000419047223 */ /* 0x001fc80000000009 */ /*0510*/ FFMA R4, R22, R5, R4 ; /* 0x0000000516047223 */ /* 0x010fc80000000004 */ /*0520*/ FFMA R21, R21, R6, R4 ; /* 0x0000000615157223 */ /* 0x002fc80000000004 */ /*0530*/ FFMA R21, R8, R7, R21 ; /* 0x0000000708157223 */ /* 0x020fe20000000015 */ /*0540*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0550*/ @!P1 BRA 0x1d0 ; /* 0xfffffc7000009947 */ /* 0x000fea000383ffff */ /*0560*/ IMAD R2, R2, UR5, R13 ; /* 0x0000000502027c24 */ /* 0x000fe2000f8e020d */ /*0570*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fec0003800000 */ /*0580*/ IMAD R5, R3, c[0x0][0x180], R0 ; /* 0x0000600003057a24 */ /* 0x000fe200078e0200 */ /*0590*/ SHF.R.S32.HI R0, RZ, 0x1f, R2 ; /* 0x0000001fff007819 */ /* 0x000fc80000011402 */ /*05a0*/ IADD3 R3, P0, R5, R2, RZ ; /* 0x0000000205037210 */ /* 0x000fc80007f1e0ff */ /*05b0*/ LEA.HI.X.SX32 R0, R5, R0, 0x1, P0 ; /* 0x0000000005007211 */ /* 0x000fe400000f0eff */ /*05c0*/ LEA R2, P0, R3, c[0x0][0x170], 0x2 ; /* 0x00005c0003027a11 */ /* 0x000fc800078010ff */ /*05d0*/ LEA.HI.X R3, R3, c[0x0][0x174], R0, 0x2, P0 ; /* 0x00005d0003037a11 */ /* 0x000fca00000f1400 */ /*05e0*/ STG.E [R2.64], R21 ; /* 0x0000001502007986 */ /* 0x000fe2000c101906 */ /*05f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0600*/ BRA 0x600; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0610*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
.file "tmpxft_0015829e_00000000-6_gemm.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 _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii .type _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii, @function _Z30__device_stub__Z4gemmPfS_S_iiiPfS_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 _Z4gemmPfS_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 _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii, .-_Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii .globl _Z4gemmPfS_S_iii .type _Z4gemmPfS_S_iii, @function _Z4gemmPfS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4gemmPfS_S_iii, .-_Z4gemmPfS_S_iii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4gemmPfS_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 _Z4gemmPfS_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 gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
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 gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4gemmPfS_S_iii .globl _Z4gemmPfS_S_iii .p2align 8 .type _Z4gemmPfS_S_iii,@function _Z4gemmPfS_S_iii: s_load_b64 s[2:3], s[0:1], 0x1c v_mov_b32_e32 v2, 0 v_bfe_u32 v3, v0, 10, 10 v_and_b32_e32 v0, 0x3ff, v0 s_lshl_b32 s4, s14, 4 s_waitcnt lgkmcnt(0) s_lshl_b32 s5, s3, 4 s_cmp_lt_i32 s2, 16 s_cbranch_scc1 .LBB0_5 s_load_b128 s[8:11], s[0:1], 0x0 v_mad_u64_u32 v[6:7], null, v3, s3, v[0:1] v_mad_u64_u32 v[1:2], null, v3, s2, v[0:1] v_dual_mov_b32 v2, 0 :: v_dual_lshlrev_b32 v5, 2, v0 v_lshlrev_b32_e32 v4, 6, v3 s_ashr_i32 s6, s2, 31 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_ashrrev_i32_e32 v7, 31, v6 s_lshr_b32 s6, s6, 28 v_lshlrev_b64 v[8:9], 2, v[1:2] v_add_nc_u32_e32 v1, v4, v5 v_add_nc_u32_e32 v5, 0x400, v5 v_lshlrev_b64 v[10:11], 2, v[6:7] s_add_i32 s6, s2, s6 s_mul_i32 s2, s15, s2 s_ashr_i32 s6, s6, 4 s_mov_b32 s7, 0 s_waitcnt lgkmcnt(0) v_add_co_u32 v6, vcc_lo, s8, v8 v_add_co_ci_u32_e32 v7, vcc_lo, s9, v9, vcc_lo v_add_co_u32 v8, vcc_lo, s10, v10 v_add_co_ci_u32_e32 v9, vcc_lo, s11, v11, vcc_lo v_add_nc_u32_e32 v10, v5, v4 s_set_inst_prefetch_distance 0x1 .p2align 6 .LBB0_2: s_add_i32 s8, s7, s2 s_mul_i32 s9, s7, s5 s_lshl_b32 s8, s8, 4 s_add_i32 s10, s9, s4 s_ashr_i32 s9, s8, 31 s_ashr_i32 s11, s10, 31 s_lshl_b64 s[8:9], s[8:9], 2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(SALU_CYCLE_1) v_add_co_u32 v11, vcc_lo, v6, s8 v_add_co_ci_u32_e32 v12, vcc_lo, s9, v7, vcc_lo s_lshl_b64 s[8:9], s[10:11], 2 v_add_co_u32 v13, vcc_lo, v8, s8 v_add_co_ci_u32_e32 v14, vcc_lo, s9, v9, vcc_lo global_load_b32 v12, v[11:12], off global_load_b32 v13, v[13:14], off v_mov_b32_e32 v11, v5 s_mov_b32 s8, 0 s_waitcnt vmcnt(1) ds_store_b32 v1, v12 s_waitcnt vmcnt(0) ds_store_b32 v10, v13 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_3: v_add_nc_u32_e32 v12, s8, v4 s_add_i32 s8, s8, 4 ds_load_b32 v13, v11 ds_load_b32 v12, v12 v_add_nc_u32_e32 v11, 64, v11 s_cmp_eq_u32 s8, 64 s_waitcnt lgkmcnt(0) v_fmac_f32_e32 v2, v12, v13 s_cbranch_scc0 .LBB0_3 s_add_i32 s7, s7, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s7, s6 s_barrier buffer_gl0_inv s_cbranch_scc0 .LBB0_2 .LBB0_5: s_set_inst_prefetch_distance 0x2 s_load_b32 s2, s[0:1], 0x18 v_add_nc_u32_e32 v1, s4, v0 v_lshl_add_u32 v4, s15, 4, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cmp_gt_i32_e32 vcc_lo, s3, v1 s_waitcnt lgkmcnt(0) v_cmp_gt_i32_e64 s2, s2, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, vcc_lo, s2 s_and_saveexec_b32 s6, s2 s_cbranch_execz .LBB0_7 s_load_b64 s[0:1], s[0:1], 0x10 v_mad_u64_u32 v[4:5], null, v3, s3, v[0:1] s_mul_i32 s2, s5, s15 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s2, s2, s4 s_ashr_i32 s3, s2, 31 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v5, 31, v4 s_lshl_b64 s[2:3], s[2:3], 2 v_lshlrev_b64 v[0:1], 2, v[4:5] s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 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_7: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4gemmPfS_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 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 _Z4gemmPfS_S_iii, .Lfunc_end0-_Z4gemmPfS_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: _Z4gemmPfS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4gemmPfS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void gemm(float* A, float* B, float* C, int m, int n, int k) { // Block row and column int blockRow = blockIdx.y; int blockCol = blockIdx.x; // Thread row and column within Csub int row = threadIdx.y; int col = threadIdx.x; // Each thread block computes one sub-matrix Csub of C float* Csub = &C[BLOCK_SIZE * k * blockRow + BLOCK_SIZE * blockCol]; // Shared memory used to store Asub and Bsub respectively __shared__ float As[BLOCK_SIZE][BLOCK_SIZE]; __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE]; // Each thread computes one element of Csub // by accumulating results into Cvalue // block_size = 16 -> 256 threads, one per Csub element float Cvalue = 0.0; // Loop over all the sub-matrices of A and B that are // required to compute Csub // Multiply each pair of sub-matrices together // and accumulate the results for (int i = 0; i < (n / BLOCK_SIZE); ++i) { // Get sub-matrix Asub of A float* Asub = &A[BLOCK_SIZE * blockRow * n + BLOCK_SIZE * i]; // Get sub-matrix Bsub of B float* Bsub = &B[BLOCK_SIZE * k * i + BLOCK_SIZE * blockCol]; // Load Asub and Bsub from device memory to shared memory // Each thread loads one element of each sub-matrix As[row][col] = Asub[row*n+col]; Bs[row][col] = Bsub[row*k+col]; // Synchronize to make sure the sub-matrices are loaded // before starting the computation __syncthreads(); // Multiply Asub and Bsub together for (int j = 0; j < BLOCK_SIZE; ++j) Cvalue += As[row][j] * Bs[j][col]; // Synchronize to make sure that the preceding // computation is done before loading two new // sub-matrices of A and B in the next iteration __syncthreads(); } // Write Csub to device memory // Each thread writes one element if(col + blockCol* BLOCK_SIZE< k && row + blockRow* BLOCK_SIZE< m) Csub[row*k+col] = Cvalue; }
.text .file "gemm.hip" .globl _Z19__device_stub__gemmPfS_S_iii # -- Begin function _Z19__device_stub__gemmPfS_S_iii .p2align 4, 0x90 .type _Z19__device_stub__gemmPfS_S_iii,@function _Z19__device_stub__gemmPfS_S_iii: # @_Z19__device_stub__gemmPfS_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 $_Z4gemmPfS_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 _Z19__device_stub__gemmPfS_S_iii, .Lfunc_end0-_Z19__device_stub__gemmPfS_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 $_Z4gemmPfS_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 _Z4gemmPfS_S_iii,@object # @_Z4gemmPfS_S_iii .section .rodata,"a",@progbits .globl _Z4gemmPfS_S_iii .p2align 3, 0x0 _Z4gemmPfS_S_iii: .quad _Z19__device_stub__gemmPfS_S_iii .size _Z4gemmPfS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4gemmPfS_S_iii" .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 _Z19__device_stub__gemmPfS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4gemmPfS_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 : _Z4gemmPfS_S_iii .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 R13, SR_CTAID.X ; /* 0x00000000000d7919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x17c] ; /* 0x00005f00ff047624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC UR5, c[0x0][0x180] ; /* 0x0000600000057ab9 */ /* 0x000fe20000000800 */ /*0040*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0050*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e620000002100 */ /*0060*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe20000000a00 */ /*0070*/ ISETP.GE.AND P1, PT, R4, 0x10, PT ; /* 0x000000100400780c */ /* 0x000fe20003f26270 */ /*0080*/ USHF.L.U32 UR5, UR5, 0x4, URZ ; /* 0x0000000405057899 */ /* 0x000fe2000800063f */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000ea80000002200 */ /*00a0*/ S2R R2, SR_CTAID.Y ; /* 0x0000000000027919 */ /* 0x000ea20000002600 */ /*00b0*/ IMAD.SHL.U32 R13, R13, 0x10, RZ ; /* 0x000000100d0d7824 */ /* 0x001fca00078e00ff */ /*00c0*/ IADD3 R5, R13, R0, RZ ; /* 0x000000000d057210 */ /* 0x002fc80007ffe0ff */ /*00d0*/ ISETP.GE.AND P0, PT, R5, c[0x0][0x180], PT ; /* 0x0000600005007a0c */ /* 0x000fe40003f06270 */ /*00e0*/ LEA R6, R2, R3, 0x4 ; /* 0x0000000302067211 */ /* 0x004fc800078e20ff */ /*00f0*/ ISETP.GE.OR P0, PT, R6, c[0x0][0x178], P0 ; /* 0x00005e0006007a0c */ /* 0x000fe20000706670 */ /*0100*/ @!P1 BRA 0x560 ; /* 0x0000045000009947 */ /* 0x000ff60003800000 */ /*0110*/ SHF.R.S32.HI R4, RZ, 0x1f, R4 ; /* 0x0000001fff047819 */ /* 0x000fe20000011404 */ /*0120*/ IMAD R19, R2, c[0x0][0x17c], RZ ; /* 0x00005f0002137a24 */ /* 0x000fe200078e02ff */ /*0130*/ SHF.L.U32 R15, R3.reuse, 0x6, RZ ; /* 0x00000006030f7819 */ /* 0x040fe200000006ff */ /*0140*/ IMAD R14, R3.reuse, c[0x0][0x17c], R0.reuse ; /* 0x00005f00030e7a24 */ /* 0x140fe200078e0200 */ /*0150*/ LEA.HI R4, R4, c[0x0][0x17c], RZ, 0x4 ; /* 0x00005f0004047a11 */ /* 0x000fe200078f20ff */ /*0160*/ IMAD R17, R3, c[0x0][0x180], R0 ; /* 0x0000600003117a24 */ /* 0x000fe200078e0200 */ /*0170*/ MOV R12, R13 ; /* 0x0000000d000c7202 */ /* 0x000fe20000000f00 */ /*0180*/ IMAD.MOV.U32 R21, RZ, RZ, RZ ; /* 0x000000ffff157224 */ /* 0x000fe200078e00ff */ /*0190*/ LEA R18, R0, R15, 0x2 ; /* 0x0000000f00127211 */ /* 0x000fe200078e10ff */ /*01a0*/ IMAD.SHL.U32 R19, R19, 0x10, RZ ; /* 0x0000001013137824 */ /* 0x000fe200078e00ff */ /*01b0*/ SHF.R.S32.HI R16, RZ, 0x4, R4 ; /* 0x00000004ff107819 */ /* 0x000fe20000011404 */ /*01c0*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fc40008000000 */ /*01d0*/ SHF.R.S32.HI R4, RZ, 0x1f, R12 ; /* 0x0000001fff047819 */ /* 0x000fe4000001140c */ /*01e0*/ IADD3 R6, P2, R17, R12, RZ ; /* 0x0000000c11067210 */ /* 0x000fc40007f5e0ff */ /*01f0*/ SHF.R.S32.HI R5, RZ, 0x1f, R19 ; /* 0x0000001fff057819 */ /* 0x000fe40000011413 */ /*0200*/ IADD3 R8, P1, R14.reuse, R19, RZ ; /* 0x000000130e087210 */ /* 0x040fe40007f3e0ff */ /*0210*/ LEA.HI.X.SX32 R7, R17, R4, 0x1, P2 ; /* 0x0000000411077211 */ /* 0x000fe400010f0eff */ /*0220*/ LEA.HI.X.SX32 R5, R14, R5, 0x1, P1 ; /* 0x000000050e057211 */ /* 0x000fe400008f0eff */ /*0230*/ LEA R24, P2, R6, c[0x0][0x168], 0x2 ; /* 0x00005a0006187a11 */ /* 0x000fe400078410ff */ /*0240*/ LEA R4, P1, R8, c[0x0][0x160], 0x2 ; /* 0x0000580008047a11 */ /* 0x000fc400078210ff */ /*0250*/ LEA.HI.X R25, R6, c[0x0][0x16c], R7, 0x2, P2 ; /* 0x00005b0006197a11 */ /* 0x000fe400010f1407 */ /*0260*/ LEA.HI.X R5, R8, c[0x0][0x164], R5, 0x2, P1 ; /* 0x0000590008057a11 */ /* 0x000fc800008f1405 */ /*0270*/ LDG.E R25, [R24.64] ; /* 0x0000000618197981 */ /* 0x000ea8000c1e1900 */ /*0280*/ LDG.E R29, [R4.64] ; /* 0x00000006041d7981 */ /* 0x000ee2000c1e1900 */ /*0290*/ UIADD3 UR4, UR4, 0x1, URZ ; /* 0x0000000104047890 */ /* 0x000fcc000fffe03f */ /*02a0*/ ISETP.LE.AND P1, PT, R16, UR4, PT ; /* 0x0000000410007c0c */ /* 0x000fe4000bf23270 */ /*02b0*/ IADD3 R12, R12, UR5, RZ ; /* 0x000000050c0c7c10 */ /* 0x000fe4000fffe0ff */ /*02c0*/ IADD3 R19, R19, 0x10, RZ ; /* 0x0000001013137810 */ /* 0x000fe20007ffe0ff */ /*02d0*/ STS [R18+0x400], R25 ; /* 0x0004001912007388 */ /* 0x004fe80000000800 */ /*02e0*/ STS [R18], R29 ; /* 0x0000001d12007388 */ /* 0x008fe80000000800 */ /*02f0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0300*/ LDS R22, [R0.X4+0x400] ; /* 0x0004000000167984 */ /* 0x000fe80000004800 */ /*0310*/ LDS.128 R8, [R15] ; /* 0x000000000f087984 */ /* 0x000e280000000c00 */ /*0320*/ LDS R28, [R0.X4+0x440] ; /* 0x00044000001c7984 */ /* 0x000e680000004800 */ /*0330*/ LDS R27, [R0.X4+0x480] ; /* 0x00048000001b7984 */ /* 0x000ea80000004800 */ /*0340*/ LDS R26, [R0.X4+0x4c0] ; /* 0x0004c000001a7984 */ /* 0x000ee80000004800 */ /*0350*/ LDS R23, [R0.X4+0x500] ; /* 0x0005000000177984 */ /* 0x000fe80000004800 */ /*0360*/ LDS.128 R4, [R15+0x10] ; /* 0x000010000f047984 */ /* 0x000f280000000c00 */ /*0370*/ LDS R20, [R0.X4+0x540] ; /* 0x0005400000147984 */ /* 0x000f680000004800 */ /*0380*/ LDS R25, [R0.X4+0x580] ; /* 0x0005800000197984 */ /* 0x000f680000004800 */ /*0390*/ LDS R24, [R0.X4+0x640] ; /* 0x0006400000187984 */ /* 0x000fe20000004800 */ /*03a0*/ FFMA R8, R22, R8, R21 ; /* 0x0000000816087223 */ /* 0x001fc60000000015 */ /*03b0*/ LDS R22, [R0.X4+0x5c0] ; /* 0x0005c00000167984 */ /* 0x000e220000004800 */ /*03c0*/ FFMA R8, R28, R9, R8 ; /* 0x000000091c087223 */ /* 0x002fc60000000008 */ /*03d0*/ LDS R21, [R0.X4+0x600] ; /* 0x0006000000157984 */ /* 0x000fe20000004800 */ /*03e0*/ FFMA R8, R27, R10, R8 ; /* 0x0000000a1b087223 */ /* 0x004fc80000000008 */ /*03f0*/ FFMA R26, R26, R11, R8 ; /* 0x0000000b1a1a7223 */ /* 0x008fe40000000008 */ /*0400*/ LDS.128 R8, [R15+0x20] ; /* 0x000020000f087984 */ /* 0x000e640000000c00 */ /*0410*/ FFMA R4, R23, R4, R26 ; /* 0x0000000417047223 */ /* 0x010fe4000000001a */ /*0420*/ LDS R23, [R0.X4+0x680] ; /* 0x0006800000177984 */ /* 0x000ea40000004800 */ /*0430*/ FFMA R4, R20, R5, R4 ; /* 0x0000000514047223 */ /* 0x020fe40000000004 */ /*0440*/ LDS R20, [R0.X4+0x6c0] ; /* 0x0006c00000147984 */ /* 0x000ee40000004800 */ /*0450*/ FFMA R4, R25, R6, R4 ; /* 0x0000000619047223 */ /* 0x000fc40000000004 */ /*0460*/ LDS R25, [R0.X4+0x700] ; /* 0x0007000000197984 */ /* 0x000fe40000004800 */ /*0470*/ FFMA R26, R22, R7, R4 ; /* 0x00000007161a7223 */ /* 0x001fe40000000004 */ /*0480*/ LDS.128 R4, [R15+0x30] ; /* 0x000030000f047984 */ /* 0x000e280000000c00 */ /*0490*/ LDS R22, [R0.X4+0x740] ; /* 0x0007400000167984 */ /* 0x000f220000004800 */ /*04a0*/ FFMA R26, R21, R8, R26 ; /* 0x00000008151a7223 */ /* 0x002fc6000000001a */ /*04b0*/ LDS R21, [R0.X4+0x780] ; /* 0x0007800000157984 */ /* 0x000e680000004800 */ /*04c0*/ LDS R8, [R0.X4+0x7c0] ; /* 0x0007c00000087984 */ /* 0x000f620000004800 */ /*04d0*/ FFMA R9, R24, R9, R26 ; /* 0x0000000918097223 */ /* 0x000fc8000000001a */ /*04e0*/ FFMA R9, R23, R10, R9 ; /* 0x0000000a17097223 */ /* 0x004fc80000000009 */ /*04f0*/ FFMA R9, R20, R11, R9 ; /* 0x0000000b14097223 */ /* 0x008fc80000000009 */ /*0500*/ FFMA R4, R25, R4, R9 ; /* 0x0000000419047223 */ /* 0x001fc80000000009 */ /*0510*/ FFMA R4, R22, R5, R4 ; /* 0x0000000516047223 */ /* 0x010fc80000000004 */ /*0520*/ FFMA R21, R21, R6, R4 ; /* 0x0000000615157223 */ /* 0x002fc80000000004 */ /*0530*/ FFMA R21, R8, R7, R21 ; /* 0x0000000708157223 */ /* 0x020fe20000000015 */ /*0540*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0550*/ @!P1 BRA 0x1d0 ; /* 0xfffffc7000009947 */ /* 0x000fea000383ffff */ /*0560*/ IMAD R2, R2, UR5, R13 ; /* 0x0000000502027c24 */ /* 0x000fe2000f8e020d */ /*0570*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fec0003800000 */ /*0580*/ IMAD R5, R3, c[0x0][0x180], R0 ; /* 0x0000600003057a24 */ /* 0x000fe200078e0200 */ /*0590*/ SHF.R.S32.HI R0, RZ, 0x1f, R2 ; /* 0x0000001fff007819 */ /* 0x000fc80000011402 */ /*05a0*/ IADD3 R3, P0, R5, R2, RZ ; /* 0x0000000205037210 */ /* 0x000fc80007f1e0ff */ /*05b0*/ LEA.HI.X.SX32 R0, R5, R0, 0x1, P0 ; /* 0x0000000005007211 */ /* 0x000fe400000f0eff */ /*05c0*/ LEA R2, P0, R3, c[0x0][0x170], 0x2 ; /* 0x00005c0003027a11 */ /* 0x000fc800078010ff */ /*05d0*/ LEA.HI.X R3, R3, c[0x0][0x174], R0, 0x2, P0 ; /* 0x00005d0003037a11 */ /* 0x000fca00000f1400 */ /*05e0*/ STG.E [R2.64], R21 ; /* 0x0000001502007986 */ /* 0x000fe2000c101906 */ /*05f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0600*/ BRA 0x600; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0610*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 _Z4gemmPfS_S_iii .globl _Z4gemmPfS_S_iii .p2align 8 .type _Z4gemmPfS_S_iii,@function _Z4gemmPfS_S_iii: s_load_b64 s[2:3], s[0:1], 0x1c v_mov_b32_e32 v2, 0 v_bfe_u32 v3, v0, 10, 10 v_and_b32_e32 v0, 0x3ff, v0 s_lshl_b32 s4, s14, 4 s_waitcnt lgkmcnt(0) s_lshl_b32 s5, s3, 4 s_cmp_lt_i32 s2, 16 s_cbranch_scc1 .LBB0_5 s_load_b128 s[8:11], s[0:1], 0x0 v_mad_u64_u32 v[6:7], null, v3, s3, v[0:1] v_mad_u64_u32 v[1:2], null, v3, s2, v[0:1] v_dual_mov_b32 v2, 0 :: v_dual_lshlrev_b32 v5, 2, v0 v_lshlrev_b32_e32 v4, 6, v3 s_ashr_i32 s6, s2, 31 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_ashrrev_i32_e32 v7, 31, v6 s_lshr_b32 s6, s6, 28 v_lshlrev_b64 v[8:9], 2, v[1:2] v_add_nc_u32_e32 v1, v4, v5 v_add_nc_u32_e32 v5, 0x400, v5 v_lshlrev_b64 v[10:11], 2, v[6:7] s_add_i32 s6, s2, s6 s_mul_i32 s2, s15, s2 s_ashr_i32 s6, s6, 4 s_mov_b32 s7, 0 s_waitcnt lgkmcnt(0) v_add_co_u32 v6, vcc_lo, s8, v8 v_add_co_ci_u32_e32 v7, vcc_lo, s9, v9, vcc_lo v_add_co_u32 v8, vcc_lo, s10, v10 v_add_co_ci_u32_e32 v9, vcc_lo, s11, v11, vcc_lo v_add_nc_u32_e32 v10, v5, v4 s_set_inst_prefetch_distance 0x1 .p2align 6 .LBB0_2: s_add_i32 s8, s7, s2 s_mul_i32 s9, s7, s5 s_lshl_b32 s8, s8, 4 s_add_i32 s10, s9, s4 s_ashr_i32 s9, s8, 31 s_ashr_i32 s11, s10, 31 s_lshl_b64 s[8:9], s[8:9], 2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(SALU_CYCLE_1) v_add_co_u32 v11, vcc_lo, v6, s8 v_add_co_ci_u32_e32 v12, vcc_lo, s9, v7, vcc_lo s_lshl_b64 s[8:9], s[10:11], 2 v_add_co_u32 v13, vcc_lo, v8, s8 v_add_co_ci_u32_e32 v14, vcc_lo, s9, v9, vcc_lo global_load_b32 v12, v[11:12], off global_load_b32 v13, v[13:14], off v_mov_b32_e32 v11, v5 s_mov_b32 s8, 0 s_waitcnt vmcnt(1) ds_store_b32 v1, v12 s_waitcnt vmcnt(0) ds_store_b32 v10, v13 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB0_3: v_add_nc_u32_e32 v12, s8, v4 s_add_i32 s8, s8, 4 ds_load_b32 v13, v11 ds_load_b32 v12, v12 v_add_nc_u32_e32 v11, 64, v11 s_cmp_eq_u32 s8, 64 s_waitcnt lgkmcnt(0) v_fmac_f32_e32 v2, v12, v13 s_cbranch_scc0 .LBB0_3 s_add_i32 s7, s7, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s7, s6 s_barrier buffer_gl0_inv s_cbranch_scc0 .LBB0_2 .LBB0_5: s_set_inst_prefetch_distance 0x2 s_load_b32 s2, s[0:1], 0x18 v_add_nc_u32_e32 v1, s4, v0 v_lshl_add_u32 v4, s15, 4, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_cmp_gt_i32_e32 vcc_lo, s3, v1 s_waitcnt lgkmcnt(0) v_cmp_gt_i32_e64 s2, s2, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, vcc_lo, s2 s_and_saveexec_b32 s6, s2 s_cbranch_execz .LBB0_7 s_load_b64 s[0:1], s[0:1], 0x10 v_mad_u64_u32 v[4:5], null, v3, s3, v[0:1] s_mul_i32 s2, s5, s15 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s2, s2, s4 s_ashr_i32 s3, s2, 31 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v5, 31, v4 s_lshl_b64 s[2:3], s[2:3], 2 v_lshlrev_b64 v[0:1], 2, v[4:5] s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 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_7: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4gemmPfS_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 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 _Z4gemmPfS_S_iii, .Lfunc_end0-_Z4gemmPfS_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: _Z4gemmPfS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4gemmPfS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0015829e_00000000-6_gemm.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 _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii .type _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii, @function _Z30__device_stub__Z4gemmPfS_S_iiiPfS_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 _Z4gemmPfS_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 _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii, .-_Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii .globl _Z4gemmPfS_S_iii .type _Z4gemmPfS_S_iii, @function _Z4gemmPfS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z4gemmPfS_S_iiiPfS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4gemmPfS_S_iii, .-_Z4gemmPfS_S_iii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4gemmPfS_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 _Z4gemmPfS_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 "gemm.hip" .globl _Z19__device_stub__gemmPfS_S_iii # -- Begin function _Z19__device_stub__gemmPfS_S_iii .p2align 4, 0x90 .type _Z19__device_stub__gemmPfS_S_iii,@function _Z19__device_stub__gemmPfS_S_iii: # @_Z19__device_stub__gemmPfS_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 $_Z4gemmPfS_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 _Z19__device_stub__gemmPfS_S_iii, .Lfunc_end0-_Z19__device_stub__gemmPfS_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 $_Z4gemmPfS_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 _Z4gemmPfS_S_iii,@object # @_Z4gemmPfS_S_iii .section .rodata,"a",@progbits .globl _Z4gemmPfS_S_iii .p2align 3, 0x0 _Z4gemmPfS_S_iii: .quad _Z19__device_stub__gemmPfS_S_iii .size _Z4gemmPfS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4gemmPfS_S_iii" .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 _Z19__device_stub__gemmPfS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4gemmPfS_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.
#include "includes.h" __global__ void add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
code for sm_80 Function : _Z4add3PfS_Pi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_TID.X ; /* 0x0000000000047919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fc800078e0205 */ /*0050*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e0205 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*0080*/ FADD R7, R2, R7 ; /* 0x0000000702077221 */ /* 0x004fca0000000000 */ /*0090*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00b0*/ BRA 0xb0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
.file "tmpxft_000c6dfe_00000000-6_add3.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z4add3PfS_PiPfS_Pi .type _Z27__device_stub__Z4add3PfS_PiPfS_Pi, @function _Z27__device_stub__Z4add3PfS_PiPfS_Pi: .LFB2051: .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 .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 _Z4add3PfS_Pi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z27__device_stub__Z4add3PfS_PiPfS_Pi, .-_Z27__device_stub__Z4add3PfS_PiPfS_Pi .globl _Z4add3PfS_Pi .type _Z4add3PfS_Pi, @function _Z4add3PfS_Pi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z4add3PfS_PiPfS_Pi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4add3PfS_Pi, .-_Z4add3PfS_Pi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4add3PfS_Pi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z4add3PfS_Pi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4add3PfS_Pi .globl _Z4add3PfS_Pi .p2align 8 .type _Z4add3PfS_Pi,@function _Z4add3PfS_Pi: s_load_b128 s[0:3], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_waitcnt lgkmcnt(0) s_clause 0x1 global_load_b32 v1, v0, s[2:3] global_load_b32 v2, v0, s[0:1] s_waitcnt vmcnt(0) v_add_f32_e32 v1, v1, v2 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4add3PfS_Pi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 4 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z4add3PfS_Pi, .Lfunc_end0-_Z4add3PfS_Pi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4add3PfS_Pi .private_segment_fixed_size: 0 .sgpr_count: 4 .sgpr_spill_count: 0 .symbol: _Z4add3PfS_Pi.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 add3(float *val1, float *val2, int *num_elem) { int i = threadIdx.x; val1[i] += val2[i]; }
.text .file "add3.hip" .globl _Z19__device_stub__add3PfS_Pi # -- Begin function _Z19__device_stub__add3PfS_Pi .p2align 4, 0x90 .type _Z19__device_stub__add3PfS_Pi,@function _Z19__device_stub__add3PfS_Pi: # @_Z19__device_stub__add3PfS_Pi .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 $_Z4add3PfS_Pi, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z19__device_stub__add3PfS_Pi, .Lfunc_end0-_Z19__device_stub__add3PfS_Pi .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4add3PfS_Pi, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z4add3PfS_Pi,@object # @_Z4add3PfS_Pi .section .rodata,"a",@progbits .globl _Z4add3PfS_Pi .p2align 3, 0x0 _Z4add3PfS_Pi: .quad _Z19__device_stub__add3PfS_Pi .size _Z4add3PfS_Pi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4add3PfS_Pi" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__add3PfS_Pi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4add3PfS_Pi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z4add3PfS_Pi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_TID.X ; /* 0x0000000000047919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fc800078e0205 */ /*0050*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e0205 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*0080*/ FADD R7, R2, R7 ; /* 0x0000000702077221 */ /* 0x004fca0000000000 */ /*0090*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00b0*/ BRA 0xb0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 _Z4add3PfS_Pi .globl _Z4add3PfS_Pi .p2align 8 .type _Z4add3PfS_Pi,@function _Z4add3PfS_Pi: s_load_b128 s[0:3], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_waitcnt lgkmcnt(0) s_clause 0x1 global_load_b32 v1, v0, s[2:3] global_load_b32 v2, v0, s[0:1] s_waitcnt vmcnt(0) v_add_f32_e32 v1, v1, v2 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4add3PfS_Pi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 4 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z4add3PfS_Pi, .Lfunc_end0-_Z4add3PfS_Pi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z4add3PfS_Pi .private_segment_fixed_size: 0 .sgpr_count: 4 .sgpr_spill_count: 0 .symbol: _Z4add3PfS_Pi.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_000c6dfe_00000000-6_add3.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z4add3PfS_PiPfS_Pi .type _Z27__device_stub__Z4add3PfS_PiPfS_Pi, @function _Z27__device_stub__Z4add3PfS_PiPfS_Pi: .LFB2051: .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 .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 _Z4add3PfS_Pi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z27__device_stub__Z4add3PfS_PiPfS_Pi, .-_Z27__device_stub__Z4add3PfS_PiPfS_Pi .globl _Z4add3PfS_Pi .type _Z4add3PfS_Pi, @function _Z4add3PfS_Pi: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z4add3PfS_PiPfS_Pi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z4add3PfS_Pi, .-_Z4add3PfS_Pi .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z4add3PfS_Pi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z4add3PfS_Pi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "add3.hip" .globl _Z19__device_stub__add3PfS_Pi # -- Begin function _Z19__device_stub__add3PfS_Pi .p2align 4, 0x90 .type _Z19__device_stub__add3PfS_Pi,@function _Z19__device_stub__add3PfS_Pi: # @_Z19__device_stub__add3PfS_Pi .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 $_Z4add3PfS_Pi, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z19__device_stub__add3PfS_Pi, .Lfunc_end0-_Z19__device_stub__add3PfS_Pi .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4add3PfS_Pi, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z4add3PfS_Pi,@object # @_Z4add3PfS_Pi .section .rodata,"a",@progbits .globl _Z4add3PfS_Pi .p2align 3, 0x0 _Z4add3PfS_Pi: .quad _Z19__device_stub__add3PfS_Pi .size _Z4add3PfS_Pi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4add3PfS_Pi" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__add3PfS_Pi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z4add3PfS_Pi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); cudaMalloc((void**)&device_A,buffer_size ); cudaMalloc((void**)&device_B,buffer_size ); cudaMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); cudaFree(device_A); cudaFree(device_B); cudaFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
code for sm_80 Function : _Z8exam_mulPiS_S_i .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2UR UR6, SR_CTAID.X ; /* 0x00000000000679c3 */ /* 0x000e220000002500 */ /*0020*/ S2R R0, SR_TID.Y ; /* 0x0000000000007919 */ /* 0x000e620000002200 */ /*0030*/ ISETP.LT.AND P0, PT, RZ, c[0x0][0x178], PT ; /* 0x00005e00ff007a0c */ /* 0x000fe20003f01270 */ /*0040*/ ULDC UR4, c[0x0][0x178] ; /* 0x00005e0000047ab9 */ /* 0x000fe40000000800 */ /*0050*/ USHF.L.U32 UR4, UR4, 0x6, URZ ; /* 0x0000000604047899 */ /* 0x000fe4000800063f */ /*0060*/ ULDC.64 UR10, c[0x0][0x118] ; /* 0x00004600000a7ab9 */ /* 0x000fe20000000a00 */ /*0070*/ S2UR UR5, SR_CTAID.Y ; /* 0x00000000000579c3 */ /* 0x000eae0000002600 */ /*0080*/ @!P0 MOV R9, RZ ; /* 0x000000ff00098202 */ /* 0x000fe20000000f00 */ /*0090*/ USHF.L.U32 UR6, UR6, 0x6, URZ ; /* 0x0000000606067899 */ /* 0x001fc4000800063f */ /*00a0*/ UIMAD UR5, UR4, UR5, URZ ; /* 0x00000005040572a4 */ /* 0x004fe2000f8e023f */ /*00b0*/ @!P0 BRA 0xb40 ; /* 0x00000a8000008947 */ /* 0x000fea0003800000 */ /*00c0*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x002e220000002100 */ /*00d0*/ SHF.L.U32 R18, R0.reuse, 0x8, RZ ; /* 0x0000000800127819 */ /* 0x040fe200000006ff */ /*00e0*/ HFMA2.MMA R9, -RZ, RZ, 0, 0 ; /* 0x00000000ff097435 */ /* 0x000fe200000001ff */ /*00f0*/ MOV R2, UR5 ; /* 0x0000000500027c02 */ /* 0x000fe20008000f00 */ /*0100*/ ULDC UR7, c[0x0][0x178] ; /* 0x00005e0000077ab9 */ /* 0x000fe40000000800 */ /*0110*/ UIADD3 UR7, UR5, UR7, URZ ; /* 0x0000000705077290 */ /* 0x000fe4000fffe03f */ /*0120*/ UMOV UR8, UR6 ; /* 0x0000000600087c82 */ /* 0x000fe20008000000 */ /*0130*/ IMAD R17, R0, c[0x0][0x178], R3 ; /* 0x00005e0000117a24 */ /* 0x001fe200078e0203 */ /*0140*/ LEA R16, R3, R18, 0x2 ; /* 0x0000001203107211 */ /* 0x000fc800078e10ff */ /*0150*/ IADD3 R4, R17, R2, RZ ; /* 0x0000000211047210 */ /* 0x000fe40007ffe0ff */ /*0160*/ MOV R11, 0x4 ; /* 0x00000004000b7802 */ /* 0x000fc40000000f00 */ /*0170*/ IADD3 R10, R17, UR8, RZ ; /* 0x00000008110a7c10 */ /* 0x000fc6000fffe0ff */ /*0180*/ IMAD.WIDE R4, R4, R11, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fc800078e020b */ /*0190*/ IMAD.WIDE R10, R10, R11, c[0x0][0x168] ; /* 0x00005a000a0a7625 */ /* 0x000fe200078e020b */ /*01a0*/ LDG.E R19, [R4.64] ; /* 0x0000000a04137981 */ /* 0x000eaa000c1e1900 */ /*01b0*/ LDG.E R11, [R10.64] ; /* 0x0000000a0a0b7981 */ /* 0x000ee2000c1e1900 */ /*01c0*/ IADD3 R2, R2, 0x40, RZ ; /* 0x0000004002027810 */ /* 0x000fe20007ffe0ff */ /*01d0*/ UIADD3 UR8, UR4, UR8, URZ ; /* 0x0000000804087290 */ /* 0x000fc6000fffe03f */ /*01e0*/ ISETP.GE.AND P0, PT, R2, UR7, PT ; /* 0x0000000702007c0c */ /* 0x000fe2000bf06270 */ /*01f0*/ STS [R16], R19 ; /* 0x0000001310007388 */ /* 0x004fe80000000800 */ /*0200*/ STS [R16+0x4000], R11 ; /* 0x0040000b10007388 */ /* 0x008fe80000000800 */ /*0210*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0220*/ LDS R8, [R3.X4+0x4000] ; /* 0x0040000003087984 */ /* 0x000fe80000004800 */ /*0230*/ LDS.128 R12, [R18] ; /* 0x00000000120c7984 */ /* 0x000e280000000c00 */ /*0240*/ LDS R21, [R3.X4+0x4100] ; /* 0x0041000003157984 */ /* 0x000e680000004800 */ /*0250*/ LDS R23, [R3.X4+0x4200] ; /* 0x0042000003177984 */ /* 0x000ea80000004800 */ /*0260*/ LDS R24, [R3.X4+0x4300] ; /* 0x0043000003187984 */ /* 0x000ee80000004800 */ /*0270*/ LDS R25, [R3.X4+0x4400] ; /* 0x0044000003197984 */ /* 0x000fe80000004800 */ /*0280*/ LDS.128 R4, [R18+0x10] ; /* 0x0000100012047984 */ /* 0x000f280000000c00 */ /*0290*/ LDS R22, [R3.X4+0x4500] ; /* 0x0045000003167984 */ /* 0x000f680000004800 */ /*02a0*/ LDS R19, [R3.X4+0x4600] ; /* 0x0046000003137984 */ /* 0x000f680000004800 */ /*02b0*/ LDS R20, [R3.X4+0x4700] ; /* 0x0047000003147984 */ /* 0x000f680000004800 */ /*02c0*/ LDS R26, [R3.X4+0x4900] ; /* 0x00490000031a7984 */ /* 0x000fe20000004800 */ /*02d0*/ IMAD R8, R8, R12, R9 ; /* 0x0000000c08087224 */ /* 0x001fc800078e0209 */ /*02e0*/ IMAD R13, R21, R13, R8 ; /* 0x0000000d150d7224 */ /* 0x002fe400078e0208 */ /*02f0*/ LDS R21, [R3.X4+0x4800] ; /* 0x0048000003157984 */ /* 0x000fe40000004800 */ /*0300*/ IMAD R13, R23, R14, R13 ; /* 0x0000000e170d7224 */ /* 0x004fe400078e020d */ /*0310*/ LDS.128 R8, [R18+0x20] ; /* 0x0000200012087984 */ /* 0x000e240000000c00 */ /*0320*/ IMAD R13, R24, R15, R13 ; /* 0x0000000f180d7224 */ /* 0x008fe400078e020d */ /*0330*/ LDS R23, [R3.X4+0x4a00] ; /* 0x004a000003177984 */ /* 0x000e680000004800 */ /*0340*/ LDS R24, [R3.X4+0x4b00] ; /* 0x004b000003187984 */ /* 0x000ea20000004800 */ /*0350*/ IMAD R4, R25, R4, R13 ; /* 0x0000000419047224 */ /* 0x010fc600078e020d */ /*0360*/ LDS R25, [R3.X4+0x4c00] ; /* 0x004c000003197984 */ /* 0x000fe20000004800 */ /*0370*/ IMAD R4, R22, R5, R4 ; /* 0x0000000516047224 */ /* 0x020fc600078e0204 */ /*0380*/ LDS.128 R12, [R18+0x30] ; /* 0x00003000120c7984 */ /* 0x000ee20000000c00 */ /*0390*/ IMAD R4, R19, R6, R4 ; /* 0x0000000613047224 */ /* 0x000fc600078e0204 */ /*03a0*/ LDS R22, [R3.X4+0x4d00] ; /* 0x004d000003167984 */ /* 0x000f220000004800 */ /*03b0*/ IMAD R4, R20, R7, R4 ; /* 0x0000000714047224 */ /* 0x000fc600078e0204 */ /*03c0*/ LDS R19, [R3.X4+0x4e00] ; /* 0x004e000003137984 */ /* 0x000f680000004800 */ /*03d0*/ LDS R20, [R3.X4+0x4f00] ; /* 0x004f000003147984 */ /* 0x000f620000004800 */ /*03e0*/ IMAD R4, R21, R8, R4 ; /* 0x0000000815047224 */ /* 0x001fc600078e0204 */ /*03f0*/ LDS R21, [R3.X4+0x5000] ; /* 0x0050000003157984 */ /* 0x000fe20000004800 */ /*0400*/ IMAD R9, R26, R9, R4 ; /* 0x000000091a097224 */ /* 0x000fc600078e0204 */ /*0410*/ LDS.128 R4, [R18+0x40] ; /* 0x0000400012047984 */ /* 0x000e220000000c00 */ /*0420*/ IMAD R9, R23, R10, R9 ; /* 0x0000000a17097224 */ /* 0x002fc600078e0209 */ /*0430*/ LDS R26, [R3.X4+0x5100] ; /* 0x00510000031a7984 */ /* 0x000e620000004800 */ /*0440*/ IMAD R9, R24, R11, R9 ; /* 0x0000000b18097224 */ /* 0x004fc600078e0209 */ /*0450*/ LDS R23, [R3.X4+0x5200] ; /* 0x0052000003177984 */ /* 0x000ea20000004800 */ /*0460*/ IMAD R9, R25, R12, R9 ; /* 0x0000000c19097224 */ /* 0x008fc600078e0209 */ /*0470*/ LDS R24, [R3.X4+0x5300] ; /* 0x0053000003187984 */ /* 0x000ee20000004800 */ /*0480*/ IMAD R13, R22, R13, R9 ; /* 0x0000000d160d7224 */ /* 0x010fc600078e0209 */ /*0490*/ LDS R25, [R3.X4+0x5400] ; /* 0x0054000003197984 */ /* 0x000fe20000004800 */ /*04a0*/ IMAD R13, R19, R14, R13 ; /* 0x0000000e130d7224 */ /* 0x020fc600078e020d */ /*04b0*/ LDS.128 R8, [R18+0x50] ; /* 0x0000500012087984 */ /* 0x000f220000000c00 */ /*04c0*/ IMAD R13, R20, R15, R13 ; /* 0x0000000f140d7224 */ /* 0x000fc600078e020d */ /*04d0*/ LDS R22, [R3.X4+0x5500] ; /* 0x0055000003167984 */ /* 0x000f680000004800 */ /*04e0*/ LDS R19, [R3.X4+0x5600] ; /* 0x0056000003137984 */ /* 0x000f680000004800 */ /*04f0*/ LDS R20, [R3.X4+0x5700] ; /* 0x0057000003147984 */ /* 0x000f620000004800 */ /*0500*/ IMAD R4, R21, R4, R13 ; /* 0x0000000415047224 */ /* 0x001fc600078e020d */ /*0510*/ LDS R21, [R3.X4+0x5800] ; /* 0x0058000003157984 */ /* 0x000fe20000004800 */ /*0520*/ IMAD R4, R26, R5, R4 ; /* 0x000000051a047224 */ /* 0x002fc600078e0204 */ /*0530*/ LDS.128 R12, [R18+0x60] ; /* 0x00006000120c7984 */ /* 0x000e220000000c00 */ /*0540*/ IMAD R4, R23, R6, R4 ; /* 0x0000000617047224 */ /* 0x004fc600078e0204 */ /*0550*/ LDS R26, [R3.X4+0x5900] ; /* 0x00590000031a7984 */ /* 0x000e620000004800 */ /*0560*/ IMAD R4, R24, R7, R4 ; /* 0x0000000718047224 */ /* 0x008fc600078e0204 */ /*0570*/ LDS R23, [R3.X4+0x5a00] ; /* 0x005a000003177984 */ /* 0x000ea80000004800 */ /*0580*/ LDS R24, [R3.X4+0x5b00] ; /* 0x005b000003187984 */ /* 0x000ee20000004800 */ /*0590*/ IMAD R4, R25, R8, R4 ; /* 0x0000000819047224 */ /* 0x010fc600078e0204 */ /*05a0*/ LDS R25, [R3.X4+0x5c00] ; /* 0x005c000003197984 */ /* 0x000fe20000004800 */ /*05b0*/ IMAD R9, R22, R9, R4 ; /* 0x0000000916097224 */ /* 0x020fc600078e0204 */ /*05c0*/ LDS.128 R4, [R18+0x70] ; /* 0x0000700012047984 */ /* 0x000f220000000c00 */ /*05d0*/ IMAD R9, R19, R10, R9 ; /* 0x0000000a13097224 */ /* 0x000fc600078e0209 */ /*05e0*/ LDS R22, [R3.X4+0x5d00] ; /* 0x005d000003167984 */ /* 0x000f620000004800 */ /*05f0*/ IMAD R9, R20, R11, R9 ; /* 0x0000000b14097224 */ /* 0x000fc600078e0209 */ /*0600*/ LDS R19, [R3.X4+0x5e00] ; /* 0x005e000003137984 */ /* 0x000f680000004800 */ /*0610*/ LDS R20, [R3.X4+0x5f00] ; /* 0x005f000003147984 */ /* 0x000f620000004800 */ /*0620*/ IMAD R9, R21, R12, R9 ; /* 0x0000000c15097224 */ /* 0x001fc600078e0209 */ /*0630*/ LDS R21, [R3.X4+0x6000] ; /* 0x0060000003157984 */ /* 0x000fe20000004800 */ /*0640*/ IMAD R13, R26, R13, R9 ; /* 0x0000000d1a0d7224 */ /* 0x002fc600078e0209 */ /*0650*/ LDS.128 R8, [R18+0x80] ; /* 0x0000800012087984 */ /* 0x000e220000000c00 */ /*0660*/ IMAD R13, R23, R14, R13 ; /* 0x0000000e170d7224 */ /* 0x004fc600078e020d */ /*0670*/ LDS R26, [R3.X4+0x6100] ; /* 0x00610000031a7984 */ /* 0x000e620000004800 */ /*0680*/ IMAD R13, R24, R15, R13 ; /* 0x0000000f180d7224 */ /* 0x008fc600078e020d */ /*0690*/ LDS R23, [R3.X4+0x6200] ; /* 0x0062000003177984 */ /* 0x000ea80000004800 */ /*06a0*/ LDS R24, [R3.X4+0x6300] ; /* 0x0063000003187984 */ /* 0x000ee20000004800 */ /*06b0*/ IMAD R4, R25, R4, R13 ; /* 0x0000000419047224 */ /* 0x010fc600078e020d */ /*06c0*/ LDS R25, [R3.X4+0x6400] ; /* 0x0064000003197984 */ /* 0x000fe20000004800 */ /*06d0*/ IMAD R4, R22, R5, R4 ; /* 0x0000000516047224 */ /* 0x020fc600078e0204 */ /*06e0*/ LDS.128 R12, [R18+0x90] ; /* 0x00009000120c7984 */ /* 0x000f220000000c00 */ /*06f0*/ IMAD R4, R19, R6, R4 ; /* 0x0000000613047224 */ /* 0x000fc600078e0204 */ /*0700*/ LDS R22, [R3.X4+0x6500] ; /* 0x0065000003167984 */ /* 0x000f620000004800 */ /*0710*/ IMAD R4, R20, R7, R4 ; /* 0x0000000714047224 */ /* 0x000fc600078e0204 */ /*0720*/ LDS R19, [R3.X4+0x6600] ; /* 0x0066000003137984 */ /* 0x000f680000004800 */ /*0730*/ LDS R20, [R3.X4+0x6700] ; /* 0x0067000003147984 */ /* 0x000f620000004800 */ /*0740*/ IMAD R4, R21, R8, R4 ; /* 0x0000000815047224 */ /* 0x001fc600078e0204 */ /*0750*/ LDS R21, [R3.X4+0x6800] ; /* 0x0068000003157984 */ /* 0x000fe20000004800 */ /*0760*/ IMAD R9, R26, R9, R4 ; /* 0x000000091a097224 */ /* 0x002fc600078e0204 */ /*0770*/ LDS.128 R4, [R18+0xa0] ; /* 0x0000a00012047984 */ /* 0x000e220000000c00 */ /*0780*/ IMAD R9, R23, R10, R9 ; /* 0x0000000a17097224 */ /* 0x004fc600078e0209 */ /*0790*/ LDS R26, [R3.X4+0x6900] ; /* 0x00690000031a7984 */ /* 0x000e620000004800 */ /*07a0*/ IMAD R9, R24, R11, R9 ; /* 0x0000000b18097224 */ /* 0x008fc600078e0209 */ /*07b0*/ LDS R23, [R3.X4+0x6a00] ; /* 0x006a000003177984 */ /* 0x000ea80000004800 */ /*07c0*/ LDS R24, [R3.X4+0x6b00] ; /* 0x006b000003187984 */ /* 0x000ee20000004800 */ /*07d0*/ IMAD R9, R25, R12, R9 ; /* 0x0000000c19097224 */ /* 0x010fc600078e0209 */ /*07e0*/ LDS R25, [R3.X4+0x6c00] ; /* 0x006c000003197984 */ /* 0x000fe20000004800 */ /*07f0*/ IMAD R13, R22, R13, R9 ; /* 0x0000000d160d7224 */ /* 0x020fc600078e0209 */ /*0800*/ LDS.128 R8, [R18+0xb0] ; /* 0x0000b00012087984 */ /* 0x000f220000000c00 */ /*0810*/ IMAD R13, R19, R14, R13 ; /* 0x0000000e130d7224 */ /* 0x000fc600078e020d */ /*0820*/ LDS R22, [R3.X4+0x6d00] ; /* 0x006d000003167984 */ /* 0x000f620000004800 */ /*0830*/ IMAD R13, R20, R15, R13 ; /* 0x0000000f140d7224 */ /* 0x000fc600078e020d */ /*0840*/ LDS R19, [R3.X4+0x6e00] ; /* 0x006e000003137984 */ /* 0x000f680000004800 */ /*0850*/ LDS R20, [R3.X4+0x6f00] ; /* 0x006f000003147984 */ /* 0x000f620000004800 */ /*0860*/ IMAD R4, R21, R4, R13 ; /* 0x0000000415047224 */ /* 0x001fc600078e020d */ /*0870*/ LDS R21, [R3.X4+0x7000] ; /* 0x0070000003157984 */ /* 0x000fe20000004800 */ /*0880*/ IMAD R4, R26, R5, R4 ; /* 0x000000051a047224 */ /* 0x002fc600078e0204 */ /*0890*/ LDS.128 R12, [R18+0xc0] ; /* 0x0000c000120c7984 */ /* 0x000e220000000c00 */ /*08a0*/ IMAD R4, R23, R6, R4 ; /* 0x0000000617047224 */ /* 0x004fc600078e0204 */ /*08b0*/ LDS R26, [R3.X4+0x7100] ; /* 0x00710000031a7984 */ /* 0x000e620000004800 */ /*08c0*/ IMAD R4, R24, R7, R4 ; /* 0x0000000718047224 */ /* 0x008fc600078e0204 */ /*08d0*/ LDS R23, [R3.X4+0x7200] ; /* 0x0072000003177984 */ /* 0x000ea80000004800 */ /*08e0*/ LDS R24, [R3.X4+0x7300] ; /* 0x0073000003187984 */ /* 0x000ee20000004800 */ /*08f0*/ IMAD R4, R25, R8, R4 ; /* 0x0000000819047224 */ /* 0x010fc600078e0204 */ /*0900*/ LDS R25, [R3.X4+0x7400] ; /* 0x0074000003197984 */ /* 0x000fe20000004800 */ /*0910*/ IMAD R9, R22, R9, R4 ; /* 0x0000000916097224 */ /* 0x020fc600078e0204 */ /*0920*/ LDS.128 R4, [R18+0xd0] ; /* 0x0000d00012047984 */ /* 0x000f220000000c00 */ /*0930*/ IMAD R9, R19, R10, R9 ; /* 0x0000000a13097224 */ /* 0x000fc600078e0209 */ /*0940*/ LDS R22, [R3.X4+0x7500] ; /* 0x0075000003167984 */ /* 0x000f620000004800 */ /*0950*/ IMAD R9, R20, R11, R9 ; /* 0x0000000b14097224 */ /* 0x000fc600078e0209 */ /*0960*/ LDS R19, [R3.X4+0x7600] ; /* 0x0076000003137984 */ /* 0x000f680000004800 */ /*0970*/ LDS R20, [R3.X4+0x7700] ; /* 0x0077000003147984 */ /* 0x000f620000004800 */ /*0980*/ IMAD R9, R21, R12, R9 ; /* 0x0000000c15097224 */ /* 0x001fc600078e0209 */ /*0990*/ LDS R21, [R3.X4+0x7800] ; /* 0x0078000003157984 */ /* 0x000fe20000004800 */ /*09a0*/ IMAD R13, R26, R13, R9 ; /* 0x0000000d1a0d7224 */ /* 0x002fc600078e0209 */ /*09b0*/ LDS.128 R8, [R18+0xe0] ; /* 0x0000e00012087984 */ /* 0x000e220000000c00 */ /*09c0*/ IMAD R13, R23, R14, R13 ; /* 0x0000000e170d7224 */ /* 0x004fc600078e020d */ /*09d0*/ LDS R26, [R3.X4+0x7900] ; /* 0x00790000031a7984 */ /* 0x000e620000004800 */ /*09e0*/ IMAD R13, R24, R15, R13 ; /* 0x0000000f180d7224 */ /* 0x008fc600078e020d */ /*09f0*/ LDS R23, [R3.X4+0x7a00] ; /* 0x007a000003177984 */ /* 0x000ea20000004800 */ /*0a00*/ IMAD R13, R25, R4, R13 ; /* 0x00000004190d7224 */ /* 0x010fc600078e020d */ /*0a10*/ LDS R4, [R3.X4+0x7b00] ; /* 0x007b000003047984 */ /* 0x000ee20000004800 */ /*0a20*/ IMAD R24, R22, R5, R13 ; /* 0x0000000516187224 */ /* 0x020fc600078e020d */ /*0a30*/ LDS R5, [R3.X4+0x7c00] ; /* 0x007c000003057984 */ /* 0x000fe20000004800 */ /*0a40*/ IMAD R24, R19, R6, R24 ; /* 0x0000000613187224 */ /* 0x000fc600078e0218 */ /*0a50*/ LDS.128 R12, [R18+0xf0] ; /* 0x0000f000120c7984 */ /* 0x000f220000000c00 */ /*0a60*/ IMAD R7, R20, R7, R24 ; /* 0x0000000714077224 */ /* 0x000fc600078e0218 */ /*0a70*/ LDS R22, [R3.X4+0x7d00] ; /* 0x007d000003167984 */ /* 0x000f680000004800 */ /*0a80*/ LDS R19, [R3.X4+0x7e00] ; /* 0x007e000003137984 */ /* 0x000f680000004800 */ /*0a90*/ LDS R6, [R3.X4+0x7f00] ; /* 0x007f000003067984 */ /* 0x000f620000004800 */ /*0aa0*/ IMAD R7, R21, R8, R7 ; /* 0x0000000815077224 */ /* 0x001fc800078e0207 */ /*0ab0*/ IMAD R7, R26, R9, R7 ; /* 0x000000091a077224 */ /* 0x002fc800078e0207 */ /*0ac0*/ IMAD R7, R23, R10, R7 ; /* 0x0000000a17077224 */ /* 0x004fc800078e0207 */ /*0ad0*/ IMAD R4, R4, R11, R7 ; /* 0x0000000b04047224 */ /* 0x008fc800078e0207 */ /*0ae0*/ IMAD R4, R5, R12, R4 ; /* 0x0000000c05047224 */ /* 0x010fc800078e0204 */ /*0af0*/ IMAD R4, R22, R13, R4 ; /* 0x0000000d16047224 */ /* 0x020fc800078e0204 */ /*0b00*/ IMAD R4, R19, R14, R4 ; /* 0x0000000e13047224 */ /* 0x000fc800078e0204 */ /*0b10*/ IMAD R9, R6, R15, R4 ; /* 0x0000000f06097224 */ /* 0x000fe200078e0204 */ /*0b20*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0b30*/ @!P0 BRA 0x150 ; /* 0xfffff61000008947 */ /* 0x000fea000383ffff */ /*0b40*/ IADD3 R3, R0, UR6, RZ ; /* 0x0000000600037c10 */ /* 0x002fe2000fffe0ff */ /*0b50*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fc800000001ff */ /*0b60*/ IMAD R3, R0, c[0x0][0x178], R3 ; /* 0x00005e0000037a24 */ /* 0x000fca00078e0203 */ /*0b70*/ IADD3 R3, R3, UR5, RZ ; /* 0x0000000503037c10 */ /* 0x000fca000fffe0ff */ /*0b80*/ IMAD.WIDE R2, R3, R2, c[0x0][0x170] ; /* 0x00005c0003027625 */ /* 0x000fca00078e0202 */ /*0b90*/ STG.E [R2.64], R9 ; /* 0x0000000902007986 */ /* 0x000fe2000c10190a */ /*0ba0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0bb0*/ BRA 0xbb0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0bc0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0bd0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0be0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0bf0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c00*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c10*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c20*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0c70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z10shared_mulPiS_S_i .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z8cuda_mulPiS_S_i .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ ISETP.LT.AND P0, PT, RZ, c[0x0][0x178], PT ; /* 0x00005e00ff007a0c */ /* 0x000fe20003f01270 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd80000000a00 */ /*0030*/ @!P0 MOV R16, RZ ; /* 0x000000ff00108202 */ /* 0x000fe20000000f00 */ /*0040*/ @!P0 BRA 0x440 ; /* 0x000003f000008947 */ /* 0x000fea0003800000 */ /*0050*/ MOV R2, c[0x0][0x178] ; /* 0x00005e0000027a02 */ /* 0x000fe20000000f00 */ /*0060*/ S2R R17, SR_TID.Y ; /* 0x0000000000117919 */ /* 0x000e220000002200 */ /*0070*/ MOV R16, RZ ; /* 0x000000ff00107202 */ /* 0x000fe40000000f00 */ /*0080*/ IADD3 R0, R2.reuse, -0x1, RZ ; /* 0xffffffff02007810 */ /* 0x040fe20007ffe0ff */ /*0090*/ S2R R19, SR_TID.X ; /* 0x0000000000137919 */ /* 0x000e620000002100 */ /*00a0*/ LOP3.LUT R15, R2, 0x3, RZ, 0xc0, !PT ; /* 0x00000003020f7812 */ /* 0x000fe400078ec0ff */ /*00b0*/ ISETP.GE.U32.AND P1, PT, R0, 0x3, PT ; /* 0x000000030000780c */ /* 0x000fe20003f26070 */ /*00c0*/ HFMA2.MMA R0, -RZ, RZ, 0, 0 ; /* 0x00000000ff007435 */ /* 0x000fe200000001ff */ /*00d0*/ ISETP.NE.AND P0, PT, R15, RZ, PT ; /* 0x000000ff0f00720c */ /* 0x000fd60003f05270 */ /*00e0*/ @!P1 BRA 0x360 ; /* 0x0000027000009947 */ /* 0x000fea0003800000 */ /*00f0*/ IMAD R14, R19, R2.reuse, 0x3 ; /* 0x00000003130e7424 */ /* 0x082fe200078e0202 */ /*0100*/ IADD3 R21, R15, -c[0x0][0x178], RZ ; /* 0x80005e000f157a10 */ /* 0x000fe20007ffe0ff */ /*0110*/ IMAD R18, R17, R2.reuse, 0x3 ; /* 0x0000000311127424 */ /* 0x081fe200078e0202 */ /*0120*/ MOV R0, RZ ; /* 0x000000ff00007202 */ /* 0x000fe20000000f00 */ /*0130*/ IMAD R23, R19, R2.reuse, 0x1 ; /* 0x0000000113177424 */ /* 0x080fe400078e0202 */ /*0140*/ IMAD R25, R17, R2, 0x1 ; /* 0x0000000111197424 */ /* 0x000fc600078e0202 */ /*0150*/ HFMA2.MMA R13, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff0d7435 */ /* 0x000fe200000001ff */ /*0160*/ IMAD R26, R19, c[0x0][0x178], R0.reuse ; /* 0x00005e00131a7a24 */ /* 0x100fe200078e0200 */ /*0170*/ IADD3 R8, R25, R0.reuse, RZ ; /* 0x0000000019087210 */ /* 0x080fe20007ffe0ff */ /*0180*/ IMAD R10, R17, c[0x0][0x178], R0 ; /* 0x00005e00110a7a24 */ /* 0x000fe200078e0200 */ /*0190*/ IADD3 R4, R23, R0, RZ ; /* 0x0000000017047210 */ /* 0x000fc40007ffe0ff */ /*01a0*/ IADD3 R6, R18, -0x1, RZ ; /* 0xffffffff12067810 */ /* 0x000fc80007ffe0ff */ /*01b0*/ IMAD.WIDE.U32 R26, R26, R13, c[0x0][0x168] ; /* 0x00005a001a1a7625 */ /* 0x000fe200078e000d */ /*01c0*/ IADD3 R2, R14, -0x1, RZ ; /* 0xffffffff0e027810 */ /* 0x000fc60007ffe0ff */ /*01d0*/ IMAD.WIDE.U32 R10, R10, R13.reuse, c[0x0][0x160] ; /* 0x000058000a0a7625 */ /* 0x080fe400078e000d */ /*01e0*/ LDG.E R26, [R26.64] ; /* 0x000000041a1a7981 */ /* 0x000ea4000c1e1900 */ /*01f0*/ IMAD.WIDE.U32 R8, R8, R13.reuse, c[0x0][0x160] ; /* 0x0000580008087625 */ /* 0x080fe400078e000d */ /*0200*/ LDG.E R29, [R10.64] ; /* 0x000000040a1d7981 */ /* 0x0000a4000c1e1900 */ /*0210*/ IMAD.WIDE.U32 R4, R4, R13.reuse, c[0x0][0x168] ; /* 0x00005a0004047625 */ /* 0x080fe400078e000d */ /*0220*/ LDG.E R8, [R8.64] ; /* 0x0000000408087981 */ /* 0x0002e4000c1e1900 */ /*0230*/ IMAD.WIDE.U32 R6, R6, R13, c[0x0][0x160] ; /* 0x0000580006067625 */ /* 0x000fc400078e000d */ /*0240*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ee4000c1e1900 */ /*0250*/ IMAD.WIDE.U32 R2, R2, R13.reuse, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x080fe400078e000d */ /*0260*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000f24000c1e1900 */ /*0270*/ IMAD.WIDE.U32 R10, R18, R13.reuse, c[0x0][0x160] ; /* 0x00005800120a7625 */ /* 0x081fe400078e000d */ /*0280*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000f24000c1e1900 */ /*0290*/ IMAD.WIDE.U32 R12, R14, R13, c[0x0][0x168] ; /* 0x00005a000e0c7625 */ /* 0x000fc400078e000d */ /*02a0*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000f68000c1e1900 */ /*02b0*/ LDG.E R13, [R12.64] ; /* 0x000000040c0d7981 */ /* 0x000f62000c1e1900 */ /*02c0*/ IADD3 R0, R0, 0x4, RZ ; /* 0x0000000400007810 */ /* 0x000fc80007ffe0ff */ /*02d0*/ IADD3 R9, R21, R0, RZ ; /* 0x0000000015097210 */ /* 0x002fc80007ffe0ff */ /*02e0*/ ISETP.NE.AND P1, PT, R9, RZ, PT ; /* 0x000000ff0900720c */ /* 0x000fe40003f25270 */ /*02f0*/ IADD3 R18, R18, 0x4, RZ ; /* 0x0000000412127810 */ /* 0x000fe40007ffe0ff */ /*0300*/ IADD3 R14, R14, 0x4, RZ ; /* 0x000000040e0e7810 */ /* 0x000fe20007ffe0ff */ /*0310*/ IMAD R29, R26, R29, R16 ; /* 0x0000001d1a1d7224 */ /* 0x004fc800078e0210 */ /*0320*/ IMAD R8, R4, R8, R29 ; /* 0x0000000804087224 */ /* 0x008fc800078e021d */ /*0330*/ IMAD R6, R3, R6, R8 ; /* 0x0000000603067224 */ /* 0x010fc800078e0208 */ /*0340*/ IMAD R16, R13, R10, R6 ; /* 0x0000000a0d107224 */ /* 0x020fe200078e0206 */ /*0350*/ @P1 BRA 0x150 ; /* 0xfffffdf000001947 */ /* 0x000fea000383ffff */ /*0360*/ @!P0 BRA 0x440 ; /* 0x000000d000008947 */ /* 0x000fea0003800000 */ /*0370*/ IMAD R19, R19, c[0x0][0x178], R0.reuse ; /* 0x00005e0013137a24 */ /* 0x102fe400078e0200 */ /*0380*/ IMAD R0, R17, c[0x0][0x178], R0 ; /* 0x00005e0011007a24 */ /* 0x001fe400078e0200 */ /*0390*/ MOV R4, 0x4 ; /* 0x0000000400047802 */ /* 0x000fca0000000f00 */ /*03a0*/ IMAD.WIDE.U32 R2, R0, R4, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x000fc800078e0004 */ /*03b0*/ IMAD.WIDE.U32 R4, R19, R4, c[0x0][0x168] ; /* 0x00005a0013047625 */ /* 0x000fe400078e0004 */ /*03c0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*03d0*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*03e0*/ IADD3 R15, R15, -0x1, RZ ; /* 0xffffffff0f0f7810 */ /* 0x000fe40007ffe0ff */ /*03f0*/ IADD3 R0, R0, 0x1, RZ ; /* 0x0000000100007810 */ /* 0x000fe40007ffe0ff */ /*0400*/ ISETP.NE.AND P0, PT, R15, RZ, PT ; /* 0x000000ff0f00720c */ /* 0x000fc40003f05270 */ /*0410*/ IADD3 R19, R19, 0x1, RZ ; /* 0x0000000113137810 */ /* 0x000fe20007ffe0ff */ /*0420*/ IMAD R16, R5, R2, R16 ; /* 0x0000000205107224 */ /* 0x004fd400078e0210 */ /*0430*/ @P0 BRA 0x390 ; /* 0xffffff5000000947 */ /* 0x000fea000383ffff */ /*0440*/ S2R R2, SR_TID.X ; /* 0x0000000000027919 */ /* 0x000ea20000002100 */ /*0450*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fc600000001ff */ /*0460*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000ea40000002200 */ /*0470*/ IMAD R2, R2, c[0x0][0x178], R3 ; /* 0x00005e0002027a24 */ /* 0x004fca00078e0203 */ /*0480*/ IMAD.WIDE.U32 R2, R2, R5, c[0x0][0x170] ; /* 0x00005c0002027625 */ /* 0x000fca00078e0005 */ /*0490*/ STG.E [R2.64], R16 ; /* 0x0000001002007986 */ /* 0x000fe2000c101904 */ /*04a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*04b0*/ BRA 0x4b0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*04c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*04d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*04e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*04f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0500*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0510*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0520*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0530*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0540*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0550*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0560*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0570*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); cudaMalloc((void**)&device_A,buffer_size ); cudaMalloc((void**)&device_B,buffer_size ); cudaMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); cudaFree(device_A); cudaFree(device_B); cudaFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
.file "tmpxft_001149e5_00000000-6_4_sharedMul.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Failed to call clock_gettime\n" .LC1: .string "elapsed time : % lld micros\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "wrong flag | 0 : start, 1 : end\n" .text .globl _Z9stopwatchi .type _Z9stopwatchi, @function _Z9stopwatchi: .LFB2058: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 testl %edi, %edi je .L9 cmpl $1, %edi je .L10 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L3: addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .cfi_restore_state leaq _ZZ9stopwatchiE7startTS(%rip), %rsi movl $1, %edi call clock_gettime@PLT cmpl $-1, %eax jne .L3 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L3 .L10: leaq _ZZ9stopwatchiE5endTS(%rip), %rsi call clock_gettime@PLT cmpl $-1, %eax je .L11 .L7: movq _ZZ9stopwatchiE5endTS(%rip), %rcx subq _ZZ9stopwatchiE7startTS(%rip), %rcx imulq $1000000000, %rcx, %rcx addq 8+_ZZ9stopwatchiE5endTS(%rip), %rcx subq 8+_ZZ9stopwatchiE7startTS(%rip), %rcx movabsq $2361183241434822607, %rdx movq %rcx, %rax imulq %rdx sarq $7, %rdx sarq $63, %rcx subq %rcx, %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L3 .L11: leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L7 .cfi_endproc .LFE2058: .size _Z9stopwatchi, .-_Z9stopwatchi .globl _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i .type _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i, @function _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i: .LFB2083: .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 .L16 .L12: movq 136(%rsp), %rax subq %fs:40, %rax jne .L17 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L16: .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 _Z8cuda_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L12 .L17: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i, .-_Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i .globl _Z8cuda_mulPiS_S_i .type _Z8cuda_mulPiS_S_i, @function _Z8cuda_mulPiS_S_i: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z8cuda_mulPiS_S_i, .-_Z8cuda_mulPiS_S_i .globl _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i .type _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i, @function _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i: .LFB2085: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L24 .L20: movq 136(%rsp), %rax subq %fs:40, %rax jne .L25 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L24: .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 _Z10shared_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L20 .L25: call __stack_chk_fail@PLT .cfi_endproc .LFE2085: .size _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i, .-_Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i .globl _Z10shared_mulPiS_S_i .type _Z10shared_mulPiS_S_i, @function _Z10shared_mulPiS_S_i: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _Z10shared_mulPiS_S_i, .-_Z10shared_mulPiS_S_i .globl _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i .type _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i, @function _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L32 .L28: movq 136(%rsp), %rax subq %fs:40, %rax jne .L33 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L32: .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 _Z8exam_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L28 .L33: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i, .-_Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i .globl _Z8exam_mulPiS_S_i .type _Z8exam_mulPiS_S_i, @function _Z8exam_mulPiS_S_i: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z8exam_mulPiS_S_i, .-_Z8exam_mulPiS_S_i .section .rodata.str1.8 .align 8 .LC3: .string "Multiply matrix (%dX%d ) * (%dX%d)\n" .section .rodata.str1.1 .LC4: .string "cuda_mul\n" .LC5: .string "shared_mul\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r12 .cfi_def_cfa_offset 16 .cfi_offset 12, -16 pushq %rbp .cfi_def_cfa_offset 24 .cfi_offset 6, -24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset 3, -32 subq $64, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $4194304, %edi call malloc@PLT movq %rax, %r12 movl $4194304, %edi call malloc@PLT movq %rax, %rbp movl $4194304, %edi call malloc@PLT movq %rax, %rbx movl $0, %eax .L37: movl %eax, (%r12,%rax,4) movl %eax, 0(%rbp,%rax,4) movl $0, (%rbx,%rax,4) addq $1, %rax cmpq $1048576, %rax jne .L37 movl $1024, %r9d movl $1024, %r8d movl $1024, %ecx movl $1024, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1024, 32(%rsp) movl $1024, 36(%rsp) movl $1, 40(%rsp) leaq 8(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call _Z9stopwatchi movl $1, %ecx movl $4194304, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4194304, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L45 .L38: movl $2, %ecx movl $4194304, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl $1, %edi call _Z9stopwatchi movl $0, %eax .L39: movl %eax, (%r12,%rax,4) movl %eax, 0(%rbp,%rax,4) movl $0, (%rbx,%rax,4) addq $1, %rax cmpq $1048576, %rax jne .L39 leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call _Z9stopwatchi movl $1, %ecx movl $4194304, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4194304, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L46 .L40: movl $2, %ecx movl $4194304, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl $1, %edi call _Z9stopwatchi movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L47 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .L45: .cfi_restore_state movl $1024, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i jmp .L38 .L46: movl $1024, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i jmp .L40 .L47: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC6: .string "_Z8exam_mulPiS_S_i" .LC7: .string "_Z10shared_mulPiS_S_i" .LC8: .string "_Z8cuda_mulPiS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2090: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z8exam_mulPiS_S_i(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z10shared_mulPiS_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 _Z8cuda_mulPiS_S_i(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2090: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZZ9stopwatchiE5endTS .comm _ZZ9stopwatchiE5endTS,16,16 .local _ZZ9stopwatchiE7startTS .comm _ZZ9stopwatchiE7startTS,16,16 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); cudaMalloc((void**)&device_A,buffer_size ); cudaMalloc((void**)&device_B,buffer_size ); cudaMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); cudaMemcpy(device_A,host_A,buffer_size,cudaMemcpyHostToDevice); cudaMemcpy(device_B,host_B,buffer_size,cudaMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); cudaMemcpy(host_C,device_C,buffer_size,cudaMemcpyDeviceToHost); stopwatch(1); cudaFree(device_A); cudaFree(device_B); cudaFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); hipMalloc((void**)&device_A,buffer_size ); hipMalloc((void**)&device_B,buffer_size ); hipMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); hipFree(device_A); hipFree(device_B); hipFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); hipMalloc((void**)&device_A,buffer_size ); hipMalloc((void**)&device_B,buffer_size ); hipMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); hipFree(device_A); hipFree(device_B); hipFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z8cuda_mulPiS_S_i .globl _Z8cuda_mulPiS_S_i .p2align 8 .type _Z8cuda_mulPiS_S_i,@function _Z8cuda_mulPiS_S_i: s_load_b32 s2, s[0:1], 0x18 v_bfe_u32 v1, v0, 10, 10 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s2, 1 s_cbranch_scc1 .LBB0_3 s_load_b128 s[4:7], s[0:1], 0x0 v_and_b32_e32 v2, 0x3ff, v0 v_mul_lo_u32 v3, v1, s2 v_mov_b32_e32 v4, 0 s_mov_b32 s3, s2 s_delay_alu instid0(VALU_DEP_3) v_mul_lo_u32 v5, v2, s2 v_mov_b32_e32 v2, 0 .p2align 6 .LBB0_2: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_mov_b32_e32 v6, v4 v_lshlrev_b64 v[7:8], 2, v[3:4] s_add_i32 s3, s3, -1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_2) s_cmp_eq_u32 s3, 0 v_lshlrev_b64 v[9:10], 2, v[5:6] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_add_co_u32 v6, vcc_lo, s4, v7 v_add_co_ci_u32_e32 v7, vcc_lo, s5, v8, vcc_lo v_add_co_u32 v8, vcc_lo, s6, v9 s_delay_alu instid0(VALU_DEP_4) v_add_co_ci_u32_e32 v9, vcc_lo, s7, v10, vcc_lo global_load_b32 v10, v[6:7], off global_load_b32 v8, v[8:9], off s_waitcnt vmcnt(0) v_mad_u64_u32 v[6:7], null, v8, v10, v[2:3] v_add_nc_u32_e32 v3, 1, v3 s_delay_alu instid0(VALU_DEP_2) v_dual_mov_b32 v2, v6 :: v_dual_add_nc_u32 v5, 1, v5 s_cbranch_scc0 .LBB0_2 s_branch .LBB0_4 .LBB0_3: v_mov_b32_e32 v2, 0 .LBB0_4: s_load_b64 s[0:1], s[0:1], 0x10 v_and_b32_e32 v0, 0x3ff, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[3:4], null, v0, s2, v[1:2] v_mov_b32_e32 v4, 0 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 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z8cuda_mulPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 28 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 11 .amdhsa_next_free_sgpr 8 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z8cuda_mulPiS_S_i, .Lfunc_end0-_Z8cuda_mulPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z10shared_mulPiS_S_i .globl _Z10shared_mulPiS_S_i .p2align 8 .type _Z10shared_mulPiS_S_i,@function _Z10shared_mulPiS_S_i: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z10shared_mulPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 28 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end1: .size _Z10shared_mulPiS_S_i, .Lfunc_end1-_Z10shared_mulPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .protected _Z8exam_mulPiS_S_i .globl _Z8exam_mulPiS_S_i .p2align 8 .type _Z8exam_mulPiS_S_i,@function _Z8exam_mulPiS_S_i: s_load_b32 s2, s[0:1], 0x18 v_bfe_u32 v3, v0, 10, 10 s_lshl_b32 s8, s14, 6 s_waitcnt lgkmcnt(0) s_lshl_b32 s9, s2, 6 s_cmp_lt_i32 s2, 1 s_mul_i32 s3, s9, s15 s_cbranch_scc1 .LBB2_5 v_and_b32_e32 v0, 0x3ff, v0 s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v4, 8, v3 s_add_i32 s10, s3, s2 s_mov_b32 s11, s8 v_lshlrev_b32_e32 v6, 2, v0 v_mad_u64_u32 v[1:2], null, v3, s2, v[0:1] v_mov_b32_e32 v0, 0 s_mov_b32 s12, s3 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_or_b32_e32 v5, 0x4000, v6 v_add_nc_u32_e32 v2, v4, v6 v_add_nc_u32_e32 v6, v5, v4 s_set_inst_prefetch_distance 0x1 .p2align 6 .LBB2_2: v_add_nc_u32_e32 v7, s12, v1 v_add_nc_u32_e32 v9, s11, v1 s_mov_b32 s13, 0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v8, 31, v7 v_ashrrev_i32_e32 v10, 31, v9 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_lshlrev_b64 v[7:8], 2, v[7:8] v_lshlrev_b64 v[9:10], 2, v[9:10] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_3) v_add_co_u32 v7, vcc_lo, s4, v7 v_add_co_ci_u32_e32 v8, vcc_lo, s5, v8, vcc_lo s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_u32 v9, vcc_lo, s6, v9 v_add_co_ci_u32_e32 v10, vcc_lo, s7, v10, vcc_lo global_load_b32 v8, v[7:8], off global_load_b32 v9, v[9:10], off v_mov_b32_e32 v7, v5 s_waitcnt vmcnt(1) ds_store_b32 v2, v8 s_waitcnt vmcnt(0) ds_store_b32 v6, v9 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv .LBB2_3: v_add_nc_u32_e32 v8, s13, v4 s_add_i32 s13, s13, 4 ds_load_b32 v10, v7 ds_load_b32 v11, v8 s_cmpk_eq_i32 s13, 0x100 s_waitcnt lgkmcnt(0) v_mad_u64_u32 v[8:9], null, v10, v11, v[0:1] s_delay_alu instid0(VALU_DEP_1) v_dual_mov_b32 v0, v8 :: v_dual_add_nc_u32 v7, 0x100, v7 s_cbranch_scc0 .LBB2_3 s_add_i32 s12, s12, 64 s_add_i32 s11, s11, s9 s_cmp_ge_i32 s12, s10 s_barrier buffer_gl0_inv s_cbranch_scc0 .LBB2_2 s_branch .LBB2_6 .LBB2_5: v_mov_b32_e32 v0, 0 .LBB2_6: s_set_inst_prefetch_distance 0x2 v_mul_lo_u32 v1, v3, s2 v_add_nc_u32_e32 v2, s8, v3 s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add3_u32 v1, v2, v1, s3 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[1:2], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v1, vcc_lo, s0, v1 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v2, vcc_lo, s1, v2, vcc_lo global_store_b32 v[1:2], v0, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z8exam_mulPiS_S_i .amdhsa_group_segment_fixed_size 32768 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 28 .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 12 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end2: .size _Z8exam_mulPiS_S_i, .Lfunc_end2-_Z8exam_mulPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 28 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z8cuda_mulPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 10 .sgpr_spill_count: 0 .symbol: _Z8cuda_mulPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 11 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 28 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z10shared_mulPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z10shared_mulPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value .group_segment_fixed_size: 32768 .kernarg_segment_align: 8 .kernarg_segment_size: 28 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z8exam_mulPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z8exam_mulPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 12 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #define MATRIX_TYPE int #define BLOCK_SIZE 8 #define THREAD_SIZE 64 #define TILE 64 void stopwatch(int); //CUDA 배열 곱 __global__ void cuda_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void shared_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); __global__ void exam_mul(MATRIX_TYPE*,MATRIX_TYPE*,MATRIX_TYPE*,int); int main() { //1024 by 1024 행렬 const int width = 1024; const int height = width; const int matrix_size = width*height; const int buffer_size = matrix_size*sizeof(MATRIX_TYPE); MATRIX_TYPE *host_A,*host_B,*host_C; host_A = (MATRIX_TYPE*)malloc(buffer_size); host_B = (MATRIX_TYPE*)malloc(buffer_size); host_C = (MATRIX_TYPE*)malloc(buffer_size); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } printf("Multiply matrix (%dX%d ) * (%dX%d)\n",width,width,width,width); MATRIX_TYPE *device_A,*device_B,*device_C; dim3 Db(1024,1024,1); hipMalloc((void**)&device_A,buffer_size ); hipMalloc((void**)&device_B,buffer_size ); hipMalloc((void**)&device_C,buffer_size ); printf("cuda_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); cuda_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); for(int i=0;i<matrix_size;i++) { host_A[i] = i; host_B[i] = i; host_C[i] =0; } dim3 Sg(BLOCK_SIZE,BLOCK_SIZE,1); dim3 Sb(THREAD_SIZE,THREAD_SIZE,1); printf("shared_mul\n"); stopwatch(0); hipMemcpy(device_A,host_A,buffer_size,hipMemcpyHostToDevice); hipMemcpy(device_B,host_B,buffer_size,hipMemcpyHostToDevice); exam_mul<<<1,Db>>>(device_A,device_B,device_C,width); hipMemcpy(host_C,device_C,buffer_size,hipMemcpyDeviceToHost); stopwatch(1); hipFree(device_A); hipFree(device_B); hipFree(device_C); free(host_A); free(host_B); free(host_C); return 0; } __global__ void cuda_mul(MATRIX_TYPE* A, MATRIX_TYPE* B, MATRIX_TYPE* C, int w) { MATRIX_TYPE v; v = 0; for(int i =0;i<w;i++) { v += A[threadIdx.y*w + i] * B[threadIdx.x *w + i]; } C[threadIdx.x *w + threadIdx.y] = v; } __global__ void shared_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { /* Dg(16,16,1) Db(64,64,1) 0,0 1,0 --------- | 0,1 | 1,1 | 1 0 2 4 1 0 2 4 */ __shared__ MATRIX_TYPE SA[THREAD_SIZE][THREAD_SIZE]; __shared__ MATRIX_TYPE SB[THREAD_SIZE][THREAD_SIZE]; MATRIX_TYPE v; SA[threadIdx.x][threadIdx.y] = A[blockIdx.y *w +blockIdx.x]; SB[threadIdx.x][threadIdx.y] = B[blockIdx.x *w +blockIdx.y]; v = 0; /* A 의 한 타일을 사용하는 모든 B의 타일들을 연산 O O O O X X X X O O O O O O O O X O O O O O O O O O O O O O O O */ } void stopwatch(int flag) { const long long NANOS = 1000000000LL; static struct timespec startTS,endTS; static long long Diff = 0; //start if(flag == 0) { Diff = 0; if(-1 == clock_gettime(CLOCK_MONOTONIC,&startTS)) printf("Failed to call clock_gettime\n"); } //end else if(flag == 1) { if(-1 == clock_gettime(CLOCK_MONOTONIC,&endTS)) printf("Failed to call clock_gettime\n"); Diff = NANOS * (endTS.tv_sec - startTS.tv_sec) + (endTS.tv_nsec - startTS.tv_nsec); printf("elapsed time : % lld micros\n",Diff/1000); } else { printf("wrong flag | 0 : start, 1 : end\n"); } } __global__ void exam_mul(MATRIX_TYPE*A,MATRIX_TYPE*B,MATRIX_TYPE*C,int w) { int bx = blockIdx.x; int by = blockIdx.y; int tx = threadIdx.x; int ty = threadIdx.y; int aBegin = w * TILE * by; int aEnd = aBegin +w -1; int aStep = TILE; int bBegin = TILE *bx; int bStep = TILE * w; MATRIX_TYPE Csub = 0; for(int a = aBegin, b = bBegin; a <= aEnd; a += aStep, b+= bStep) { __shared__ MATRIX_TYPE As[TILE][TILE]; __shared__ MATRIX_TYPE Bs[TILE][TILE]; As[ty][tx] = A[a + w * ty + tx]; Bs[ty][tx] = B[b + w * ty + tx]; __syncthreads(); for(int k=0;k<TILE;k++) Csub += As[ty][k] * Bs[k][tx]; __syncthreads(); } int c = w * TILE * by + TILE * bx; C[c + w * ty + ty] = Csub; }
.text .file "4_sharedMul.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 %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 $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %rbx movl $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %r14 movl $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %r15 xorl %r12d, %r12d movl $4194304, %edx # imm = 0x400000 movq %rax, %rdi xorl %esi, %esi callq memset@PLT .p2align 4, 0x90 .LBB0_1: # =>This Inner Loop Header: Depth=1 movl %r12d, (%rbx,%r12,4) movl %r12d, (%r14,%r12,4) incq %r12 cmpq $1048576, %r12 # imm = 0x100000 jne .LBB0_1 # %bb.2: movl $.L.str, %edi movl $1024, %esi # imm = 0x400 movl $1024, %edx # imm = 0x400 movl $1024, %ecx # imm = 0x400 movl $1024, %r8d # imm = 0x400 xorl %eax, %eax callq printf leaq 32(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc leaq 24(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc leaq 16(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc movl $.Lstr, %edi callq puts@PLT movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_4 # %bb.3: movl $.Lstr.4, %edi callq puts@PLT .LBB0_4: # %_Z9stopwatchi.exit movabsq $4398046512128, %r12 # imm = 0x40000000400 movabsq $4294967297, %r13 # imm = 0x100000001 movq 32(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq %r13, %rdi movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_6 # %bb.5: movq 32(%rsp), %rax movq 24(%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl $1024, 12(%rsp) # imm = 0x400 leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z8cuda_mulPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_6: movq 16(%rsp), %rsi movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_8 # %bb.7: movl $.Lstr.4, %edi callq puts@PLT .LBB0_8: # %_Z9stopwatchi.exit41 movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax movabsq $2361183241434822607, %rbp # imm = 0x20C49BA5E353F7CF imulq %rbp movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi xorl %r12d, %r12d movl $.L.str.4, %edi xorl %eax, %eax callq printf movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi xorl %esi, %esi callq memset@PLT .p2align 4, 0x90 .LBB0_9: # =>This Inner Loop Header: Depth=1 movl %r12d, (%rbx,%r12,4) movl %r12d, (%r14,%r12,4) incq %r12 cmpq $1048576, %r12 # imm = 0x100000 jne .LBB0_9 # %bb.10: movl $.Lstr.1, %edi callq puts@PLT movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_12 # %bb.11: movl $.Lstr.4, %edi callq puts@PLT .LBB0_12: # %_Z9stopwatchi.exit43 movq 32(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq %r13, %rdi movl $1, %esi movabsq $4398046512128, %rdx # imm = 0x40000000400 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_14 # %bb.13: movq 32(%rsp), %rax movq 24(%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl $1024, 12(%rsp) # imm = 0x400 leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z8exam_mulPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_14: movq 16(%rsp), %rsi movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_16 # %bb.15: movl $.Lstr.4, %edi callq puts@PLT .LBB0_16: # %_Z9stopwatchi.exit51 movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax imulq %rbp movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi movl $.L.str.4, %edi xorl %eax, %eax callq printf movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size main, .Lfunc_end0-main .cfi_endproc # -- End function .globl _Z9stopwatchi # -- Begin function _Z9stopwatchi .p2align 4, 0x90 .type _Z9stopwatchi,@function _Z9stopwatchi: # @_Z9stopwatchi .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 cmpl $1, %edi je .LBB1_4 # %bb.1: testl %edi, %edi jne .LBB1_7 # %bb.2: movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax je .LBB1_3 # %bb.8: popq %rax .cfi_def_cfa_offset 8 retq .LBB1_4: .cfi_def_cfa_offset 16 movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB1_6 # %bb.5: movl $.Lstr.4, %edi callq puts@PLT .LBB1_6: movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax movabsq $2361183241434822607, %rcx # imm = 0x20C49BA5E353F7CF imulq %rcx movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi movl $.L.str.4, %edi xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 jmp printf # TAILCALL .LBB1_7: .cfi_def_cfa_offset 16 movl $.Lstr.2, %edi popq %rax .cfi_def_cfa_offset 8 jmp puts@PLT # TAILCALL .LBB1_3: .cfi_def_cfa_offset 16 movl $.Lstr.4, %edi popq %rax .cfi_def_cfa_offset 8 jmp puts@PLT # TAILCALL .Lfunc_end1: .size _Z9stopwatchi, .Lfunc_end1-_Z9stopwatchi .cfi_endproc # -- End function .globl _Z23__device_stub__cuda_mulPiS_S_i # -- Begin function _Z23__device_stub__cuda_mulPiS_S_i .p2align 4, 0x90 .type _Z23__device_stub__cuda_mulPiS_S_i,@function _Z23__device_stub__cuda_mulPiS_S_i: # @_Z23__device_stub__cuda_mulPiS_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 $_Z8cuda_mulPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z23__device_stub__cuda_mulPiS_S_i, .Lfunc_end2-_Z23__device_stub__cuda_mulPiS_S_i .cfi_endproc # -- End function .globl _Z23__device_stub__exam_mulPiS_S_i # -- Begin function _Z23__device_stub__exam_mulPiS_S_i .p2align 4, 0x90 .type _Z23__device_stub__exam_mulPiS_S_i,@function _Z23__device_stub__exam_mulPiS_S_i: # @_Z23__device_stub__exam_mulPiS_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 $_Z8exam_mulPiS_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_end3: .size _Z23__device_stub__exam_mulPiS_S_i, .Lfunc_end3-_Z23__device_stub__exam_mulPiS_S_i .cfi_endproc # -- End function .globl _Z25__device_stub__shared_mulPiS_S_i # -- Begin function _Z25__device_stub__shared_mulPiS_S_i .p2align 4, 0x90 .type _Z25__device_stub__shared_mulPiS_S_i,@function _Z25__device_stub__shared_mulPiS_S_i: # @_Z25__device_stub__shared_mulPiS_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 $_Z10shared_mulPiS_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_end4: .size _Z25__device_stub__shared_mulPiS_S_i, .Lfunc_end4-_Z25__device_stub__shared_mulPiS_S_i .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8cuda_mulPiS_S_i, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z10shared_mulPiS_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 $_Z8exam_mulPiS_S_i, %esi movl $.L__unnamed_3, %edx movl $.L__unnamed_3, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end5: .size __hip_module_ctor, .Lfunc_end5-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB6_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB6_2: retq .Lfunc_end6: .size __hip_module_dtor, .Lfunc_end6-__hip_module_dtor .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Multiply matrix (%dX%d ) * (%dX%d)\n" .size .L.str, 36 .type _Z8cuda_mulPiS_S_i,@object # @_Z8cuda_mulPiS_S_i .section .rodata,"a",@progbits .globl _Z8cuda_mulPiS_S_i .p2align 3, 0x0 _Z8cuda_mulPiS_S_i: .quad _Z23__device_stub__cuda_mulPiS_S_i .size _Z8cuda_mulPiS_S_i, 8 .type _Z8exam_mulPiS_S_i,@object # @_Z8exam_mulPiS_S_i .globl _Z8exam_mulPiS_S_i .p2align 3, 0x0 _Z8exam_mulPiS_S_i: .quad _Z23__device_stub__exam_mulPiS_S_i .size _Z8exam_mulPiS_S_i, 8 .type _Z10shared_mulPiS_S_i,@object # @_Z10shared_mulPiS_S_i .globl _Z10shared_mulPiS_S_i .p2align 3, 0x0 _Z10shared_mulPiS_S_i: .quad _Z25__device_stub__shared_mulPiS_S_i .size _Z10shared_mulPiS_S_i, 8 .type _ZZ9stopwatchiE7startTS,@object # @_ZZ9stopwatchiE7startTS .local _ZZ9stopwatchiE7startTS .comm _ZZ9stopwatchiE7startTS,16,8 .type _ZZ9stopwatchiE5endTS,@object # @_ZZ9stopwatchiE5endTS .local _ZZ9stopwatchiE5endTS .comm _ZZ9stopwatchiE5endTS,16,8 .type .L.str.4,@object # @.str.4 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.4: .asciz "elapsed time : % lld micros\n" .size .L.str.4, 29 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z8cuda_mulPiS_S_i" .size .L__unnamed_1, 19 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z10shared_mulPiS_S_i" .size .L__unnamed_2, 22 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z8exam_mulPiS_S_i" .size .L__unnamed_3, 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 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "cuda_mul" .size .Lstr, 9 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "shared_mul" .size .Lstr.1, 11 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "wrong flag | 0 : start, 1 : end" .size .Lstr.2, 32 .type .Lstr.4,@object # @str.4 .Lstr.4: .asciz "Failed to call clock_gettime" .size .Lstr.4, 29 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__cuda_mulPiS_S_i .addrsig_sym _Z23__device_stub__exam_mulPiS_S_i .addrsig_sym _Z25__device_stub__shared_mulPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8cuda_mulPiS_S_i .addrsig_sym _Z8exam_mulPiS_S_i .addrsig_sym _Z10shared_mulPiS_S_i .addrsig_sym _ZZ9stopwatchiE7startTS .addrsig_sym _ZZ9stopwatchiE5endTS .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_001149e5_00000000-6_4_sharedMul.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Failed to call clock_gettime\n" .LC1: .string "elapsed time : % lld micros\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "wrong flag | 0 : start, 1 : end\n" .text .globl _Z9stopwatchi .type _Z9stopwatchi, @function _Z9stopwatchi: .LFB2058: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 testl %edi, %edi je .L9 cmpl $1, %edi je .L10 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L3: addq $8, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .cfi_restore_state leaq _ZZ9stopwatchiE7startTS(%rip), %rsi movl $1, %edi call clock_gettime@PLT cmpl $-1, %eax jne .L3 leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L3 .L10: leaq _ZZ9stopwatchiE5endTS(%rip), %rsi call clock_gettime@PLT cmpl $-1, %eax je .L11 .L7: movq _ZZ9stopwatchiE5endTS(%rip), %rcx subq _ZZ9stopwatchiE7startTS(%rip), %rcx imulq $1000000000, %rcx, %rcx addq 8+_ZZ9stopwatchiE5endTS(%rip), %rcx subq 8+_ZZ9stopwatchiE7startTS(%rip), %rcx movabsq $2361183241434822607, %rdx movq %rcx, %rax imulq %rdx sarq $7, %rdx sarq $63, %rcx subq %rcx, %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L3 .L11: leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L7 .cfi_endproc .LFE2058: .size _Z9stopwatchi, .-_Z9stopwatchi .globl _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i .type _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i, @function _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i: .LFB2083: .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 .L16 .L12: movq 136(%rsp), %rax subq %fs:40, %rax jne .L17 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L16: .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 _Z8cuda_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L12 .L17: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i, .-_Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i .globl _Z8cuda_mulPiS_S_i .type _Z8cuda_mulPiS_S_i, @function _Z8cuda_mulPiS_S_i: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z8cuda_mulPiS_S_i, .-_Z8cuda_mulPiS_S_i .globl _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i .type _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i, @function _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i: .LFB2085: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L24 .L20: movq 136(%rsp), %rax subq %fs:40, %rax jne .L25 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L24: .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 _Z10shared_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L20 .L25: call __stack_chk_fail@PLT .cfi_endproc .LFE2085: .size _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i, .-_Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i .globl _Z10shared_mulPiS_S_i .type _Z10shared_mulPiS_S_i, @function _Z10shared_mulPiS_S_i: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10shared_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _Z10shared_mulPiS_S_i, .-_Z10shared_mulPiS_S_i .globl _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i .type _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i, @function _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i: .LFB2087: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L32 .L28: movq 136(%rsp), %rax subq %fs:40, %rax jne .L33 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L32: .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 _Z8exam_mulPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L28 .L33: call __stack_chk_fail@PLT .cfi_endproc .LFE2087: .size _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i, .-_Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i .globl _Z8exam_mulPiS_S_i .type _Z8exam_mulPiS_S_i, @function _Z8exam_mulPiS_S_i: .LFB2088: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2088: .size _Z8exam_mulPiS_S_i, .-_Z8exam_mulPiS_S_i .section .rodata.str1.8 .align 8 .LC3: .string "Multiply matrix (%dX%d ) * (%dX%d)\n" .section .rodata.str1.1 .LC4: .string "cuda_mul\n" .LC5: .string "shared_mul\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r12 .cfi_def_cfa_offset 16 .cfi_offset 12, -16 pushq %rbp .cfi_def_cfa_offset 24 .cfi_offset 6, -24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset 3, -32 subq $64, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $4194304, %edi call malloc@PLT movq %rax, %r12 movl $4194304, %edi call malloc@PLT movq %rax, %rbp movl $4194304, %edi call malloc@PLT movq %rax, %rbx movl $0, %eax .L37: movl %eax, (%r12,%rax,4) movl %eax, 0(%rbp,%rax,4) movl $0, (%rbx,%rax,4) addq $1, %rax cmpq $1048576, %rax jne .L37 movl $1024, %r9d movl $1024, %r8d movl $1024, %ecx movl $1024, %edx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $1024, 32(%rsp) movl $1024, 36(%rsp) movl $1, 40(%rsp) leaq 8(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $4194304, %esi call cudaMalloc@PLT leaq .LC4(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call _Z9stopwatchi movl $1, %ecx movl $4194304, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4194304, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L45 .L38: movl $2, %ecx movl $4194304, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl $1, %edi call _Z9stopwatchi movl $0, %eax .L39: movl %eax, (%r12,%rax,4) movl %eax, 0(%rbp,%rax,4) movl $0, (%rbx,%rax,4) addq $1, %rax cmpq $1048576, %rax jne .L39 leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call _Z9stopwatchi movl $1, %ecx movl $4194304, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movl $4194304, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl 40(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 32(%rsp), %rdx movq 44(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L46 .L40: movl $2, %ecx movl $4194304, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl $1, %edi call _Z9stopwatchi movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L47 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %rbp .cfi_def_cfa_offset 16 popq %r12 .cfi_def_cfa_offset 8 ret .L45: .cfi_restore_state movl $1024, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z32__device_stub__Z8cuda_mulPiS_S_iPiS_S_i jmp .L38 .L46: movl $1024, %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z32__device_stub__Z8exam_mulPiS_S_iPiS_S_i jmp .L40 .L47: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC6: .string "_Z8exam_mulPiS_S_i" .LC7: .string "_Z10shared_mulPiS_S_i" .LC8: .string "_Z8cuda_mulPiS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2090: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z8exam_mulPiS_S_i(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _Z10shared_mulPiS_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 _Z8cuda_mulPiS_S_i(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2090: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZZ9stopwatchiE5endTS .comm _ZZ9stopwatchiE5endTS,16,16 .local _ZZ9stopwatchiE7startTS .comm _ZZ9stopwatchiE7startTS,16,16 .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 "4_sharedMul.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 %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 $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %rbx movl $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %r14 movl $4194304, %edi # imm = 0x400000 callq malloc movq %rax, %r15 xorl %r12d, %r12d movl $4194304, %edx # imm = 0x400000 movq %rax, %rdi xorl %esi, %esi callq memset@PLT .p2align 4, 0x90 .LBB0_1: # =>This Inner Loop Header: Depth=1 movl %r12d, (%rbx,%r12,4) movl %r12d, (%r14,%r12,4) incq %r12 cmpq $1048576, %r12 # imm = 0x100000 jne .LBB0_1 # %bb.2: movl $.L.str, %edi movl $1024, %esi # imm = 0x400 movl $1024, %edx # imm = 0x400 movl $1024, %ecx # imm = 0x400 movl $1024, %r8d # imm = 0x400 xorl %eax, %eax callq printf leaq 32(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc leaq 24(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc leaq 16(%rsp), %rdi movl $4194304, %esi # imm = 0x400000 callq hipMalloc movl $.Lstr, %edi callq puts@PLT movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_4 # %bb.3: movl $.Lstr.4, %edi callq puts@PLT .LBB0_4: # %_Z9stopwatchi.exit movabsq $4398046512128, %r12 # imm = 0x40000000400 movabsq $4294967297, %r13 # imm = 0x100000001 movq 32(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq %r13, %rdi movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_6 # %bb.5: movq 32(%rsp), %rax movq 24(%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl $1024, 12(%rsp) # imm = 0x400 leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z8cuda_mulPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_6: movq 16(%rsp), %rsi movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_8 # %bb.7: movl $.Lstr.4, %edi callq puts@PLT .LBB0_8: # %_Z9stopwatchi.exit41 movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax movabsq $2361183241434822607, %rbp # imm = 0x20C49BA5E353F7CF imulq %rbp movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi xorl %r12d, %r12d movl $.L.str.4, %edi xorl %eax, %eax callq printf movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi xorl %esi, %esi callq memset@PLT .p2align 4, 0x90 .LBB0_9: # =>This Inner Loop Header: Depth=1 movl %r12d, (%rbx,%r12,4) movl %r12d, (%r14,%r12,4) incq %r12 cmpq $1048576, %r12 # imm = 0x100000 jne .LBB0_9 # %bb.10: movl $.Lstr.1, %edi callq puts@PLT movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_12 # %bb.11: movl $.Lstr.4, %edi callq puts@PLT .LBB0_12: # %_Z9stopwatchi.exit43 movq 32(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 24(%rsp), %rdi movl $4194304, %edx # imm = 0x400000 movq %r14, %rsi movl $1, %ecx callq hipMemcpy movq %r13, %rdi movl $1, %esi movabsq $4398046512128, %rdx # imm = 0x40000000400 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB0_14 # %bb.13: movq 32(%rsp), %rax movq 24(%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) movl $1024, 12(%rsp) # imm = 0x400 leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z8exam_mulPiS_S_i, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB0_14: movq 16(%rsp), %rsi movl $4194304, %edx # imm = 0x400000 movq %r15, %rdi movl $2, %ecx callq hipMemcpy movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB0_16 # %bb.15: movl $.Lstr.4, %edi callq puts@PLT .LBB0_16: # %_Z9stopwatchi.exit51 movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax imulq %rbp movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi movl $.L.str.4, %edi xorl %eax, %eax callq printf movq 32(%rsp), %rdi callq hipFree movq 24(%rsp), %rdi callq hipFree movq 16(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free xorl %eax, %eax addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size main, .Lfunc_end0-main .cfi_endproc # -- End function .globl _Z9stopwatchi # -- Begin function _Z9stopwatchi .p2align 4, 0x90 .type _Z9stopwatchi,@function _Z9stopwatchi: # @_Z9stopwatchi .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 cmpl $1, %edi je .LBB1_4 # %bb.1: testl %edi, %edi jne .LBB1_7 # %bb.2: movl $_ZZ9stopwatchiE7startTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax je .LBB1_3 # %bb.8: popq %rax .cfi_def_cfa_offset 8 retq .LBB1_4: .cfi_def_cfa_offset 16 movl $_ZZ9stopwatchiE5endTS, %esi movl $1, %edi callq clock_gettime cmpl $-1, %eax jne .LBB1_6 # %bb.5: movl $.Lstr.4, %edi callq puts@PLT .LBB1_6: movq _ZZ9stopwatchiE5endTS(%rip), %rax subq _ZZ9stopwatchiE7startTS(%rip), %rax imulq $1000000000, %rax, %rcx # imm = 0x3B9ACA00 movq _ZZ9stopwatchiE5endTS+8(%rip), %rax subq _ZZ9stopwatchiE7startTS+8(%rip), %rax addq %rcx, %rax movabsq $2361183241434822607, %rcx # imm = 0x20C49BA5E353F7CF imulq %rcx movq %rdx, %rsi shrq $63, %rsi sarq $7, %rdx addq %rdx, %rsi movl $.L.str.4, %edi xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 jmp printf # TAILCALL .LBB1_7: .cfi_def_cfa_offset 16 movl $.Lstr.2, %edi popq %rax .cfi_def_cfa_offset 8 jmp puts@PLT # TAILCALL .LBB1_3: .cfi_def_cfa_offset 16 movl $.Lstr.4, %edi popq %rax .cfi_def_cfa_offset 8 jmp puts@PLT # TAILCALL .Lfunc_end1: .size _Z9stopwatchi, .Lfunc_end1-_Z9stopwatchi .cfi_endproc # -- End function .globl _Z23__device_stub__cuda_mulPiS_S_i # -- Begin function _Z23__device_stub__cuda_mulPiS_S_i .p2align 4, 0x90 .type _Z23__device_stub__cuda_mulPiS_S_i,@function _Z23__device_stub__cuda_mulPiS_S_i: # @_Z23__device_stub__cuda_mulPiS_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 $_Z8cuda_mulPiS_S_i, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end2: .size _Z23__device_stub__cuda_mulPiS_S_i, .Lfunc_end2-_Z23__device_stub__cuda_mulPiS_S_i .cfi_endproc # -- End function .globl _Z23__device_stub__exam_mulPiS_S_i # -- Begin function _Z23__device_stub__exam_mulPiS_S_i .p2align 4, 0x90 .type _Z23__device_stub__exam_mulPiS_S_i,@function _Z23__device_stub__exam_mulPiS_S_i: # @_Z23__device_stub__exam_mulPiS_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 $_Z8exam_mulPiS_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_end3: .size _Z23__device_stub__exam_mulPiS_S_i, .Lfunc_end3-_Z23__device_stub__exam_mulPiS_S_i .cfi_endproc # -- End function .globl _Z25__device_stub__shared_mulPiS_S_i # -- Begin function _Z25__device_stub__shared_mulPiS_S_i .p2align 4, 0x90 .type _Z25__device_stub__shared_mulPiS_S_i,@function _Z25__device_stub__shared_mulPiS_S_i: # @_Z25__device_stub__shared_mulPiS_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 $_Z10shared_mulPiS_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_end4: .size _Z25__device_stub__shared_mulPiS_S_i, .Lfunc_end4-_Z25__device_stub__shared_mulPiS_S_i .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8cuda_mulPiS_S_i, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z10shared_mulPiS_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 $_Z8exam_mulPiS_S_i, %esi movl $.L__unnamed_3, %edx movl $.L__unnamed_3, %ecx movq %rbx, %rdi movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $32, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end5: .size __hip_module_ctor, .Lfunc_end5-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB6_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB6_2: retq .Lfunc_end6: .size __hip_module_dtor, .Lfunc_end6-__hip_module_dtor .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Multiply matrix (%dX%d ) * (%dX%d)\n" .size .L.str, 36 .type _Z8cuda_mulPiS_S_i,@object # @_Z8cuda_mulPiS_S_i .section .rodata,"a",@progbits .globl _Z8cuda_mulPiS_S_i .p2align 3, 0x0 _Z8cuda_mulPiS_S_i: .quad _Z23__device_stub__cuda_mulPiS_S_i .size _Z8cuda_mulPiS_S_i, 8 .type _Z8exam_mulPiS_S_i,@object # @_Z8exam_mulPiS_S_i .globl _Z8exam_mulPiS_S_i .p2align 3, 0x0 _Z8exam_mulPiS_S_i: .quad _Z23__device_stub__exam_mulPiS_S_i .size _Z8exam_mulPiS_S_i, 8 .type _Z10shared_mulPiS_S_i,@object # @_Z10shared_mulPiS_S_i .globl _Z10shared_mulPiS_S_i .p2align 3, 0x0 _Z10shared_mulPiS_S_i: .quad _Z25__device_stub__shared_mulPiS_S_i .size _Z10shared_mulPiS_S_i, 8 .type _ZZ9stopwatchiE7startTS,@object # @_ZZ9stopwatchiE7startTS .local _ZZ9stopwatchiE7startTS .comm _ZZ9stopwatchiE7startTS,16,8 .type _ZZ9stopwatchiE5endTS,@object # @_ZZ9stopwatchiE5endTS .local _ZZ9stopwatchiE5endTS .comm _ZZ9stopwatchiE5endTS,16,8 .type .L.str.4,@object # @.str.4 .section .rodata.str1.1,"aMS",@progbits,1 .L.str.4: .asciz "elapsed time : % lld micros\n" .size .L.str.4, 29 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z8cuda_mulPiS_S_i" .size .L__unnamed_1, 19 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z10shared_mulPiS_S_i" .size .L__unnamed_2, 22 .type .L__unnamed_3,@object # @2 .L__unnamed_3: .asciz "_Z8exam_mulPiS_S_i" .size .L__unnamed_3, 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 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "cuda_mul" .size .Lstr, 9 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "shared_mul" .size .Lstr.1, 11 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "wrong flag | 0 : start, 1 : end" .size .Lstr.2, 32 .type .Lstr.4,@object # @str.4 .Lstr.4: .asciz "Failed to call clock_gettime" .size .Lstr.4, 29 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__cuda_mulPiS_S_i .addrsig_sym _Z23__device_stub__exam_mulPiS_S_i .addrsig_sym _Z25__device_stub__shared_mulPiS_S_i .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8cuda_mulPiS_S_i .addrsig_sym _Z8exam_mulPiS_S_i .addrsig_sym _Z10shared_mulPiS_S_i .addrsig_sym _ZZ9stopwatchiE7startTS .addrsig_sym _ZZ9stopwatchiE5endTS .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <thrust/host_vector.h> #include <thrust/device_vector.h> #include <thrust/scan.h> #include <thrust/functional.h> // #include <iomanip> #include <iostream> int main(int argc, char *argv[]) { long n = atol(argv[1]); cudaEvent_t start; cudaEvent_t stop; cudaEventCreate(&start); cudaEventCreate(&stop); thrust::host_vector<float> h_vec(n); for (long i = 0; i < n; i++) { h_vec[i] = 1.1; } thrust::device_vector<float> d_vec(h_vec.size()); thrust::device_vector<float> d_vec1(h_vec.size()); thrust::copy(h_vec.begin(), h_vec.end(), d_vec.begin()); float init = 0.0; cudaEventRecord(start); // No need to allocate another device vector. Do it in place thrust::exclusive_scan(d_vec.begin(), d_vec.end(), d_vec1.begin(), init, thrust::plus<float>()); cudaEventRecord(stop); cudaEventSynchronize(stop); float ms; cudaEventElapsedTime(&ms, start, stop); thrust::copy(d_vec1.begin(), d_vec1.end(), h_vec.begin()); //for (long i = 0; i < n; i++) { // std::cout << setprecision(12) << h_vec[i] << " "; //} //std::cout << std::endl; std::cout << h_vec[n-1] << std::endl; std::cout << ms << std::endl; return 0; }
#include <hip/hip_runtime.h> #include <thrust/host_vector.h> #include <thrust/device_vector.h> #include <thrust/scan.h> #include <thrust/functional.h> // #include <iomanip> #include <iostream> int main(int argc, char *argv[]) { long n = atol(argv[1]); hipEvent_t start; hipEvent_t stop; hipEventCreate(&start); hipEventCreate(&stop); thrust::host_vector<float> h_vec(n); for (long i = 0; i < n; i++) { h_vec[i] = 1.1; } thrust::device_vector<float> d_vec(h_vec.size()); thrust::device_vector<float> d_vec1(h_vec.size()); thrust::copy(h_vec.begin(), h_vec.end(), d_vec.begin()); float init = 0.0; hipEventRecord(start); // No need to allocate another device vector. Do it in place thrust::exclusive_scan(d_vec.begin(), d_vec.end(), d_vec1.begin(), init, thrust::plus<float>()); hipEventRecord(stop); hipEventSynchronize(stop); float ms; hipEventElapsedTime(&ms, start, stop); thrust::copy(d_vec1.begin(), d_vec1.end(), h_vec.begin()); //for (long i = 0; i < n; i++) { // std::cout << setprecision(12) << h_vec[i] << " "; //} //std::cout << std::endl; std::cout << h_vec[n-1] << std::endl; std::cout << ms << std::endl; return 0; }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
// includes, system #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { cudaSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); cudaMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed cudaThreadSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy cudaMemcpy( h_a, d_a, memSize ,cudaMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("cudaMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory cudaFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { cudaError_t err = cudaGetLastError(); if( cudaSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, cudaGetErrorString( err) ); exit(-1); } }
code for sm_80 Function : _Z13myFirstKernelPi .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 R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R5.reuse, c[0x0][0x0], R0 ; /* 0x0000000005027a24 */ /* 0x041fe200078e0200 */ /*0060*/ IADD3 R5, R5, R0, RZ ; /* 0x0000000005057210 */ /* 0x000fc80007ffe0ff */ /*0070*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*0080*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0090*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00a0*/ BRA 0xa0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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.
// includes, system #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { cudaSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); cudaMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed cudaThreadSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy cudaMemcpy( h_a, d_a, memSize ,cudaMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("cudaMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory cudaFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { cudaError_t err = cudaGetLastError(); if( cudaSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, cudaGetErrorString( err) ); exit(-1); } }
.file "tmpxft_0015b402_00000000-6_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Cuda error: %s: %s.\n" .text .globl _Z14checkCUDAErrorPKc .type _Z14checkCUDAErrorPKc, @function _Z14checkCUDAErrorPKc: .LFB2058: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movq %rdi, %rbx call cudaGetLastError@PLT testl %eax, %eax jne .L6 popq %rbx .cfi_remember_state .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movq %rbx, %rcx leaq .LC0(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $-1, %edi call exit@PLT .cfi_endproc .LFE2058: .size _Z14checkCUDAErrorPKc, .-_Z14checkCUDAErrorPKc .globl _Z33__device_stub__Z13myFirstKernelPiPi .type _Z33__device_stub__Z13myFirstKernelPiPi, @function _Z33__device_stub__Z13myFirstKernelPiPi: .LFB2083: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z13myFirstKernelPi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z33__device_stub__Z13myFirstKernelPiPi, .-_Z33__device_stub__Z13myFirstKernelPiPi .globl _Z13myFirstKernelPi .type _Z13myFirstKernelPi, @function _Z13myFirstKernelPi: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z33__device_stub__Z13myFirstKernelPiPi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z13myFirstKernelPi, .-_Z13myFirstKernelPi .section .rodata.str1.1 .LC1: .string "kernel execution" .LC2: .string "cudaMemcpy" .LC3: .string "Correct!\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $48, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0, %edi call cudaSetDevice@PLT movl $256, %edi call malloc@PLT movq %rax, %rbx leaq 8(%rsp), %rdi movl $256, %esi call cudaMalloc@PLT movl $8, 16(%rsp) movl $1, 20(%rsp) movl $8, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L19 .L16: call cudaThreadSynchronize@PLT leaq .LC1(%rip), %rdi call _Z14checkCUDAErrorPKc movl $2, %ecx movl $256, %edx movq 8(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rdi call _Z14checkCUDAErrorPKc movq 8(%rsp), %rdi call cudaFree@PLT movq %rbx, %rdi call free@PLT leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L20 movl $0, %eax addq $48, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state movq 8(%rsp), %rdi call _Z33__device_stub__Z13myFirstKernelPiPi jmp .L16 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z13myFirstKernelPi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z13myFirstKernelPi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
// includes, system #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { cudaSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); cudaMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed cudaThreadSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy cudaMemcpy( h_a, d_a, memSize ,cudaMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("cudaMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory cudaFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { cudaError_t err = cudaGetLastError(); if( cudaSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, cudaGetErrorString( err) ); exit(-1); } }
// includes, system #include <hip/hip_runtime.h> #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { hipSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); hipMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed hipDeviceSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy hipMemcpy( h_a, d_a, memSize ,hipMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("hipMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory hipFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { hipError_t err = hipGetLastError(); if( hipSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, hipGetErrorString( err) ); exit(-1); } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
// includes, system #include <hip/hip_runtime.h> #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { hipSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); hipMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed hipDeviceSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy hipMemcpy( h_a, d_a, memSize ,hipMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("hipMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory hipFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { hipError_t err = hipGetLastError(); if( hipSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, hipGetErrorString( err) ); exit(-1); } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z13myFirstKernelPi .globl _Z13myFirstKernelPi .p2align 8 .type _Z13myFirstKernelPi,@function _Z13myFirstKernelPi: s_clause 0x1 s_load_b32 s2, s[0:1], 0x14 s_load_b64 s[0:1], s[0:1], 0x0 v_add_nc_u32_e32 v3, s15, v0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[1:2], 2, v[1:2] v_add_co_u32 v0, vcc_lo, s0, v1 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v2, vcc_lo global_store_b32 v[0:1], v3, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13myFirstKernelPi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 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 _Z13myFirstKernelPi, .Lfunc_end0-_Z13myFirstKernelPi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z13myFirstKernelPi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13myFirstKernelPi.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.
// includes, system #include <hip/hip_runtime.h> #include <stdio.h> #include <assert.h> // Here you can set the device ID that was assigned to you #define MYDEVICE 0 // Simple utility function to check for CUDA runtime errors void checkCUDAError(const char *msg); // Part 3 of 5: implement the kernel __global__ void myFirstKernel(int *d_a) { int i = blockIdx.x*blockDim.x + threadIdx.x; d_a[i] = blockIdx.x + threadIdx.x; } //////////////////////////////////////////////////////////////////////////////// // Program main //////////////////////////////////////////////////////////////////////////////// int main( int argc, char** argv) { hipSetDevice(MYDEVICE); // pointer for host memory int *h_a; // pointer for device memory int *d_a; // define grid and block size int numBlocks = 8; int numThreadsPerBlock = 8; // Part 1 of 5: allocate host and device memory size_t memSize = numBlocks * numThreadsPerBlock * sizeof(int); h_a = (int *) malloc(memSize); hipMalloc((void **) &d_a, memSize); // Part 2 of 5: configure and launch kernel dim3 dimGrid( numBlocks ); dim3 dimBlock( numThreadsPerBlock ); myFirstKernel<<< dimGrid , dimBlock >>>(d_a); // block until the device has completed hipDeviceSynchronize(); // check if kernel execution generated an error checkCUDAError("kernel execution"); // Part 4 of 5: device to host copy hipMemcpy( h_a, d_a, memSize ,hipMemcpyDeviceToHost); // Check for any CUDA errors checkCUDAError("hipMemcpy"); // Part 5 of 5: verify the data returned to the host is correct for (int i = 0; i < numBlocks ; i++) { for (int j = 0; j < numThreadsPerBlock ; j++) { assert(h_a[i * numThreadsPerBlock + j] == i + j); } } // free device memory hipFree(d_a); // free host memory free(h_a); // If the program makes it this far, then the results are correct and // there are no run-time errors. Good work! printf("Correct!\n"); return 0; } void checkCUDAError(const char *msg) { hipError_t err = hipGetLastError(); if( hipSuccess != err) { fprintf(stderr, "Cuda error: %s: %s.\n", msg, hipGetErrorString( err) ); exit(-1); } }
.text .file "kernel.hip" .globl _Z28__device_stub__myFirstKernelPi # -- Begin function _Z28__device_stub__myFirstKernelPi .p2align 4, 0x90 .type _Z28__device_stub__myFirstKernelPi,@function _Z28__device_stub__myFirstKernelPi: # @_Z28__device_stub__myFirstKernelPi .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 $_Z13myFirstKernelPi, %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 _Z28__device_stub__myFirstKernelPi, .Lfunc_end0-_Z28__device_stub__myFirstKernelPi .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $80, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -16 xorl %edi, %edi callq hipSetDevice movl $256, %edi # imm = 0x100 callq malloc movq %rax, %rbx leaq 8(%rsp), %rdi movl $256, %esi # imm = 0x100 callq hipMalloc movabsq $4294967304, %rdi # imm = 0x100000008 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 8(%rsp), %rax movq %rax, 72(%rsp) leaq 72(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z13myFirstKernelPi, %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 .LBB1_2: callq hipDeviceSynchronize callq hipGetLastError testl %eax, %eax jne .LBB1_3 # %bb.5: # %_Z14checkCUDAErrorPKc.exit movq 8(%rsp), %rsi movl $256, %edx # imm = 0x100 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy callq hipGetLastError testl %eax, %eax jne .LBB1_6 # %bb.7: # %.preheader.preheader movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movl $.Lstr, %edi callq puts@PLT xorl %eax, %eax addq $80, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_3: .cfi_def_cfa_offset 96 movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movl $.L.str, %edx jmp .LBB1_4 .LBB1_6: movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movl $.L.str.1, %edx .LBB1_4: movq %rbx, %rdi movq %rax, %rcx xorl %eax, %eax callq fprintf movl $-1, %edi callq exit .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .globl _Z14checkCUDAErrorPKc # -- Begin function _Z14checkCUDAErrorPKc .p2align 4, 0x90 .type _Z14checkCUDAErrorPKc,@function _Z14checkCUDAErrorPKc: # @_Z14checkCUDAErrorPKc .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 %rdi, %rbx callq hipGetLastError testl %eax, %eax jne .LBB2_2 # %bb.1: addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .LBB2_2: .cfi_def_cfa_offset 32 movq stderr(%rip), %r14 movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movq %r14, %rdi movq %rbx, %rdx movq %rax, %rcx xorl %eax, %eax callq fprintf movl $-1, %edi callq exit .Lfunc_end2: .size _Z14checkCUDAErrorPKc, .Lfunc_end2-_Z14checkCUDAErrorPKc .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 $_Z13myFirstKernelPi, %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 _Z13myFirstKernelPi,@object # @_Z13myFirstKernelPi .section .rodata,"a",@progbits .globl _Z13myFirstKernelPi .p2align 3, 0x0 _Z13myFirstKernelPi: .quad _Z28__device_stub__myFirstKernelPi .size _Z13myFirstKernelPi, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "kernel execution" .size .L.str, 17 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "hipMemcpy" .size .L.str.1, 10 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Cuda error: %s: %s.\n" .size .L.str.3, 21 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z13myFirstKernelPi" .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 "Correct!" .size .Lstr, 9 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z28__device_stub__myFirstKernelPi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z13myFirstKernelPi .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 : _Z13myFirstKernelPi .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 R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R5.reuse, c[0x0][0x0], R0 ; /* 0x0000000005027a24 */ /* 0x041fe200078e0200 */ /*0060*/ IADD3 R5, R5, R0, RZ ; /* 0x0000000005057210 */ /* 0x000fc80007ffe0ff */ /*0070*/ IMAD.WIDE R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0203 */ /*0080*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0090*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00a0*/ BRA 0xa0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 _Z13myFirstKernelPi .globl _Z13myFirstKernelPi .p2align 8 .type _Z13myFirstKernelPi,@function _Z13myFirstKernelPi: s_clause 0x1 s_load_b32 s2, s[0:1], 0x14 s_load_b64 s[0:1], s[0:1], 0x0 v_add_nc_u32_e32 v3, s15, v0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[1:2], 2, v[1:2] v_add_co_u32 v0, vcc_lo, s0, v1 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v2, vcc_lo global_store_b32 v[0:1], v3, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z13myFirstKernelPi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 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 _Z13myFirstKernelPi, .Lfunc_end0-_Z13myFirstKernelPi .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z13myFirstKernelPi .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z13myFirstKernelPi.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_0015b402_00000000-6_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "Cuda error: %s: %s.\n" .text .globl _Z14checkCUDAErrorPKc .type _Z14checkCUDAErrorPKc, @function _Z14checkCUDAErrorPKc: .LFB2058: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movq %rdi, %rbx call cudaGetLastError@PLT testl %eax, %eax jne .L6 popq %rbx .cfi_remember_state .cfi_def_cfa_offset 8 ret .L6: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movq %rbx, %rcx leaq .LC0(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $-1, %edi call exit@PLT .cfi_endproc .LFE2058: .size _Z14checkCUDAErrorPKc, .-_Z14checkCUDAErrorPKc .globl _Z33__device_stub__Z13myFirstKernelPiPi .type _Z33__device_stub__Z13myFirstKernelPiPi, @function _Z33__device_stub__Z13myFirstKernelPiPi: .LFB2083: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z13myFirstKernelPi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z33__device_stub__Z13myFirstKernelPiPi, .-_Z33__device_stub__Z13myFirstKernelPiPi .globl _Z13myFirstKernelPi .type _Z13myFirstKernelPi, @function _Z13myFirstKernelPi: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z33__device_stub__Z13myFirstKernelPiPi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z13myFirstKernelPi, .-_Z13myFirstKernelPi .section .rodata.str1.1 .LC1: .string "kernel execution" .LC2: .string "cudaMemcpy" .LC3: .string "Correct!\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $48, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0, %edi call cudaSetDevice@PLT movl $256, %edi call malloc@PLT movq %rax, %rbx leaq 8(%rsp), %rdi movl $256, %esi call cudaMalloc@PLT movl $8, 16(%rsp) movl $1, 20(%rsp) movl $8, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L19 .L16: call cudaThreadSynchronize@PLT leaq .LC1(%rip), %rdi call _Z14checkCUDAErrorPKc movl $2, %ecx movl $256, %edx movq 8(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rdi call _Z14checkCUDAErrorPKc movq 8(%rsp), %rdi call cudaFree@PLT movq %rbx, %rdi call free@PLT leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L20 movl $0, %eax addq $48, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state movq 8(%rsp), %rdi call _Z33__device_stub__Z13myFirstKernelPiPi jmp .L16 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z13myFirstKernelPi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z13myFirstKernelPi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "kernel.hip" .globl _Z28__device_stub__myFirstKernelPi # -- Begin function _Z28__device_stub__myFirstKernelPi .p2align 4, 0x90 .type _Z28__device_stub__myFirstKernelPi,@function _Z28__device_stub__myFirstKernelPi: # @_Z28__device_stub__myFirstKernelPi .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 $_Z13myFirstKernelPi, %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 _Z28__device_stub__myFirstKernelPi, .Lfunc_end0-_Z28__device_stub__myFirstKernelPi .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $80, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -16 xorl %edi, %edi callq hipSetDevice movl $256, %edi # imm = 0x100 callq malloc movq %rax, %rbx leaq 8(%rsp), %rdi movl $256, %esi # imm = 0x100 callq hipMalloc movabsq $4294967304, %rdi # imm = 0x100000008 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 8(%rsp), %rax movq %rax, 72(%rsp) leaq 72(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z13myFirstKernelPi, %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 .LBB1_2: callq hipDeviceSynchronize callq hipGetLastError testl %eax, %eax jne .LBB1_3 # %bb.5: # %_Z14checkCUDAErrorPKc.exit movq 8(%rsp), %rsi movl $256, %edx # imm = 0x100 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy callq hipGetLastError testl %eax, %eax jne .LBB1_6 # %bb.7: # %.preheader.preheader movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free movl $.Lstr, %edi callq puts@PLT xorl %eax, %eax addq $80, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_3: .cfi_def_cfa_offset 96 movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movl $.L.str, %edx jmp .LBB1_4 .LBB1_6: movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movl $.L.str.1, %edx .LBB1_4: movq %rbx, %rdi movq %rax, %rcx xorl %eax, %eax callq fprintf movl $-1, %edi callq exit .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .globl _Z14checkCUDAErrorPKc # -- Begin function _Z14checkCUDAErrorPKc .p2align 4, 0x90 .type _Z14checkCUDAErrorPKc,@function _Z14checkCUDAErrorPKc: # @_Z14checkCUDAErrorPKc .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 %rdi, %rbx callq hipGetLastError testl %eax, %eax jne .LBB2_2 # %bb.1: addq $8, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .LBB2_2: .cfi_def_cfa_offset 32 movq stderr(%rip), %r14 movl %eax, %edi callq hipGetErrorString movl $.L.str.3, %esi movq %r14, %rdi movq %rbx, %rdx movq %rax, %rcx xorl %eax, %eax callq fprintf movl $-1, %edi callq exit .Lfunc_end2: .size _Z14checkCUDAErrorPKc, .Lfunc_end2-_Z14checkCUDAErrorPKc .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 $_Z13myFirstKernelPi, %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 _Z13myFirstKernelPi,@object # @_Z13myFirstKernelPi .section .rodata,"a",@progbits .globl _Z13myFirstKernelPi .p2align 3, 0x0 _Z13myFirstKernelPi: .quad _Z28__device_stub__myFirstKernelPi .size _Z13myFirstKernelPi, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "kernel execution" .size .L.str, 17 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "hipMemcpy" .size .L.str.1, 10 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Cuda error: %s: %s.\n" .size .L.str.3, 21 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z13myFirstKernelPi" .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 "Correct!" .size .Lstr, 9 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z28__device_stub__myFirstKernelPi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z13myFirstKernelPi .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
code for sm_80 Function : add_n .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_TID.X ; /* 0x0000000000027919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x001fca00078e0003 */ /*0050*/ LDG.E R0, [R2.64] ; /* 0x0000000402007981 */ /* 0x000ea4000c1e1900 */ /*0060*/ IADD3 R5, R0, c[0x0][0x168], RZ ; /* 0x00005a0000057a10 */ /* 0x004fca0007ffe0ff */ /*0070*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0080*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0090*/ BRA 0x90; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
.file "tmpxft_00195b0a_00000000-6_addn.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z5add_nPiiiPiii .type _Z26__device_stub__Z5add_nPiiiPiii, @function _Z26__device_stub__Z5add_nPiiiPiii: .LFB2081: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq add_n(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z26__device_stub__Z5add_nPiiiPiii, .-_Z26__device_stub__Z5add_nPiiiPiii .globl add_n .type add_n, @function add_n: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z5add_nPiiiPiii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size add_n, .-add_n .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "add_n" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq add_n(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
#include <hip/hip_runtime.h> #include <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected add_n .globl add_n .p2align 8 .type add_n,@function add_n: s_load_b64 s[2:3], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_load_b32 s0, s[0:1], 0x8 s_waitcnt lgkmcnt(0) global_load_b32 v1, v0, s[2:3] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, s0, v1 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 add_n .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 16 .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 2 .amdhsa_next_free_sgpr 4 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size add_n, .Lfunc_end0-add_n .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 16 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: add_n .private_segment_fixed_size: 0 .sgpr_count: 4 .sgpr_spill_count: 0 .symbol: add_n.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 2 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> extern "C" __global__ void add_n(int *nums, int n, int size) { nums[threadIdx.x] += n; }
.text .file "addn.hip" .globl __device_stub__add_n # -- Begin function __device_stub__add_n .p2align 4, 0x90 .type __device_stub__add_n,@function __device_stub__add_n: # @__device_stub__add_n .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 $add_n, %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 __device_stub__add_n, .Lfunc_end0-__device_stub__add_n .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_n, %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_n,@object # @add_n .section .rodata,"a",@progbits .globl add_n .p2align 3, 0x0 add_n: .quad __device_stub__add_n .size add_n, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "add_n" .size .L__unnamed_1, 6 .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_n .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym add_n .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_n .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_TID.X ; /* 0x0000000000027919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x001fca00078e0003 */ /*0050*/ LDG.E R0, [R2.64] ; /* 0x0000000402007981 */ /* 0x000ea4000c1e1900 */ /*0060*/ IADD3 R5, R0, c[0x0][0x168], RZ ; /* 0x00005a0000057a10 */ /* 0x004fca0007ffe0ff */ /*0070*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 0x000fe2000c101904 */ /*0080*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0090*/ BRA 0x90; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*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 add_n .globl add_n .p2align 8 .type add_n,@function add_n: s_load_b64 s[2:3], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_load_b32 s0, s[0:1], 0x8 s_waitcnt lgkmcnt(0) global_load_b32 v1, v0, s[2:3] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, s0, v1 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 add_n .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 16 .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 2 .amdhsa_next_free_sgpr 4 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size add_n, .Lfunc_end0-add_n .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 16 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: add_n .private_segment_fixed_size: 0 .sgpr_count: 4 .sgpr_spill_count: 0 .symbol: add_n.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 2 .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_00195b0a_00000000-6_addn.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z5add_nPiiiPiii .type _Z26__device_stub__Z5add_nPiiiPiii, @function _Z26__device_stub__Z5add_nPiiiPiii: .LFB2081: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movl %esi, 4(%rsp) movl %edx, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) leaq 4(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq add_n(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z26__device_stub__Z5add_nPiiiPiii, .-_Z26__device_stub__Z5add_nPiiiPiii .globl add_n .type add_n, @function add_n: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z5add_nPiiiPiii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size add_n, .-add_n .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "add_n" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq add_n(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "addn.hip" .globl __device_stub__add_n # -- Begin function __device_stub__add_n .p2align 4, 0x90 .type __device_stub__add_n,@function __device_stub__add_n: # @__device_stub__add_n .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 $add_n, %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 __device_stub__add_n, .Lfunc_end0-__device_stub__add_n .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_n, %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_n,@object # @add_n .section .rodata,"a",@progbits .globl add_n .p2align 3, 0x0 add_n: .quad __device_stub__add_n .size add_n, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "add_n" .size .L__unnamed_1, 6 .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_n .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym add_n .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); cudaMalloc((void**)&d_a,size1); cudaMalloc((void**)&d_b,size2); cudaMalloc((void**)&d_t,size3); cudaMemcpy(d_a,a,size1,cudaMemcpyHostToDevice); cudaMemcpy(d_b,b,size2,cudaMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); cudaMemcpy(t,d_t,size3,cudaMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); cudaFree(d_a); cudaFree(d_t); return 0; }
code for sm_80 Function : _Z10mulElementPiS_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*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); cudaMalloc((void**)&d_a,size1); cudaMalloc((void**)&d_b,size2); cudaMalloc((void**)&d_t,size3); cudaMemcpy(d_a,a,size1,cudaMemcpyHostToDevice); cudaMemcpy(d_b,b,size2,cudaMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); cudaMemcpy(t,d_t,size3,cudaMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); cudaFree(d_a); cudaFree(d_t); return 0; }
.file "tmpxft_000450e8_00000000-6_first.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z6getTidv .type _Z6getTidv, @function _Z6getTidv: .LFB2057: .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 .LFE2057: .size _Z6getTidv, .-_Z6getTidv .globl _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii .type _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii, @function _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii: .LFB2083: .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 .L9 .L5: movq 168(%rsp), %rax subq %fs:40, %rax jne .L10 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .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 _Z10mulElementPiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L5 .L10: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii, .-_Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii .globl _Z10mulElementPiS_S_iii .type _Z10mulElementPiS_S_iii, @function _Z10mulElementPiS_S_iii: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z10mulElementPiS_S_iii, .-_Z10mulElementPiS_S_iii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Enter the dimensions of first matrix \n " .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "%d %d" .section .rodata.str1.8 .align 8 .LC2: .string "Enter the dimensions of second matrix \n" .section .rodata.str1.1 .LC3: .string "Enter input matrix 1 : \n" .LC4: .string "%d" .LC5: .string "Enter input matrix 2 : \n" .section .rodata.str1.8 .align 8 .LC6: .string "Enter the dimension of the grid \n" .align 8 .LC10: .string "The dimensions of block are : \n %d %d \n" .section .rodata.str1.1 .LC11: .string "Result vector is :\n" .LC12: .string "%d " .LC13: .string "\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $120, %rsp .cfi_def_cfa_offset 176 movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq .LC0(%rip), %rsi movl $2, %edi call __printf_chk@PLT leaq 36(%rsp), %rdx leaq 32(%rsp), %rsi leaq .LC1(%rip), %rbx movq %rbx, %rdi movl $0, %eax call __isoc23_scanf@PLT leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 44(%rsp), %rdx leaq 40(%rsp), %rsi movq %rbx, %rdi movl $0, %eax call __isoc23_scanf@PLT movl 32(%rsp), %ebx movl 36(%rsp), %eax movl %ebx, %r14d imull %eax, %r14d sall $2, %r14d movl 40(%rsp), %ebp movl 44(%rsp), %r12d movl %r12d, %r13d imull %ebp, %r13d sall $2, %r13d movl %ebx, %edx imull %r12d, %edx leal 0(,%rdx,4), %ecx movl %ecx, 20(%rsp) imull %ebx, %eax movslq %eax, %rdi salq $2, %rdi call malloc@PLT movq %rax, %r15 movq %rax, 24(%rsp) imull %r12d, %ebp movslq %ebp, %rdi salq $2, %rdi call malloc@PLT movq %rax, 8(%rsp) imull %r12d, %ebx movslq %ebx, %rdi salq $2, %rdi call malloc@PLT movq %rax, %rbx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 32(%rsp), %eax imull 36(%rsp), %eax testl %eax, %eax jle .L14 movq %r15, %r12 movl $0, %ebp leaq .LC4(%rip), %r15 .L15: movq %r12, %rsi movq %r15, %rdi movl $0, %eax call __isoc23_scanf@PLT addl $1, %ebp addq $4, %r12 movl 32(%rsp), %eax imull 36(%rsp), %eax cmpl %ebp, %eax jg .L15 .L14: leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 40(%rsp), %eax imull 44(%rsp), %eax testl %eax, %eax jle .L16 movq 8(%rsp), %r12 movl $0, %ebp leaq .LC4(%rip), %r15 .L17: movq %r12, %rsi movq %r15, %rdi movl $0, %eax call __isoc23_scanf@PLT addl $1, %ebp addq $4, %r12 movl 40(%rsp), %eax imull 44(%rsp), %eax cmpl %ebp, %eax jg .L17 .L16: movslq %r14d, %r14 leaq 56(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movslq %r13d, %r13 leaq 64(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT movslq 20(%rsp), %rbp leaq 72(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r14, %rdx movq 24(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %r13, %rdx movq 8(%rsp), %rsi movq 64(%rsp), %rdi call cudaMemcpy@PLT leaq .LC6(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 52(%rsp), %rdx leaq 48(%rsp), %rsi leaq .LC1(%rip), %rdi movl $0, %eax call __isoc23_scanf@PLT pxor %xmm0, %xmm0 cvtsi2sdl 32(%rsp), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdl 48(%rsp), %xmm1 divsd %xmm1, %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L18 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L18: cvttsd2sil %xmm3, %r12d pxor %xmm0, %xmm0 cvtsi2sdl 44(%rsp), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdl 52(%rsp), %xmm1 divsd %xmm1, %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L19 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L19: cvttsd2sil %xmm3, %r13d movl %r13d, %ecx movl %r12d, %edx leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 48(%rsp), %eax movl %eax, 80(%rsp) movl 52(%rsp), %eax movl %eax, 84(%rsp) movl %r12d, 92(%rsp) movl %r13d, 96(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L34 .L20: movl $2, %ecx movq %rbp, %rdx movq 72(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %r12d leaq .LC12(%rip), %r13 leaq .LC13(%rip), %r14 cmpl $0, 32(%rsp) jg .L21 .L22: movq stdin(%rip), %rdi call getc@PLT movq 56(%rsp), %rdi call cudaFree@PLT movq 72(%rsp), %rdi call cudaFree@PLT movq 104(%rsp), %rax subq %fs:40, %rax jne .L35 movl $0, %eax addq $120, %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 .L34: .cfi_restore_state movl 36(%rsp), %r9d movl 44(%rsp), %r8d movl 32(%rsp), %ecx movq 72(%rsp), %rdx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii jmp .L20 .L23: imull %r12d, %eax addl %ebp, %eax cltq movl (%rbx,%rax,4), %edx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %ebp movl 44(%rsp), %eax cmpl %ebp, %eax jg .L23 .L24: movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %r12d cmpl %r12d, 32(%rsp) jle .L22 .L21: movl 44(%rsp), %eax movl $0, %ebp testl %eax, %eax jg .L23 jmp .L24 .L35: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC15: .string "_Z10mulElementPiS_S_iii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _Z10mulElementPiS_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 .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC7: .long 0 .long 1127219200 .align 8 .LC9: .long 0 .long 1072693248 .align 8 .LC14: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); cudaMalloc((void**)&d_a,size1); cudaMalloc((void**)&d_b,size2); cudaMalloc((void**)&d_t,size3); cudaMemcpy(d_a,a,size1,cudaMemcpyHostToDevice); cudaMemcpy(d_b,b,size2,cudaMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); cudaMemcpy(t,d_t,size3,cudaMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); cudaFree(d_a); cudaFree(d_t); return 0; }
#include "hip/hip_runtime.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); hipMalloc((void**)&d_a,size1); hipMalloc((void**)&d_b,size2); hipMalloc((void**)&d_t,size3); hipMemcpy(d_a,a,size1,hipMemcpyHostToDevice); hipMemcpy(d_b,b,size2,hipMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); hipMemcpy(t,d_t,size3,hipMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); hipFree(d_a); hipFree(d_t); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include "hip/hip_runtime.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); hipMalloc((void**)&d_a,size1); hipMalloc((void**)&d_b,size2); hipMalloc((void**)&d_t,size3); hipMemcpy(d_a,a,size1,hipMemcpyHostToDevice); hipMemcpy(d_b,b,size2,hipMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); hipMemcpy(t,d_t,size3,hipMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); hipFree(d_a); hipFree(d_t); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10mulElementPiS_S_iii .globl _Z10mulElementPiS_S_iii .p2align 8 .type _Z10mulElementPiS_S_iii,@function _Z10mulElementPiS_S_iii: .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z10mulElementPiS_S_iii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 36 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10mulElementPiS_S_iii, .Lfunc_end0-_Z10mulElementPiS_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: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 36 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z10mulElementPiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z10mulElementPiS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include "hip/hip_runtime.h" #include <stdio.h> #include <stdlib.h> #include <math.h> __device__ int getTid() { int bid = blockIdx.y * gridDim.x + blockIdx.x; int tid = threadIdx.y * blockDim.x + threadIdx.x; int tPB = blockDim.x * blockDim.y ; int fin = bid*tPB+tid; } __global__ void mulElement(int *a ,int *b , int *c , int ha , int wb,int wa) { int th = getTid(); if(th<(ha*wb)) { int row = th/wb; int col = th%wb; int i = 0 , sum = 0; for(i = 0;i<wa;i++) { sum += a[row*wa+i]*b[wb*i+col]; } c[th] = sum; } } int main(void) { int *a,*b,*t,i,j; int *d_a,*d_b,*d_t; int ha , wa; int hb , wb; printf("Enter the dimensions of first matrix \n "); scanf("%d %d",&ha,&wa); printf("Enter the dimensions of second matrix \n"); scanf("%d %d",&hb,&wb); int size1 = sizeof(int)*ha*wa; int size2 = sizeof(int)*hb*wb; int size3 = sizeof(int)*ha*wb; a = (int*)malloc(ha*wa*sizeof(int)); b = (int*)malloc(hb*wb*sizeof(int)); t = (int*)malloc(ha*wb*sizeof(int)); printf("Enter input matrix 1 : \n"); for(i = 0;i<ha*wa;i++) scanf("%d",&a[i]); printf("Enter input matrix 2 : \n"); for(i = 0;i<hb*wb;i++) scanf("%d",&b[i]); hipMalloc((void**)&d_a,size1); hipMalloc((void**)&d_b,size2); hipMalloc((void**)&d_t,size3); hipMemcpy(d_a,a,size1,hipMemcpyHostToDevice); hipMemcpy(d_b,b,size2,hipMemcpyHostToDevice); int gx,gy,bx,by; printf("Enter the dimension of the grid \n"); scanf("%d %d",&gx,&gy); bx = ceil((double)ha/gx); by = ceil((double)wb/gy); printf("The dimensions of block are : \n %d %d \n",bx,by); dim3 grid(gx,gy); dim3 block(bx,by); mulElement<<<grid,block>>>(d_a,d_b,d_t,ha,wb,wa); hipMemcpy(t,d_t,size3,hipMemcpyDeviceToHost); printf("Result vector is :\n"); for(i = 0;i<ha;i++) { for(j = 0;j<wb;j++) printf("%d ",t[i*wb+j]); printf("\n"); } getchar(); hipFree(d_a); hipFree(d_t); return 0; }
.text .file "first.hip" .globl _Z25__device_stub__mulElementPiS_S_iii # -- Begin function _Z25__device_stub__mulElementPiS_S_iii .p2align 4, 0x90 .type _Z25__device_stub__mulElementPiS_S_iii,@function _Z25__device_stub__mulElementPiS_S_iii: # @_Z25__device_stub__mulElementPiS_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 $_Z10mulElementPiS_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 _Z25__device_stub__mulElementPiS_S_iii, .Lfunc_end0-_Z25__device_stub__mulElementPiS_S_iii .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $216, %rsp .cfi_def_cfa_offset 272 .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 $.L.str, %edi xorl %eax, %eax callq printf leaq 12(%rsp), %rsi leaq 16(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf movl $.Lstr, %edi callq puts@PLT leaq 20(%rsp), %rsi leaq 8(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf movl 12(%rsp), %ebx leal (,%rbx,4), %eax movl %eax, 48(%rsp) # 4-byte Spill movl 16(%rsp), %r13d movl 8(%rsp), %r15d movl 20(%rsp), %ebp imull %r15d, %ebp movl %r13d, %eax imull %ebx, %eax movslq %eax, %rdi shlq $2, %rdi callq malloc movq %rax, 72(%rsp) # 8-byte Spill movslq %ebp, %rdi shlq $2, %rdi callq malloc movq %rax, 64(%rsp) # 8-byte Spill imull %r15d, %ebx movslq %ebx, %rdi shlq $2, %rdi callq malloc movq %rax, %rbx movl $.Lstr.1, %edi callq puts@PLT movl 16(%rsp), %eax imull 12(%rsp), %eax testl %eax, %eax jle .LBB1_3 # %bb.1: # %.lr.ph.preheader movq 72(%rsp), %r14 # 8-byte Reload xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq __isoc23_scanf incq %r12 movslq 12(%rsp), %rax movslq 16(%rsp), %rcx imulq %rax, %rcx addq $4, %r14 cmpq %rcx, %r12 jl .LBB1_2 .LBB1_3: # %._crit_edge movl 48(%rsp), %eax # 4-byte Reload imull %eax, %r13d shll $2, %ebp imull %eax, %r15d movl $.Lstr.2, %edi callq puts@PLT movl 8(%rsp), %eax imull 20(%rsp), %eax testl %eax, %eax jle .LBB1_6 # %bb.4: # %.lr.ph44.preheader movq 64(%rsp), %r14 # 8-byte Reload xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_5: # %.lr.ph44 # =>This Inner Loop Header: Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq __isoc23_scanf incq %r12 movslq 20(%rsp), %rax movslq 8(%rsp), %rcx imulq %rax, %rcx addq $4, %r14 cmpq %rcx, %r12 jl .LBB1_5 .LBB1_6: # %._crit_edge45 movslq %r13d, %r13 leaq 40(%rsp), %rdi movq %r13, %rsi callq hipMalloc movslq %ebp, %rbp leaq 80(%rsp), %rdi movq %rbp, %rsi callq hipMalloc movslq %r15d, %r14 leaq 32(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 40(%rsp), %rdi movq 72(%rsp), %rsi # 8-byte Reload movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 80(%rsp), %rdi movq 64(%rsp), %rsi # 8-byte Reload movq %rbp, %rdx movl $1, %ecx callq hipMemcpy movl $.Lstr.3, %edi callq puts@PLT leaq 28(%rsp), %rsi leaq 24(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf cvtsi2sdl 12(%rsp), %xmm0 cvtsi2sdl 28(%rsp), %xmm1 divsd %xmm1, %xmm0 callq ceil@PLT cvttsd2si %xmm0, %r15d xorps %xmm0, %xmm0 cvtsi2sdl 8(%rsp), %xmm0 xorps %xmm1, %xmm1 cvtsi2sdl 24(%rsp), %xmm1 divsd %xmm1, %xmm0 callq ceil@PLT cvttsd2si %xmm0, %r12d movl $.L.str.7, %edi movl %r15d, %esi movl %r12d, %edx xorl %eax, %eax callq printf movl 28(%rsp), %eax movl 24(%rsp), %edi shlq $32, %rdi orq %rax, %rdi shlq $32, %r12 orq %r15, %r12 movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_8 # %bb.7: movq 40(%rsp), %rax movq 80(%rsp), %rcx movq 32(%rsp), %rdx movl 12(%rsp), %esi movl 8(%rsp), %edi movl 16(%rsp), %r8d movq %rax, 152(%rsp) movq %rcx, 144(%rsp) movq %rdx, 136(%rsp) movl %esi, 60(%rsp) movl %edi, 56(%rsp) movl %r8d, 52(%rsp) leaq 152(%rsp), %rax movq %rax, 160(%rsp) leaq 144(%rsp), %rax movq %rax, 168(%rsp) leaq 136(%rsp), %rax movq %rax, 176(%rsp) leaq 60(%rsp), %rax movq %rax, 184(%rsp) leaq 56(%rsp), %rax movq %rax, 192(%rsp) leaq 52(%rsp), %rax movq %rax, 200(%rsp) leaq 120(%rsp), %rdi leaq 104(%rsp), %rsi leaq 96(%rsp), %rdx leaq 88(%rsp), %rcx callq __hipPopCallConfiguration movq 120(%rsp), %rsi movl 128(%rsp), %edx movq 104(%rsp), %rcx movl 112(%rsp), %r8d leaq 160(%rsp), %r9 movl $_Z10mulElementPiS_S_iii, %edi pushq 88(%rsp) .cfi_adjust_cfa_offset 8 pushq 104(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_8: movq 32(%rsp), %rsi movq %rbx, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy movl $.Lstr.4, %edi callq puts@PLT cmpl $0, 12(%rsp) jle .LBB1_14 # %bb.9: # %.preheader.preheader xorl %ebp, %ebp jmp .LBB1_10 .p2align 4, 0x90 .LBB1_13: # %._crit_edge48 # in Loop: Header=BB1_10 Depth=1 movl $10, %edi callq putchar@PLT incl %ebp cmpl 12(%rsp), %ebp jge .LBB1_14 .LBB1_10: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_12 Depth 2 movl 8(%rsp), %eax testl %eax, %eax jle .LBB1_13 # %bb.11: # %.lr.ph47.preheader # in Loop: Header=BB1_10 Depth=1 xorl %r14d, %r14d .p2align 4, 0x90 .LBB1_12: # %.lr.ph47 # Parent Loop BB1_10 Depth=1 # => This Inner Loop Header: Depth=2 imull %ebp, %eax cltq addq %r14, %rax movl (%rbx,%rax,4), %esi movl $.L.str.9, %edi xorl %eax, %eax callq printf movl 8(%rsp), %eax incq %r14 cmpl %eax, %r14d jl .LBB1_12 jmp .LBB1_13 .LBB1_14: # %._crit_edge50 movq stdin(%rip), %rdi callq getc movq 40(%rsp), %rdi callq hipFree movq 32(%rsp), %rdi callq hipFree xorl %eax, %eax addq $216, %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 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 $_Z10mulElementPiS_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_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 _Z10mulElementPiS_S_iii,@object # @_Z10mulElementPiS_S_iii .section .rodata,"a",@progbits .globl _Z10mulElementPiS_S_iii .p2align 3, 0x0 _Z10mulElementPiS_S_iii: .quad _Z25__device_stub__mulElementPiS_S_iii .size _Z10mulElementPiS_S_iii, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Enter the dimensions of first matrix \n " .size .L.str, 40 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d %d" .size .L.str.1, 6 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%d" .size .L.str.4, 3 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "The dimensions of block are : \n %d %d \n" .size .L.str.7, 40 .type .L.str.9,@object # @.str.9 .L.str.9: .asciz "%d " .size .L.str.9, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10mulElementPiS_S_iii" .size .L__unnamed_1, 24 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Enter the dimensions of second matrix " .size .Lstr, 39 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Enter input matrix 1 : " .size .Lstr.1, 24 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Enter input matrix 2 : " .size .Lstr.2, 24 .type .Lstr.3,@object # @str.3 .Lstr.3: .asciz "Enter the dimension of the grid " .size .Lstr.3, 33 .type .Lstr.4,@object # @str.4 .Lstr.4: .asciz "Result vector is :" .size .Lstr.4, 19 .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__mulElementPiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10mulElementPiS_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 : _Z10mulElementPiS_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*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10mulElementPiS_S_iii .globl _Z10mulElementPiS_S_iii .p2align 8 .type _Z10mulElementPiS_S_iii,@function _Z10mulElementPiS_S_iii: .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z10mulElementPiS_S_iii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 36 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10mulElementPiS_S_iii, .Lfunc_end0-_Z10mulElementPiS_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: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 36 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z10mulElementPiS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z10mulElementPiS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000450e8_00000000-6_first.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z6getTidv .type _Z6getTidv, @function _Z6getTidv: .LFB2057: .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 .LFE2057: .size _Z6getTidv, .-_Z6getTidv .globl _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii .type _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii, @function _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii: .LFB2083: .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 .L9 .L5: movq 168(%rsp), %rax subq %fs:40, %rax jne .L10 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .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 _Z10mulElementPiS_S_iii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 192 jmp .L5 .L10: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii, .-_Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii .globl _Z10mulElementPiS_S_iii .type _Z10mulElementPiS_S_iii, @function _Z10mulElementPiS_S_iii: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z10mulElementPiS_S_iii, .-_Z10mulElementPiS_S_iii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Enter the dimensions of first matrix \n " .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "%d %d" .section .rodata.str1.8 .align 8 .LC2: .string "Enter the dimensions of second matrix \n" .section .rodata.str1.1 .LC3: .string "Enter input matrix 1 : \n" .LC4: .string "%d" .LC5: .string "Enter input matrix 2 : \n" .section .rodata.str1.8 .align 8 .LC6: .string "Enter the dimension of the grid \n" .align 8 .LC10: .string "The dimensions of block are : \n %d %d \n" .section .rodata.str1.1 .LC11: .string "Result vector is :\n" .LC12: .string "%d " .LC13: .string "\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $120, %rsp .cfi_def_cfa_offset 176 movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq .LC0(%rip), %rsi movl $2, %edi call __printf_chk@PLT leaq 36(%rsp), %rdx leaq 32(%rsp), %rsi leaq .LC1(%rip), %rbx movq %rbx, %rdi movl $0, %eax call __isoc23_scanf@PLT leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 44(%rsp), %rdx leaq 40(%rsp), %rsi movq %rbx, %rdi movl $0, %eax call __isoc23_scanf@PLT movl 32(%rsp), %ebx movl 36(%rsp), %eax movl %ebx, %r14d imull %eax, %r14d sall $2, %r14d movl 40(%rsp), %ebp movl 44(%rsp), %r12d movl %r12d, %r13d imull %ebp, %r13d sall $2, %r13d movl %ebx, %edx imull %r12d, %edx leal 0(,%rdx,4), %ecx movl %ecx, 20(%rsp) imull %ebx, %eax movslq %eax, %rdi salq $2, %rdi call malloc@PLT movq %rax, %r15 movq %rax, 24(%rsp) imull %r12d, %ebp movslq %ebp, %rdi salq $2, %rdi call malloc@PLT movq %rax, 8(%rsp) imull %r12d, %ebx movslq %ebx, %rdi salq $2, %rdi call malloc@PLT movq %rax, %rbx leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 32(%rsp), %eax imull 36(%rsp), %eax testl %eax, %eax jle .L14 movq %r15, %r12 movl $0, %ebp leaq .LC4(%rip), %r15 .L15: movq %r12, %rsi movq %r15, %rdi movl $0, %eax call __isoc23_scanf@PLT addl $1, %ebp addq $4, %r12 movl 32(%rsp), %eax imull 36(%rsp), %eax cmpl %ebp, %eax jg .L15 .L14: leaq .LC5(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 40(%rsp), %eax imull 44(%rsp), %eax testl %eax, %eax jle .L16 movq 8(%rsp), %r12 movl $0, %ebp leaq .LC4(%rip), %r15 .L17: movq %r12, %rsi movq %r15, %rdi movl $0, %eax call __isoc23_scanf@PLT addl $1, %ebp addq $4, %r12 movl 40(%rsp), %eax imull 44(%rsp), %eax cmpl %ebp, %eax jg .L17 .L16: movslq %r14d, %r14 leaq 56(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movslq %r13d, %r13 leaq 64(%rsp), %rdi movq %r13, %rsi call cudaMalloc@PLT movslq 20(%rsp), %rbp leaq 72(%rsp), %rdi movq %rbp, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r14, %rdx movq 24(%rsp), %rsi movq 56(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %r13, %rdx movq 8(%rsp), %rsi movq 64(%rsp), %rdi call cudaMemcpy@PLT leaq .LC6(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq 52(%rsp), %rdx leaq 48(%rsp), %rsi leaq .LC1(%rip), %rdi movl $0, %eax call __isoc23_scanf@PLT pxor %xmm0, %xmm0 cvtsi2sdl 32(%rsp), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdl 48(%rsp), %xmm1 divsd %xmm1, %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L18 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L18: cvttsd2sil %xmm3, %r12d pxor %xmm0, %xmm0 cvtsi2sdl 44(%rsp), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdl 52(%rsp), %xmm1 divsd %xmm1, %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L19 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L19: cvttsd2sil %xmm3, %r13d movl %r13d, %ecx movl %r12d, %edx leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 48(%rsp), %eax movl %eax, 80(%rsp) movl 52(%rsp), %eax movl %eax, 84(%rsp) movl %r12d, 92(%rsp) movl %r13d, 96(%rsp) movl $0, %r9d movl $0, %r8d movq 92(%rsp), %rdx movl $1, %ecx movq 80(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L34 .L20: movl $2, %ecx movq %rbp, %rdx movq 72(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %r12d leaq .LC12(%rip), %r13 leaq .LC13(%rip), %r14 cmpl $0, 32(%rsp) jg .L21 .L22: movq stdin(%rip), %rdi call getc@PLT movq 56(%rsp), %rdi call cudaFree@PLT movq 72(%rsp), %rdi call cudaFree@PLT movq 104(%rsp), %rax subq %fs:40, %rax jne .L35 movl $0, %eax addq $120, %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 .L34: .cfi_restore_state movl 36(%rsp), %r9d movl 44(%rsp), %r8d movl 32(%rsp), %ecx movq 72(%rsp), %rdx movq 64(%rsp), %rsi movq 56(%rsp), %rdi call _Z37__device_stub__Z10mulElementPiS_S_iiiPiS_S_iii jmp .L20 .L23: imull %r12d, %eax addl %ebp, %eax cltq movl (%rbx,%rax,4), %edx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %ebp movl 44(%rsp), %eax cmpl %ebp, %eax jg .L23 .L24: movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %r12d cmpl %r12d, 32(%rsp) jle .L22 .L21: movl 44(%rsp), %eax movl $0, %ebp testl %eax, %eax jg .L23 jmp .L24 .L35: call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC15: .string "_Z10mulElementPiS_S_iii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _Z10mulElementPiS_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 .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC7: .long 0 .long 1127219200 .align 8 .LC9: .long 0 .long 1072693248 .align 8 .LC14: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "first.hip" .globl _Z25__device_stub__mulElementPiS_S_iii # -- Begin function _Z25__device_stub__mulElementPiS_S_iii .p2align 4, 0x90 .type _Z25__device_stub__mulElementPiS_S_iii,@function _Z25__device_stub__mulElementPiS_S_iii: # @_Z25__device_stub__mulElementPiS_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 $_Z10mulElementPiS_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 _Z25__device_stub__mulElementPiS_S_iii, .Lfunc_end0-_Z25__device_stub__mulElementPiS_S_iii .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $216, %rsp .cfi_def_cfa_offset 272 .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 $.L.str, %edi xorl %eax, %eax callq printf leaq 12(%rsp), %rsi leaq 16(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf movl $.Lstr, %edi callq puts@PLT leaq 20(%rsp), %rsi leaq 8(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf movl 12(%rsp), %ebx leal (,%rbx,4), %eax movl %eax, 48(%rsp) # 4-byte Spill movl 16(%rsp), %r13d movl 8(%rsp), %r15d movl 20(%rsp), %ebp imull %r15d, %ebp movl %r13d, %eax imull %ebx, %eax movslq %eax, %rdi shlq $2, %rdi callq malloc movq %rax, 72(%rsp) # 8-byte Spill movslq %ebp, %rdi shlq $2, %rdi callq malloc movq %rax, 64(%rsp) # 8-byte Spill imull %r15d, %ebx movslq %ebx, %rdi shlq $2, %rdi callq malloc movq %rax, %rbx movl $.Lstr.1, %edi callq puts@PLT movl 16(%rsp), %eax imull 12(%rsp), %eax testl %eax, %eax jle .LBB1_3 # %bb.1: # %.lr.ph.preheader movq 72(%rsp), %r14 # 8-byte Reload xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq __isoc23_scanf incq %r12 movslq 12(%rsp), %rax movslq 16(%rsp), %rcx imulq %rax, %rcx addq $4, %r14 cmpq %rcx, %r12 jl .LBB1_2 .LBB1_3: # %._crit_edge movl 48(%rsp), %eax # 4-byte Reload imull %eax, %r13d shll $2, %ebp imull %eax, %r15d movl $.Lstr.2, %edi callq puts@PLT movl 8(%rsp), %eax imull 20(%rsp), %eax testl %eax, %eax jle .LBB1_6 # %bb.4: # %.lr.ph44.preheader movq 64(%rsp), %r14 # 8-byte Reload xorl %r12d, %r12d .p2align 4, 0x90 .LBB1_5: # %.lr.ph44 # =>This Inner Loop Header: Depth=1 movl $.L.str.4, %edi movq %r14, %rsi xorl %eax, %eax callq __isoc23_scanf incq %r12 movslq 20(%rsp), %rax movslq 8(%rsp), %rcx imulq %rax, %rcx addq $4, %r14 cmpq %rcx, %r12 jl .LBB1_5 .LBB1_6: # %._crit_edge45 movslq %r13d, %r13 leaq 40(%rsp), %rdi movq %r13, %rsi callq hipMalloc movslq %ebp, %rbp leaq 80(%rsp), %rdi movq %rbp, %rsi callq hipMalloc movslq %r15d, %r14 leaq 32(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 40(%rsp), %rdi movq 72(%rsp), %rsi # 8-byte Reload movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 80(%rsp), %rdi movq 64(%rsp), %rsi # 8-byte Reload movq %rbp, %rdx movl $1, %ecx callq hipMemcpy movl $.Lstr.3, %edi callq puts@PLT leaq 28(%rsp), %rsi leaq 24(%rsp), %rdx movl $.L.str.1, %edi xorl %eax, %eax callq __isoc23_scanf cvtsi2sdl 12(%rsp), %xmm0 cvtsi2sdl 28(%rsp), %xmm1 divsd %xmm1, %xmm0 callq ceil@PLT cvttsd2si %xmm0, %r15d xorps %xmm0, %xmm0 cvtsi2sdl 8(%rsp), %xmm0 xorps %xmm1, %xmm1 cvtsi2sdl 24(%rsp), %xmm1 divsd %xmm1, %xmm0 callq ceil@PLT cvttsd2si %xmm0, %r12d movl $.L.str.7, %edi movl %r15d, %esi movl %r12d, %edx xorl %eax, %eax callq printf movl 28(%rsp), %eax movl 24(%rsp), %edi shlq $32, %rdi orq %rax, %rdi shlq $32, %r12 orq %r15, %r12 movl $1, %esi movq %r12, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_8 # %bb.7: movq 40(%rsp), %rax movq 80(%rsp), %rcx movq 32(%rsp), %rdx movl 12(%rsp), %esi movl 8(%rsp), %edi movl 16(%rsp), %r8d movq %rax, 152(%rsp) movq %rcx, 144(%rsp) movq %rdx, 136(%rsp) movl %esi, 60(%rsp) movl %edi, 56(%rsp) movl %r8d, 52(%rsp) leaq 152(%rsp), %rax movq %rax, 160(%rsp) leaq 144(%rsp), %rax movq %rax, 168(%rsp) leaq 136(%rsp), %rax movq %rax, 176(%rsp) leaq 60(%rsp), %rax movq %rax, 184(%rsp) leaq 56(%rsp), %rax movq %rax, 192(%rsp) leaq 52(%rsp), %rax movq %rax, 200(%rsp) leaq 120(%rsp), %rdi leaq 104(%rsp), %rsi leaq 96(%rsp), %rdx leaq 88(%rsp), %rcx callq __hipPopCallConfiguration movq 120(%rsp), %rsi movl 128(%rsp), %edx movq 104(%rsp), %rcx movl 112(%rsp), %r8d leaq 160(%rsp), %r9 movl $_Z10mulElementPiS_S_iii, %edi pushq 88(%rsp) .cfi_adjust_cfa_offset 8 pushq 104(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_8: movq 32(%rsp), %rsi movq %rbx, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy movl $.Lstr.4, %edi callq puts@PLT cmpl $0, 12(%rsp) jle .LBB1_14 # %bb.9: # %.preheader.preheader xorl %ebp, %ebp jmp .LBB1_10 .p2align 4, 0x90 .LBB1_13: # %._crit_edge48 # in Loop: Header=BB1_10 Depth=1 movl $10, %edi callq putchar@PLT incl %ebp cmpl 12(%rsp), %ebp jge .LBB1_14 .LBB1_10: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_12 Depth 2 movl 8(%rsp), %eax testl %eax, %eax jle .LBB1_13 # %bb.11: # %.lr.ph47.preheader # in Loop: Header=BB1_10 Depth=1 xorl %r14d, %r14d .p2align 4, 0x90 .LBB1_12: # %.lr.ph47 # Parent Loop BB1_10 Depth=1 # => This Inner Loop Header: Depth=2 imull %ebp, %eax cltq addq %r14, %rax movl (%rbx,%rax,4), %esi movl $.L.str.9, %edi xorl %eax, %eax callq printf movl 8(%rsp), %eax incq %r14 cmpl %eax, %r14d jl .LBB1_12 jmp .LBB1_13 .LBB1_14: # %._crit_edge50 movq stdin(%rip), %rdi callq getc movq 40(%rsp), %rdi callq hipFree movq 32(%rsp), %rdi callq hipFree xorl %eax, %eax addq $216, %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 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 $_Z10mulElementPiS_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_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 _Z10mulElementPiS_S_iii,@object # @_Z10mulElementPiS_S_iii .section .rodata,"a",@progbits .globl _Z10mulElementPiS_S_iii .p2align 3, 0x0 _Z10mulElementPiS_S_iii: .quad _Z25__device_stub__mulElementPiS_S_iii .size _Z10mulElementPiS_S_iii, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Enter the dimensions of first matrix \n " .size .L.str, 40 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d %d" .size .L.str.1, 6 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%d" .size .L.str.4, 3 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "The dimensions of block are : \n %d %d \n" .size .L.str.7, 40 .type .L.str.9,@object # @.str.9 .L.str.9: .asciz "%d " .size .L.str.9, 4 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z10mulElementPiS_S_iii" .size .L__unnamed_1, 24 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Enter the dimensions of second matrix " .size .Lstr, 39 .type .Lstr.1,@object # @str.1 .Lstr.1: .asciz "Enter input matrix 1 : " .size .Lstr.1, 24 .type .Lstr.2,@object # @str.2 .Lstr.2: .asciz "Enter input matrix 2 : " .size .Lstr.2, 24 .type .Lstr.3,@object # @str.3 .Lstr.3: .asciz "Enter the dimension of the grid " .size .Lstr.3, 33 .type .Lstr.4,@object # @str.4 .Lstr.4: .asciz "Result vector is :" .size .Lstr.4, 19 .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__mulElementPiS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10mulElementPiS_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.
#include <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; cudaMalloc(&d_total, sizeof(unsigned int) * numBlocks); cudaMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; cudaMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); cudaMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); cudaFree(d_total_scanned); } cudaFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; cudaMalloc(&d_in, memsize); cudaMalloc(&d_out, memsize); cudaMemcpy(d_in, h_in, memsize, cudaMemcpyHostToDevice); totalScan(d_in, d_out, n); cudaMemcpy(h_out, d_out, memsize, cudaMemcpyDeviceToHost); cudaFree(d_in); cudaFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
code for sm_80 Function : _Z7mapScanPjS_m .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e280000002100 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0030*/ IMAD R0, R3, 0x4, R0 ; /* 0x0000000403007824 */ /* 0x001fca00078e0200 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe40003f06070 */ /*0050*/ SHF.R.S32.HI R5, RZ, 0x1f, R0 ; /* 0x0000001fff057819 */ /* 0x000fc80000011400 */ /*0060*/ ISETP.GE.U32.AND.EX P0, PT, R5, c[0x0][0x174], PT, P0 ; /* 0x00005d0005007a0c */ /* 0x000fda0003f06100 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fe200078e00ff */ /*0090*/ LEA R4, P0, R0.reuse, c[0x0][0x160], 0x2 ; /* 0x0000580000047a11 */ /* 0x040fe200078010ff */ /*00a0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*00b0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x168] ; /* 0x00005a0003027625 */ /* 0x000fe200078e0202 */ /*00c0*/ LEA.HI.X R5, R0, c[0x0][0x164], R5, 0x2, P0 ; /* 0x0000590000057a11 */ /* 0x000fca00000f1405 */ /*00d0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea8000c1e1900 */ /*00e0*/ LDG.E R0, [R4.64] ; /* 0x0000000404007981 */ /* 0x000ea4000c1e1900 */ /*00f0*/ IMAD.IADD R7, R0, 0x1, R3 ; /* 0x0000000100077824 */ /* 0x004fca00078e0203 */ /*0100*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*0110*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0120*/ BRA 0x120; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z11partialScanPjS_S_m .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_CTAID.X ; /* 0x0000000000087919 */ /* 0x000e220000002500 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ BSSY B0, 0x110 ; /* 0x000000d000007945 */ /* 0x000fe40003800000 */ /*0040*/ S2R R13, SR_TID.X ; /* 0x00000000000d7919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R0, R8, 0x4, R13 ; /* 0x0000000408007824 */ /* 0x001fe200078e020d */ /*0060*/ ISETP.GT.AND P0, PT, R13, 0x2, PT ; /* 0x000000020d00780c */ /* 0x000fc80003f04270 */ /*0070*/ ISETP.GE.U32.AND P1, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe40003f26070 */ /*0080*/ SHF.R.S32.HI R3, RZ, 0x1f, R0 ; /* 0x0000001fff037819 */ /* 0x000fc80000011400 */ /*0090*/ ISETP.GE.U32.AND.EX P1, PT, R3, c[0x0][0x17c], PT, P1 ; /* 0x00005f0003007a0c */ /* 0x000fda0003f26110 */ /*00a0*/ @P1 STS [R13.X4], RZ ; /* 0x000000ff0d001388 */ /* 0x0001e20000004800 */ /*00b0*/ @P1 BRA 0x100 ; /* 0x0000004000001947 */ /* 0x000fea0003800000 */ /*00c0*/ LEA R2, P1, R0, c[0x0][0x160], 0x2 ; /* 0x0000580000027a11 */ /* 0x000fc800078210ff */ /*00d0*/ LEA.HI.X R3, R0, c[0x0][0x164], R3, 0x2, P1 ; /* 0x0000590000037a11 */ /* 0x000fca00008f1403 */ /*00e0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00f0*/ STS [R13.X4], R2 ; /* 0x000000020d007388 */ /* 0x0043e40000004800 */ /*0100*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0110*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0120*/ ISETP.GT.AND P1, PT, R13, 0x1, PT ; /* 0x000000010d00780c */ /* 0x000fe20003f24270 */ /*0130*/ IMAD.MOV.U32 R15, RZ, RZ, 0x4 ; /* 0x00000004ff0f7424 */ /* 0x000fe200078e00ff */ /*0140*/ IADD3 R5, R0, 0x1, RZ ; /* 0x0000000100057810 */ /* 0x000fc80007ffe0ff */ /*0150*/ SHF.R.S32.HI R10, RZ, 0x1f, R5 ; /* 0x0000001fff0a7819 */ /* 0x000fe20000011405 */ /*0160*/ @!P0 LDS R2, [R13.X4] ; /* 0x000000000d028984 */ /* 0x002fe80000004800 */ /*0170*/ @!P0 LDS R3, [R13.X4+0x4] ; /* 0x000004000d038984 */ /* 0x000e640000004800 */ /*0180*/ @!P0 IMAD.IADD R2, R2, 0x1, R3 ; /* 0x0000000102028824 */ /* 0x002fca00078e0203 */ /*0190*/ @!P0 STS [R13.X4+0x4], R2 ; /* 0x000004020d008388 */ /* 0x0003e80000004800 */ /*01a0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*01b0*/ ISETP.GE.U32.AND P0, PT, R5, c[0x0][0x178], PT ; /* 0x00005e0005007a0c */ /* 0x000fc80003f06070 */ /*01c0*/ ISETP.GE.U32.AND.EX P0, PT, R10, c[0x0][0x17c], PT, P0 ; /* 0x00005f000a007a0c */ /* 0x000fe20003f06100 */ /*01d0*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff027624 */ /* 0x002fc600078e00ff */ /*01e0*/ ISETP.GT.OR P0, PT, R13, 0x2, P0 ; /* 0x000000020d00780c */ /* 0x000fe20000704670 */ /*01f0*/ @!P1 LDS R3, [R13.X4] ; /* 0x000000000d039984 */ /* 0x000fe80000004800 */ /*0200*/ @!P1 LDS R4, [R13.X4+0x8] ; /* 0x000008000d049984 */ /* 0x000e640000004800 */ /*0210*/ @!P1 IMAD.IADD R0, R3, 0x1, R4 ; /* 0x0000000103009824 */ /* 0x002fe400078e0204 */ /*0220*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x16c] ; /* 0x00005b00ff037624 */ /* 0x000fc600078e00ff */ /*0230*/ @!P1 STS [R13.X4+0x8], R0 ; /* 0x000008000d009388 */ /* 0x000fe80000004800 */ /*0240*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0250*/ @!P0 LEA R6, P1, R5, c[0x0][0x168], 0x2 ; /* 0x00005a0005068a11 */ /* 0x000fc800078210ff */ /*0260*/ @!P0 LEA.HI.X R7, R5, c[0x0][0x16c], R10, 0x2, P1 ; /* 0x00005b0005078a11 */ /* 0x000fe200008f140a */ /*0270*/ IMAD.WIDE R4, R8, R15, c[0x0][0x170] ; /* 0x00005c0008047625 */ /* 0x000fe200078e020f */ /*0280*/ @!P0 LDS R9, [R13.X4] ; /* 0x000000000d098984 */ /* 0x000e680000004800 */ /*0290*/ LDS R11, [0xc] ; /* 0x00000c00ff0b7984 */ /* 0x000ea80000000800 */ /*02a0*/ @!P0 STG.E [R6.64], R9 ; /* 0x0000000906008986 */ /* 0x002fe8000c101904 */ /*02b0*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe8000c101904 */ /*02c0*/ STG.E [R4.64], R11 ; /* 0x0000000b04007986 */ /* 0x004fe2000c101904 */ /*02d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02e0*/ BRA 0x2e0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*02f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0300*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0310*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0320*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0330*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0340*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0350*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0360*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0370*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; cudaMalloc(&d_total, sizeof(unsigned int) * numBlocks); cudaMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; cudaMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); cudaMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); cudaFree(d_total_scanned); } cudaFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; cudaMalloc(&d_in, memsize); cudaMalloc(&d_out, memsize); cudaMemcpy(d_in, h_in, memsize, cudaMemcpyHostToDevice); totalScan(d_in, d_out, n); cudaMemcpy(h_out, d_out, memsize, cudaMemcpyDeviceToHost); cudaFree(d_in); cudaFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
.file "tmpxft_0008563e_00000000-6_scan.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 _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m .type _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m, @function _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m: .LFB2084: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11partialScanPjS_S_m(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m, .-_Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m .globl _Z11partialScanPjS_S_m .type _Z11partialScanPjS_S_m, @function _Z11partialScanPjS_S_m: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z11partialScanPjS_S_m, .-_Z11partialScanPjS_S_m .globl _Z29__device_stub__Z7mapScanPjS_mPjS_m .type _Z29__device_stub__Z7mapScanPjS_mPjS_m, @function _Z29__device_stub__Z7mapScanPjS_mPjS_m: .LFB2086: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 120(%rsp), %rax subq %fs:40, %rax jne .L16 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7mapScanPjS_m(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2086: .size _Z29__device_stub__Z7mapScanPjS_mPjS_m, .-_Z29__device_stub__Z7mapScanPjS_mPjS_m .globl _Z7mapScanPjS_m .type _Z7mapScanPjS_m, @function _Z7mapScanPjS_m: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z29__device_stub__Z7mapScanPjS_mPjS_m addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _Z7mapScanPjS_m, .-_Z7mapScanPjS_m .globl _Z9totalScanPjS_m .type _Z9totalScanPjS_m, @function _Z9totalScanPjS_m: .LFB2057: .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 $48, %rsp .cfi_def_cfa_offset 96 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax leaq -1(%rdx), %rbx shrq $2, %rbx addq $1, %rbx leaq 0(,%rbx,4), %r14 movq %rsp, %rdi movq %r14, %rsi call cudaMalloc@PLT movq %r14, %rdx movl $0, %esi movq (%rsp), %rdi call cudaMemset@PLT movl $4, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %ebx, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L20: cmpq $1, %rbx ja .L26 .L21: movq (%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L27 addq $48, %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 .L25: .cfi_restore_state movq %rbp, %rcx movq (%rsp), %rdx movq %r12, %rsi movq %r13, %rdi call _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m jmp .L20 .L26: leaq 8(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movq %r14, %rdx movl $0, %esi movq 8(%rsp), %rdi call cudaMemset@PLT movq %rbx, %rdx movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z9totalScanPjS_m movl $4, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %ebx, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L28 .L22: movq 8(%rsp), %rdi call cudaFree@PLT jmp .L21 .L28: movq %rbp, %rdx movq 8(%rsp), %rsi movq %r12, %rdi call _Z29__device_stub__Z7mapScanPjS_mPjS_m jmp .L22 .L27: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z9totalScanPjS_m, .-_Z9totalScanPjS_m .globl _Z13totalScanHostPjS_m .type _Z13totalScanHostPjS_m, @function _Z13totalScanHostPjS_m: .LFB2058: .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 $40, %rsp .cfi_def_cfa_offset 80 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax leaq 0(,%rdx,4), %rbx leaq 8(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r13, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movq %rbp, %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z9totalScanPjS_m movl $2, %ecx movq %rbx, %rdx movq 16(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rax subq %fs:40, %rax jne .L32 addq $40, %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 .L32: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size _Z13totalScanHostPjS_m, .-_Z13totalScanHostPjS_m .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "h_in = [ " .LC1: .string "%d " .LC2: .string "];\n" .LC3: .string "h_out = [ " .text .globl main .type main, @function main: .LFB2059: .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 movl $140, %edi call malloc@PLT movq %rax, %rbp movl $140, %edi call malloc@PLT movq %rax, %r13 movl $1, %eax .L34: movl %eax, -4(%rbp,%rax,4) addq $1, %rax cmpq $36, %rax jne .L34 movl $35, %edx movq %r13, %rsi movq %rbp, %rdi call _Z13totalScanHostPjS_m leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rbx leaq 140(%rbp), %r14 leaq .LC1(%rip), %r12 .L35: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L35 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r13, %rbx leaq 140(%r13), %r14 leaq .LC1(%rip), %r12 .L36: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L36 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %r13, %rdi call free@PLT movl $0, %eax popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z7mapScanPjS_m" .LC5: .string "_Z11partialScanPjS_S_m" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2089: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z7mapScanPjS_m(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z11partialScanPjS_S_m(%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 .LFE2089: .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 <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; cudaMalloc(&d_total, sizeof(unsigned int) * numBlocks); cudaMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; cudaMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); cudaMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); cudaFree(d_total_scanned); } cudaFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; cudaMalloc(&d_in, memsize); cudaMalloc(&d_out, memsize); cudaMemcpy(d_in, h_in, memsize, cudaMemcpyHostToDevice); totalScan(d_in, d_out, n); cudaMemcpy(h_out, d_out, memsize, cudaMemcpyDeviceToHost); cudaFree(d_in); cudaFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
#include <hip/hip_runtime.h> #include <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; hipMalloc(&d_total, sizeof(unsigned int) * numBlocks); hipMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; hipMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); hipMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); hipFree(d_total_scanned); } hipFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; hipMalloc(&d_in, memsize); hipMalloc(&d_out, memsize); hipMemcpy(d_in, h_in, memsize, hipMemcpyHostToDevice); totalScan(d_in, d_out, n); hipMemcpy(h_out, d_out, memsize, hipMemcpyDeviceToHost); hipFree(d_in); hipFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; hipMalloc(&d_total, sizeof(unsigned int) * numBlocks); hipMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; hipMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); hipMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); hipFree(d_total_scanned); } hipFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; hipMalloc(&d_in, memsize); hipMalloc(&d_out, memsize); hipMemcpy(d_in, h_in, memsize, hipMemcpyHostToDevice); totalScan(d_in, d_out, n); hipMemcpy(h_out, d_out, memsize, hipMemcpyDeviceToHost); hipFree(d_in); hipFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z11partialScanPjS_S_m .globl _Z11partialScanPjS_S_m .p2align 8 .type _Z11partialScanPjS_S_m,@function _Z11partialScanPjS_S_m: s_load_b64 s[2:3], s[0:1], 0x18 s_mov_b32 s4, s15 v_mov_b32_e32 v3, 0 v_lshl_add_u32 v1, s4, 2, v0 s_mov_b32 s5, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 s_waitcnt lgkmcnt(0) v_cmpx_gt_u64_e64 s[2:3], v[1:2] s_cbranch_execz .LBB0_2 s_load_b64 s[6:7], s[0:1], 0x0 v_lshlrev_b64 v[2:3], 2, v[1:2] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s6, v2 v_add_co_ci_u32_e32 v3, vcc_lo, s7, v3, vcc_lo global_load_b32 v3, v[2:3], off .LBB0_2: s_or_b32 exec_lo, exec_lo, s5 v_lshlrev_b32_e32 v2, 2, v0 s_mov_b32 s5, 1 s_waitcnt vmcnt(0) ds_store_b32 v2, v3 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_branch .LBB0_4 .p2align 6 .LBB0_3: s_or_b32 exec_lo, exec_lo, s6 s_lshl_b32 s5, s5, 1 s_waitcnt lgkmcnt(0) s_cmp_gt_i32 s5, 3 s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB0_6 .LBB0_4: v_add_nc_u32_e32 v3, s5, v0 s_mov_b32 s6, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e32 4, v3 s_cbranch_execz .LBB0_3 v_lshlrev_b32_e32 v3, 2, v3 ds_load_b32 v4, v2 ds_load_b32 v5, v3 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v4, v5, v4 ds_store_b32 v3, v4 s_branch .LBB0_3 .LBB0_6: s_load_b64 s[6:7], s[0:1], 0x8 v_add_nc_u32_e32 v1, 1, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 v_cmp_gt_u64_e32 vcc_lo, s[2:3], v[1:2] v_cmp_gt_u32_e64 s2, 3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s3, s2, vcc_lo s_and_saveexec_b32 s2, s3 s_cbranch_execz .LBB0_8 v_lshlrev_b32_e32 v0, 2, v0 ds_load_b32 v3, v0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) 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 global_store_b32 v[0:1], v3, off .LBB0_8: s_or_b32 exec_lo, exec_lo, s2 v_mov_b32_e32 v0, 0 s_load_b64 s[0:1], s[0:1], 0x10 s_ashr_i32 s5, s4, 31 s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b64 s[2:3], s[4:5], 2 ds_load_b32 v1, v0 offset:12 s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 s_clause 0x1 global_store_b32 v0, v0, s[6:7] global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11partialScanPjS_S_m .amdhsa_group_segment_fixed_size 16 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z11partialScanPjS_S_m, .Lfunc_end0-_Z11partialScanPjS_S_m .section .AMDGPU.csdata,"",@progbits .text .protected _Z7mapScanPjS_m .globl _Z7mapScanPjS_m .p2align 8 .type _Z7mapScanPjS_m,@function _Z7mapScanPjS_m: s_load_b64 s[4:5], s[0:1], 0x10 s_mov_b32 s2, s15 s_mov_b32 s3, exec_lo v_lshl_add_u32 v0, s2, 2, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v1, 31, v0 s_waitcnt lgkmcnt(0) v_cmpx_gt_u64_e64 s[4:5], v[0:1] s_cbranch_execz .LBB1_2 s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b64 v[0:1], 2, v[0:1] s_ashr_i32 s3, s2, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) s_lshl_b64 s[0:1], s[2:3], 2 s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s5, v1, vcc_lo s_add_u32 s0, s6, s0 s_addc_u32 s1, s7, s1 s_load_b32 s0, s[0:1], 0x0 global_load_b32 v2, v[0:1], off s_waitcnt vmcnt(0) lgkmcnt(0) v_add_nc_u32_e32 v2, s0, v2 global_store_b32 v[0:1], v2, off .LBB1_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7mapScanPjS_m .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 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_end1: .size _Z7mapScanPjS_m, .Lfunc_end1-_Z7mapScanPjS_m .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 8 .value_kind: by_value .group_segment_fixed_size: 16 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11partialScanPjS_S_m .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z11partialScanPjS_S_m.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z7mapScanPjS_m .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z7mapScanPjS_m.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 <stdlib.h> #include <stdio.h> #define TEST_SIZE 35 #define BLOCK_WIDTH 4 #define CEILING_DIVIDE(X, Y) (1 + (((X) - 1) / (Y))) // Computes a blockwise exclusive sum scan __global__ void partialScan(unsigned int *d_in, unsigned int *d_out, unsigned int *d_total, size_t n) { __shared__ unsigned int temp[BLOCK_WIDTH]; int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { temp[tx] = d_in[index]; } else { temp[tx] = 0; } __syncthreads(); // Perform the actual scan for(int offset = 1; offset < BLOCK_WIDTH; offset <<= 1) { if(tx + offset < BLOCK_WIDTH) { temp[tx + offset] += temp[tx]; } __syncthreads(); } // Shift when copying the result so as to make it an exclusive scan if(tx +1 < BLOCK_WIDTH && index + 1 < n) { d_out[index + 1] = temp[tx]; } d_out[0] = 0; // Store the total sum of each block d_total[bx] = temp[BLOCK_WIDTH - 1]; } // Compute a map on a partial scan to create a total scan from __global__ void mapScan(unsigned int *d_array, unsigned int *d_total, size_t n) { int tx = threadIdx.x; int bx = blockIdx.x; int index = BLOCK_WIDTH * bx + tx; if(index < n) { d_array[index] += d_total[bx]; } } // Compute exclusive sum scan for arbitrary sized array (device pointers as input) void totalScan(unsigned int *d_in, unsigned int *d_out, size_t n) { size_t numBlocks = CEILING_DIVIDE(n, BLOCK_WIDTH); unsigned int *d_total; hipMalloc(&d_total, sizeof(unsigned int) * numBlocks); hipMemset(d_total, 0, sizeof(unsigned int) * numBlocks); partialScan<<<numBlocks, BLOCK_WIDTH>>>(d_in, d_out, d_total, n); if(numBlocks > 1) { unsigned int *d_total_scanned; hipMalloc(&d_total_scanned, sizeof(unsigned int) * numBlocks); hipMemset(d_total_scanned, 0, sizeof(unsigned int) * numBlocks); totalScan(d_total, d_total_scanned, numBlocks); mapScan<<<numBlocks, BLOCK_WIDTH>>>(d_out, d_total_scanned, n); hipFree(d_total_scanned); } hipFree(d_total); } //////////////////////////////////////////////////////////////////////////////// // Wrapper for totalScan (host pointers as input) void totalScanHost(unsigned int *h_in, unsigned int *h_out, size_t n) { unsigned int *d_in; unsigned int *d_out; size_t memsize = sizeof(unsigned int) * n; hipMalloc(&d_in, memsize); hipMalloc(&d_out, memsize); hipMemcpy(d_in, h_in, memsize, hipMemcpyHostToDevice); totalScan(d_in, d_out, n); hipMemcpy(h_out, d_out, memsize, hipMemcpyDeviceToHost); hipFree(d_in); hipFree(d_out); } int main(int argc, char **argv) { unsigned int *h_in; unsigned int *h_out; size_t memsize = sizeof(unsigned int) * TEST_SIZE; h_in = (unsigned int*)malloc(memsize); h_out = (unsigned int*)malloc(memsize); // Test values 1 .. TEST_SIZE for(int i=0; i<TEST_SIZE; i++){ h_in[i] = i+1; } // Compute totalScanHost(h_in, h_out, TEST_SIZE); // Print input printf("h_in = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_in[i]); } printf("];\n"); // Print output printf("h_out = [ "); for(int i=0; i<TEST_SIZE; i++){ printf("%d ", h_out[i]); } printf("];\n"); free(h_in); free(h_out); return 0; }
.text .file "scan.hip" .globl _Z26__device_stub__partialScanPjS_S_m # -- Begin function _Z26__device_stub__partialScanPjS_S_m .p2align 4, 0x90 .type _Z26__device_stub__partialScanPjS_S_m,@function _Z26__device_stub__partialScanPjS_S_m: # @_Z26__device_stub__partialScanPjS_S_m .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movq %rcx, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z11partialScanPjS_S_m, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z26__device_stub__partialScanPjS_S_m, .Lfunc_end0-_Z26__device_stub__partialScanPjS_S_m .cfi_endproc # -- End function .globl _Z22__device_stub__mapScanPjS_m # -- Begin function _Z22__device_stub__mapScanPjS_m .p2align 4, 0x90 .type _Z22__device_stub__mapScanPjS_m,@function _Z22__device_stub__mapScanPjS_m: # @_Z22__device_stub__mapScanPjS_m .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 $_Z7mapScanPjS_m, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z22__device_stub__mapScanPjS_m, .Lfunc_end1-_Z22__device_stub__mapScanPjS_m .cfi_endproc # -- End function .globl _Z9totalScanPjS_m # -- Begin function _Z9totalScanPjS_m .p2align 4, 0x90 .type _Z9totalScanPjS_m,@function _Z9totalScanPjS_m: # @_Z9totalScanPjS_m .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 movq %rsi, 104(%rsp) # 8-byte Spill movq %rdi, %r14 movabsq $4294967300, %rbx # imm = 0x100000004 movq %rdx, 96(%rsp) # 8-byte Spill leaq -1(%rdx), %r15 movq %r15, %r12 shrq $2, %r12 leaq 4(,%r12,4), %r13 incq %r12 leaq 16(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 16(%rsp), %rdi xorl %esi, %esi movq %r13, %rdx callq hipMemset movl %r12d, %eax leaq (%rax,%rbx), %rbp addq $-4, %rbp movq %rbp, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax je .LBB2_1 # %bb.2: cmpq $4, %r15 jae .LBB2_3 jmp .LBB2_6 .LBB2_1: movq 16(%rsp), %rax movq %r14, 88(%rsp) movq 104(%rsp), %rcx # 8-byte Reload movq %rcx, 80(%rsp) movq %rax, 72(%rsp) movq 96(%rsp), %rax # 8-byte Reload movq %rax, 32(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 72(%rsp), %rax movq %rax, 128(%rsp) leaq 32(%rsp), %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 24(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z11partialScanPjS_S_m, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 cmpq $4, %r15 jb .LBB2_6 .LBB2_3: leaq 8(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 8(%rsp), %rdi xorl %esi, %esi movq %r13, %rdx callq hipMemset movq 16(%rsp), %rdi movq 8(%rsp), %rsi movq %r12, %rdx callq _Z9totalScanPjS_m movq %rbp, %rdi movl $1, %esi movabsq $4294967300, %rdx # imm = 0x100000004 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_5 # %bb.4: movq 8(%rsp), %rax movq 104(%rsp), %rcx # 8-byte Reload movq %rcx, 88(%rsp) movq %rax, 80(%rsp) movq 96(%rsp), %rax # 8-byte Reload movq %rax, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 72(%rsp), %rax movq %rax, 128(%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 112(%rsp), %r9 movl $_Z7mapScanPjS_m, %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_5: movq 8(%rsp), %rdi callq hipFree .LBB2_6: movq 16(%rsp), %rdi callq hipFree addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9totalScanPjS_m, .Lfunc_end2-_Z9totalScanPjS_m .cfi_endproc # -- End function .globl _Z13totalScanHostPjS_m # -- Begin function _Z13totalScanHostPjS_m .p2align 4, 0x90 .type _Z13totalScanHostPjS_m,@function _Z13totalScanHostPjS_m: # @_Z13totalScanHostPjS_m .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 subq $24, %rsp .cfi_def_cfa_offset 64 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdx, %r15 movq %rsi, %rbx movq %rdi, %r12 leaq (,%rdx,4), %r14 leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %r12, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy movq 16(%rsp), %rdi movq 8(%rsp), %rsi movq %r15, %rdx callq _Z9totalScanPjS_m movq 8(%rsp), %rsi movq %rbx, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree addq $24, %rsp .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end3: .size _Z13totalScanHostPjS_m, .Lfunc_end3-_Z13totalScanHostPjS_m .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 $16, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $140, %edi callq malloc movq %rax, %rbx movl $140, %edi callq malloc movq %rax, %r14 xorl %eax, %eax .p2align 4, 0x90 .LBB4_1: # =>This Inner Loop Header: Depth=1 leaq 1(%rax), %rcx movl %ecx, (%rbx,%rax,4) movq %rcx, %rax cmpq $35, %rcx jne .LBB4_1 # %bb.2: leaq 8(%rsp), %rdi movl $140, %esi callq hipMalloc movq %rsp, %rdi movl $140, %esi callq hipMalloc movq 8(%rsp), %rdi movl $140, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq (%rsp), %rsi movl $35, %edx callq _Z9totalScanPjS_m movq (%rsp), %rsi movl $140, %edx movq %r14, %rdi movl $2, %ecx callq hipMemcpy movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree movl $.L.str, %edi xorl %eax, %eax callq printf xorl %r15d, %r15d .p2align 4, 0x90 .LBB4_3: # =>This Inner Loop Header: Depth=1 movl (%rbx,%r15,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r15 cmpq $35, %r15 jne .LBB4_3 # %bb.4: movl $.Lstr.1, %edi callq puts@PLT movl $.L.str.3, %edi xorl %eax, %eax callq printf xorl %r15d, %r15d .p2align 4, 0x90 .LBB4_5: # =>This Inner Loop Header: Depth=1 movl (%r14,%r15,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r15 cmpq $35, %r15 jne .LBB4_5 # %bb.6: movl $.Lstr.1, %edi callq puts@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $16, %rsp .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size main, .Lfunc_end4-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11partialScanPjS_S_m, %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 $_Z7mapScanPjS_m, %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_end5: .size __hip_module_ctor, .Lfunc_end5-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB6_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB6_2: retq .Lfunc_end6: .size __hip_module_dtor, .Lfunc_end6-__hip_module_dtor .cfi_endproc # -- End function .type _Z11partialScanPjS_S_m,@object # @_Z11partialScanPjS_S_m .section .rodata,"a",@progbits .globl _Z11partialScanPjS_S_m .p2align 3, 0x0 _Z11partialScanPjS_S_m: .quad _Z26__device_stub__partialScanPjS_S_m .size _Z11partialScanPjS_S_m, 8 .type _Z7mapScanPjS_m,@object # @_Z7mapScanPjS_m .globl _Z7mapScanPjS_m .p2align 3, 0x0 _Z7mapScanPjS_m: .quad _Z22__device_stub__mapScanPjS_m .size _Z7mapScanPjS_m, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "h_in = [ " .size .L.str, 10 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d " .size .L.str.1, 4 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "h_out = [ " .size .L.str.3, 11 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z11partialScanPjS_S_m" .size .L__unnamed_1, 23 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z7mapScanPjS_m" .size .L__unnamed_2, 16 .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.1,@object # @str.1 .section .rodata.str1.1,"aMS",@progbits,1 .Lstr.1: .asciz "] .size .Lstr.1, 3 .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__partialScanPjS_S_m .addrsig_sym _Z22__device_stub__mapScanPjS_m .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11partialScanPjS_S_m .addrsig_sym _Z7mapScanPjS_m .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 : _Z7mapScanPjS_m .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e280000002100 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0030*/ IMAD R0, R3, 0x4, R0 ; /* 0x0000000403007824 */ /* 0x001fca00078e0200 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe40003f06070 */ /*0050*/ SHF.R.S32.HI R5, RZ, 0x1f, R0 ; /* 0x0000001fff057819 */ /* 0x000fc80000011400 */ /*0060*/ ISETP.GE.U32.AND.EX P0, PT, R5, c[0x0][0x174], PT, P0 ; /* 0x00005d0005007a0c */ /* 0x000fda0003f06100 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ IMAD.MOV.U32 R2, RZ, RZ, 0x4 ; /* 0x00000004ff027424 */ /* 0x000fe200078e00ff */ /*0090*/ LEA R4, P0, R0.reuse, c[0x0][0x160], 0x2 ; /* 0x0000580000047a11 */ /* 0x040fe200078010ff */ /*00a0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*00b0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x168] ; /* 0x00005a0003027625 */ /* 0x000fe200078e0202 */ /*00c0*/ LEA.HI.X R5, R0, c[0x0][0x164], R5, 0x2, P0 ; /* 0x0000590000057a11 */ /* 0x000fca00000f1405 */ /*00d0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea8000c1e1900 */ /*00e0*/ LDG.E R0, [R4.64] ; /* 0x0000000404007981 */ /* 0x000ea4000c1e1900 */ /*00f0*/ IMAD.IADD R7, R0, 0x1, R3 ; /* 0x0000000100077824 */ /* 0x004fca00078e0203 */ /*0100*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*0110*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0120*/ BRA 0x120; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ .......... Function : _Z11partialScanPjS_S_m .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R8, SR_CTAID.X ; /* 0x0000000000087919 */ /* 0x000e220000002500 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ BSSY B0, 0x110 ; /* 0x000000d000007945 */ /* 0x000fe40003800000 */ /*0040*/ S2R R13, SR_TID.X ; /* 0x00000000000d7919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R0, R8, 0x4, R13 ; /* 0x0000000408007824 */ /* 0x001fe200078e020d */ /*0060*/ ISETP.GT.AND P0, PT, R13, 0x2, PT ; /* 0x000000020d00780c */ /* 0x000fc80003f04270 */ /*0070*/ ISETP.GE.U32.AND P1, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe40003f26070 */ /*0080*/ SHF.R.S32.HI R3, RZ, 0x1f, R0 ; /* 0x0000001fff037819 */ /* 0x000fc80000011400 */ /*0090*/ ISETP.GE.U32.AND.EX P1, PT, R3, c[0x0][0x17c], PT, P1 ; /* 0x00005f0003007a0c */ /* 0x000fda0003f26110 */ /*00a0*/ @P1 STS [R13.X4], RZ ; /* 0x000000ff0d001388 */ /* 0x0001e20000004800 */ /*00b0*/ @P1 BRA 0x100 ; /* 0x0000004000001947 */ /* 0x000fea0003800000 */ /*00c0*/ LEA R2, P1, R0, c[0x0][0x160], 0x2 ; /* 0x0000580000027a11 */ /* 0x000fc800078210ff */ /*00d0*/ LEA.HI.X R3, R0, c[0x0][0x164], R3, 0x2, P1 ; /* 0x0000590000037a11 */ /* 0x000fca00008f1403 */ /*00e0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00f0*/ STS [R13.X4], R2 ; /* 0x000000020d007388 */ /* 0x0043e40000004800 */ /*0100*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*0110*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0120*/ ISETP.GT.AND P1, PT, R13, 0x1, PT ; /* 0x000000010d00780c */ /* 0x000fe20003f24270 */ /*0130*/ IMAD.MOV.U32 R15, RZ, RZ, 0x4 ; /* 0x00000004ff0f7424 */ /* 0x000fe200078e00ff */ /*0140*/ IADD3 R5, R0, 0x1, RZ ; /* 0x0000000100057810 */ /* 0x000fc80007ffe0ff */ /*0150*/ SHF.R.S32.HI R10, RZ, 0x1f, R5 ; /* 0x0000001fff0a7819 */ /* 0x000fe20000011405 */ /*0160*/ @!P0 LDS R2, [R13.X4] ; /* 0x000000000d028984 */ /* 0x002fe80000004800 */ /*0170*/ @!P0 LDS R3, [R13.X4+0x4] ; /* 0x000004000d038984 */ /* 0x000e640000004800 */ /*0180*/ @!P0 IMAD.IADD R2, R2, 0x1, R3 ; /* 0x0000000102028824 */ /* 0x002fca00078e0203 */ /*0190*/ @!P0 STS [R13.X4+0x4], R2 ; /* 0x000004020d008388 */ /* 0x0003e80000004800 */ /*01a0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*01b0*/ ISETP.GE.U32.AND P0, PT, R5, c[0x0][0x178], PT ; /* 0x00005e0005007a0c */ /* 0x000fc80003f06070 */ /*01c0*/ ISETP.GE.U32.AND.EX P0, PT, R10, c[0x0][0x17c], PT, P0 ; /* 0x00005f000a007a0c */ /* 0x000fe20003f06100 */ /*01d0*/ IMAD.MOV.U32 R2, RZ, RZ, c[0x0][0x168] ; /* 0x00005a00ff027624 */ /* 0x002fc600078e00ff */ /*01e0*/ ISETP.GT.OR P0, PT, R13, 0x2, P0 ; /* 0x000000020d00780c */ /* 0x000fe20000704670 */ /*01f0*/ @!P1 LDS R3, [R13.X4] ; /* 0x000000000d039984 */ /* 0x000fe80000004800 */ /*0200*/ @!P1 LDS R4, [R13.X4+0x8] ; /* 0x000008000d049984 */ /* 0x000e640000004800 */ /*0210*/ @!P1 IMAD.IADD R0, R3, 0x1, R4 ; /* 0x0000000103009824 */ /* 0x002fe400078e0204 */ /*0220*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x0][0x16c] ; /* 0x00005b00ff037624 */ /* 0x000fc600078e00ff */ /*0230*/ @!P1 STS [R13.X4+0x8], R0 ; /* 0x000008000d009388 */ /* 0x000fe80000004800 */ /*0240*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0250*/ @!P0 LEA R6, P1, R5, c[0x0][0x168], 0x2 ; /* 0x00005a0005068a11 */ /* 0x000fc800078210ff */ /*0260*/ @!P0 LEA.HI.X R7, R5, c[0x0][0x16c], R10, 0x2, P1 ; /* 0x00005b0005078a11 */ /* 0x000fe200008f140a */ /*0270*/ IMAD.WIDE R4, R8, R15, c[0x0][0x170] ; /* 0x00005c0008047625 */ /* 0x000fe200078e020f */ /*0280*/ @!P0 LDS R9, [R13.X4] ; /* 0x000000000d098984 */ /* 0x000e680000004800 */ /*0290*/ LDS R11, [0xc] ; /* 0x00000c00ff0b7984 */ /* 0x000ea80000000800 */ /*02a0*/ @!P0 STG.E [R6.64], R9 ; /* 0x0000000906008986 */ /* 0x002fe8000c101904 */ /*02b0*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x000fe8000c101904 */ /*02c0*/ STG.E [R4.64], R11 ; /* 0x0000000b04007986 */ /* 0x004fe2000c101904 */ /*02d0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*02e0*/ BRA 0x2e0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*02f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0300*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0310*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0320*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0330*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0340*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0350*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0360*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0370*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z11partialScanPjS_S_m .globl _Z11partialScanPjS_S_m .p2align 8 .type _Z11partialScanPjS_S_m,@function _Z11partialScanPjS_S_m: s_load_b64 s[2:3], s[0:1], 0x18 s_mov_b32 s4, s15 v_mov_b32_e32 v3, 0 v_lshl_add_u32 v1, s4, 2, v0 s_mov_b32 s5, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 s_waitcnt lgkmcnt(0) v_cmpx_gt_u64_e64 s[2:3], v[1:2] s_cbranch_execz .LBB0_2 s_load_b64 s[6:7], s[0:1], 0x0 v_lshlrev_b64 v[2:3], 2, v[1:2] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s6, v2 v_add_co_ci_u32_e32 v3, vcc_lo, s7, v3, vcc_lo global_load_b32 v3, v[2:3], off .LBB0_2: s_or_b32 exec_lo, exec_lo, s5 v_lshlrev_b32_e32 v2, 2, v0 s_mov_b32 s5, 1 s_waitcnt vmcnt(0) ds_store_b32 v2, v3 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_branch .LBB0_4 .p2align 6 .LBB0_3: s_or_b32 exec_lo, exec_lo, s6 s_lshl_b32 s5, s5, 1 s_waitcnt lgkmcnt(0) s_cmp_gt_i32 s5, 3 s_barrier buffer_gl0_inv s_cbranch_scc1 .LBB0_6 .LBB0_4: v_add_nc_u32_e32 v3, s5, v0 s_mov_b32 s6, exec_lo s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e32 4, v3 s_cbranch_execz .LBB0_3 v_lshlrev_b32_e32 v3, 2, v3 ds_load_b32 v4, v2 ds_load_b32 v5, v3 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v4, v5, v4 ds_store_b32 v3, v4 s_branch .LBB0_3 .LBB0_6: s_load_b64 s[6:7], s[0:1], 0x8 v_add_nc_u32_e32 v1, 1, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 v_cmp_gt_u64_e32 vcc_lo, s[2:3], v[1:2] v_cmp_gt_u32_e64 s2, 3, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s3, s2, vcc_lo s_and_saveexec_b32 s2, s3 s_cbranch_execz .LBB0_8 v_lshlrev_b32_e32 v0, 2, v0 ds_load_b32 v3, v0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) 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 global_store_b32 v[0:1], v3, off .LBB0_8: s_or_b32 exec_lo, exec_lo, s2 v_mov_b32_e32 v0, 0 s_load_b64 s[0:1], s[0:1], 0x10 s_ashr_i32 s5, s4, 31 s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b64 s[2:3], s[4:5], 2 ds_load_b32 v1, v0 offset:12 s_waitcnt lgkmcnt(0) s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 s_clause 0x1 global_store_b32 v0, v0, s[6:7] global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z11partialScanPjS_S_m .amdhsa_group_segment_fixed_size 16 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z11partialScanPjS_S_m, .Lfunc_end0-_Z11partialScanPjS_S_m .section .AMDGPU.csdata,"",@progbits .text .protected _Z7mapScanPjS_m .globl _Z7mapScanPjS_m .p2align 8 .type _Z7mapScanPjS_m,@function _Z7mapScanPjS_m: s_load_b64 s[4:5], s[0:1], 0x10 s_mov_b32 s2, s15 s_mov_b32 s3, exec_lo v_lshl_add_u32 v0, s2, 2, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v1, 31, v0 s_waitcnt lgkmcnt(0) v_cmpx_gt_u64_e64 s[4:5], v[0:1] s_cbranch_execz .LBB1_2 s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b64 v[0:1], 2, v[0:1] s_ashr_i32 s3, s2, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) s_lshl_b64 s[0:1], s[2:3], 2 s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s5, v1, vcc_lo s_add_u32 s0, s6, s0 s_addc_u32 s1, s7, s1 s_load_b32 s0, s[0:1], 0x0 global_load_b32 v2, v[0:1], off s_waitcnt vmcnt(0) lgkmcnt(0) v_add_nc_u32_e32 v2, s0, v2 global_store_b32 v[0:1], v2, off .LBB1_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7mapScanPjS_m .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 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_end1: .size _Z7mapScanPjS_m, .Lfunc_end1-_Z7mapScanPjS_m .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 8 .value_kind: by_value .group_segment_fixed_size: 16 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z11partialScanPjS_S_m .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z11partialScanPjS_S_m.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 8 .value_kind: by_value .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z7mapScanPjS_m .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z7mapScanPjS_m.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_0008563e_00000000-6_scan.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 _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m .type _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m, @function _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m: .LFB2084: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11partialScanPjS_S_m(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2084: .size _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m, .-_Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m .globl _Z11partialScanPjS_S_m .type _Z11partialScanPjS_S_m, @function _Z11partialScanPjS_S_m: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _Z11partialScanPjS_S_m, .-_Z11partialScanPjS_S_m .globl _Z29__device_stub__Z7mapScanPjS_mPjS_m .type _Z29__device_stub__Z7mapScanPjS_mPjS_m, @function _Z29__device_stub__Z7mapScanPjS_mPjS_m: .LFB2086: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L15 .L11: movq 120(%rsp), %rax subq %fs:40, %rax jne .L16 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z7mapScanPjS_m(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L11 .L16: call __stack_chk_fail@PLT .cfi_endproc .LFE2086: .size _Z29__device_stub__Z7mapScanPjS_mPjS_m, .-_Z29__device_stub__Z7mapScanPjS_mPjS_m .globl _Z7mapScanPjS_m .type _Z7mapScanPjS_m, @function _Z7mapScanPjS_m: .LFB2087: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z29__device_stub__Z7mapScanPjS_mPjS_m addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2087: .size _Z7mapScanPjS_m, .-_Z7mapScanPjS_m .globl _Z9totalScanPjS_m .type _Z9totalScanPjS_m, @function _Z9totalScanPjS_m: .LFB2057: .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 $48, %rsp .cfi_def_cfa_offset 96 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax leaq -1(%rdx), %rbx shrq $2, %rbx addq $1, %rbx leaq 0(,%rbx,4), %r14 movq %rsp, %rdi movq %r14, %rsi call cudaMalloc@PLT movq %r14, %rdx movl $0, %esi movq (%rsp), %rdi call cudaMemset@PLT movl $4, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %ebx, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L20: cmpq $1, %rbx ja .L26 .L21: movq (%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L27 addq $48, %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 .L25: .cfi_restore_state movq %rbp, %rcx movq (%rsp), %rdx movq %r12, %rsi movq %r13, %rdi call _Z36__device_stub__Z11partialScanPjS_S_mPjS_S_m jmp .L20 .L26: leaq 8(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movq %r14, %rdx movl $0, %esi movq 8(%rsp), %rdi call cudaMemset@PLT movq %rbx, %rdx movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z9totalScanPjS_m movl $4, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %ebx, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L28 .L22: movq 8(%rsp), %rdi call cudaFree@PLT jmp .L21 .L28: movq %rbp, %rdx movq 8(%rsp), %rsi movq %r12, %rdi call _Z29__device_stub__Z7mapScanPjS_mPjS_m jmp .L22 .L27: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z9totalScanPjS_m, .-_Z9totalScanPjS_m .globl _Z13totalScanHostPjS_m .type _Z13totalScanHostPjS_m, @function _Z13totalScanHostPjS_m: .LFB2058: .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 $40, %rsp .cfi_def_cfa_offset 80 movq %rdi, %r13 movq %rsi, %r12 movq %rdx, %rbp movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax leaq 0(,%rdx,4), %rbx leaq 8(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT leaq 16(%rsp), %rdi movq %rbx, %rsi call cudaMalloc@PLT movl $1, %ecx movq %rbx, %rdx movq %r13, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movq %rbp, %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z9totalScanPjS_m movl $2, %ecx movq %rbx, %rdx movq 16(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rax subq %fs:40, %rax jne .L32 addq $40, %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 .L32: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2058: .size _Z13totalScanHostPjS_m, .-_Z13totalScanHostPjS_m .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "h_in = [ " .LC1: .string "%d " .LC2: .string "];\n" .LC3: .string "h_out = [ " .text .globl main .type main, @function main: .LFB2059: .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 movl $140, %edi call malloc@PLT movq %rax, %rbp movl $140, %edi call malloc@PLT movq %rax, %r13 movl $1, %eax .L34: movl %eax, -4(%rbp,%rax,4) addq $1, %rax cmpq $36, %rax jne .L34 movl $35, %edx movq %r13, %rsi movq %rbp, %rdi call _Z13totalScanHostPjS_m leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rbx leaq 140(%rbp), %r14 leaq .LC1(%rip), %r12 .L35: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L35 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT leaq .LC3(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %r13, %rbx leaq 140(%r13), %r14 leaq .LC1(%rip), %r12 .L36: movl (%rbx), %edx movq %r12, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addq $4, %rbx cmpq %r14, %rbx jne .L36 leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %r13, %rdi call free@PLT movl $0, %eax popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size main, .-main .section .rodata.str1.1 .LC4: .string "_Z7mapScanPjS_m" .LC5: .string "_Z11partialScanPjS_S_m" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2089: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z7mapScanPjS_m(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z11partialScanPjS_S_m(%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 .LFE2089: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "scan.hip" .globl _Z26__device_stub__partialScanPjS_S_m # -- Begin function _Z26__device_stub__partialScanPjS_S_m .p2align 4, 0x90 .type _Z26__device_stub__partialScanPjS_S_m,@function _Z26__device_stub__partialScanPjS_S_m: # @_Z26__device_stub__partialScanPjS_S_m .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movq %rcx, 48(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rax movq %rax, 104(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z11partialScanPjS_S_m, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z26__device_stub__partialScanPjS_S_m, .Lfunc_end0-_Z26__device_stub__partialScanPjS_S_m .cfi_endproc # -- End function .globl _Z22__device_stub__mapScanPjS_m # -- Begin function _Z22__device_stub__mapScanPjS_m .p2align 4, 0x90 .type _Z22__device_stub__mapScanPjS_m,@function _Z22__device_stub__mapScanPjS_m: # @_Z22__device_stub__mapScanPjS_m .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 $_Z7mapScanPjS_m, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end1: .size _Z22__device_stub__mapScanPjS_m, .Lfunc_end1-_Z22__device_stub__mapScanPjS_m .cfi_endproc # -- End function .globl _Z9totalScanPjS_m # -- Begin function _Z9totalScanPjS_m .p2align 4, 0x90 .type _Z9totalScanPjS_m,@function _Z9totalScanPjS_m: # @_Z9totalScanPjS_m .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 movq %rsi, 104(%rsp) # 8-byte Spill movq %rdi, %r14 movabsq $4294967300, %rbx # imm = 0x100000004 movq %rdx, 96(%rsp) # 8-byte Spill leaq -1(%rdx), %r15 movq %r15, %r12 shrq $2, %r12 leaq 4(,%r12,4), %r13 incq %r12 leaq 16(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 16(%rsp), %rdi xorl %esi, %esi movq %r13, %rdx callq hipMemset movl %r12d, %eax leaq (%rax,%rbx), %rbp addq $-4, %rbp movq %rbp, %rdi movl $1, %esi movq %rbx, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax je .LBB2_1 # %bb.2: cmpq $4, %r15 jae .LBB2_3 jmp .LBB2_6 .LBB2_1: movq 16(%rsp), %rax movq %r14, 88(%rsp) movq 104(%rsp), %rcx # 8-byte Reload movq %rcx, 80(%rsp) movq %rax, 72(%rsp) movq 96(%rsp), %rax # 8-byte Reload movq %rax, 32(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 72(%rsp), %rax movq %rax, 128(%rsp) leaq 32(%rsp), %rax movq %rax, 136(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 24(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z11partialScanPjS_S_m, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 cmpq $4, %r15 jb .LBB2_6 .LBB2_3: leaq 8(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 8(%rsp), %rdi xorl %esi, %esi movq %r13, %rdx callq hipMemset movq 16(%rsp), %rdi movq 8(%rsp), %rsi movq %r12, %rdx callq _Z9totalScanPjS_m movq %rbp, %rdi movl $1, %esi movabsq $4294967300, %rdx # imm = 0x100000004 movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_5 # %bb.4: movq 8(%rsp), %rax movq 104(%rsp), %rcx # 8-byte Reload movq %rcx, 88(%rsp) movq %rax, 80(%rsp) movq 96(%rsp), %rax # 8-byte Reload movq %rax, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 72(%rsp), %rax movq %rax, 128(%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 112(%rsp), %r9 movl $_Z7mapScanPjS_m, %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_5: movq 8(%rsp), %rdi callq hipFree .LBB2_6: movq 16(%rsp), %rdi callq hipFree addq $152, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9totalScanPjS_m, .Lfunc_end2-_Z9totalScanPjS_m .cfi_endproc # -- End function .globl _Z13totalScanHostPjS_m # -- Begin function _Z13totalScanHostPjS_m .p2align 4, 0x90 .type _Z13totalScanHostPjS_m,@function _Z13totalScanHostPjS_m: # @_Z13totalScanHostPjS_m .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 subq $24, %rsp .cfi_def_cfa_offset 64 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdx, %r15 movq %rsi, %rbx movq %rdi, %r12 leaq (,%rdx,4), %r14 leaq 16(%rsp), %rdi movq %r14, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r14, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %r12, %rsi movq %r14, %rdx movl $1, %ecx callq hipMemcpy movq 16(%rsp), %rdi movq 8(%rsp), %rsi movq %r15, %rdx callq _Z9totalScanPjS_m movq 8(%rsp), %rsi movq %rbx, %rdi movq %r14, %rdx movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree addq $24, %rsp .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end3: .size _Z13totalScanHostPjS_m, .Lfunc_end3-_Z13totalScanHostPjS_m .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 $16, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $140, %edi callq malloc movq %rax, %rbx movl $140, %edi callq malloc movq %rax, %r14 xorl %eax, %eax .p2align 4, 0x90 .LBB4_1: # =>This Inner Loop Header: Depth=1 leaq 1(%rax), %rcx movl %ecx, (%rbx,%rax,4) movq %rcx, %rax cmpq $35, %rcx jne .LBB4_1 # %bb.2: leaq 8(%rsp), %rdi movl $140, %esi callq hipMalloc movq %rsp, %rdi movl $140, %esi callq hipMalloc movq 8(%rsp), %rdi movl $140, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq (%rsp), %rsi movl $35, %edx callq _Z9totalScanPjS_m movq (%rsp), %rsi movl $140, %edx movq %r14, %rdi movl $2, %ecx callq hipMemcpy movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree movl $.L.str, %edi xorl %eax, %eax callq printf xorl %r15d, %r15d .p2align 4, 0x90 .LBB4_3: # =>This Inner Loop Header: Depth=1 movl (%rbx,%r15,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r15 cmpq $35, %r15 jne .LBB4_3 # %bb.4: movl $.Lstr.1, %edi callq puts@PLT movl $.L.str.3, %edi xorl %eax, %eax callq printf xorl %r15d, %r15d .p2align 4, 0x90 .LBB4_5: # =>This Inner Loop Header: Depth=1 movl (%r14,%r15,4), %esi movl $.L.str.1, %edi xorl %eax, %eax callq printf incq %r15 cmpq $35, %r15 jne .LBB4_5 # %bb.6: movl $.Lstr.1, %edi callq puts@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free xorl %eax, %eax addq $16, %rsp .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size main, .Lfunc_end4-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB5_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB5_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z11partialScanPjS_S_m, %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 $_Z7mapScanPjS_m, %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_end5: .size __hip_module_ctor, .Lfunc_end5-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB6_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB6_2: retq .Lfunc_end6: .size __hip_module_dtor, .Lfunc_end6-__hip_module_dtor .cfi_endproc # -- End function .type _Z11partialScanPjS_S_m,@object # @_Z11partialScanPjS_S_m .section .rodata,"a",@progbits .globl _Z11partialScanPjS_S_m .p2align 3, 0x0 _Z11partialScanPjS_S_m: .quad _Z26__device_stub__partialScanPjS_S_m .size _Z11partialScanPjS_S_m, 8 .type _Z7mapScanPjS_m,@object # @_Z7mapScanPjS_m .globl _Z7mapScanPjS_m .p2align 3, 0x0 _Z7mapScanPjS_m: .quad _Z22__device_stub__mapScanPjS_m .size _Z7mapScanPjS_m, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "h_in = [ " .size .L.str, 10 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "%d " .size .L.str.1, 4 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "h_out = [ " .size .L.str.3, 11 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z11partialScanPjS_S_m" .size .L__unnamed_1, 23 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "_Z7mapScanPjS_m" .size .L__unnamed_2, 16 .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.1,@object # @str.1 .section .rodata.str1.1,"aMS",@progbits,1 .Lstr.1: .asciz "] .size .Lstr.1, 3 .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__partialScanPjS_S_m .addrsig_sym _Z22__device_stub__mapScanPjS_m .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z11partialScanPjS_S_m .addrsig_sym _Z7mapScanPjS_m .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.
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <cuda.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
code for sm_80 Function : _Z3fooPf .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 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x160] ; /* 0x00005800ff047624 */ /* 0x000fe200078e00ff */ /*0030*/ MOV R5, c[0x0][0x164] ; /* 0x0000590000057a02 */ /* 0x000fe40000000f00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fe400078e0203 */ /*0060*/ HFMA2.MMA R9, -RZ, RZ, 0, 0 ; /* 0x00000000ff097435 */ /* 0x000fe200000001ff */ /*0070*/ IMAD.MOV.U32 R8, RZ, RZ, 0x3 ; /* 0x00000003ff087424 */ /* 0x000fe200078e00ff */ /*0080*/ ULDC.64 UR36, c[0x0][0x118] ; /* 0x0000460000247ab9 */ /* 0x000fe20000000a00 */ /*0090*/ IMAD.MOV.U32 R6, RZ, RZ, R2 ; /* 0x000000ffff067224 */ /* 0x000fe400078e0002 */ /*00a0*/ MOV R20, 0xd0 ; /* 0x000000d000147802 */ /* 0x000fe40000000f00 */ /*00b0*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fc80000000f00 */ /*00c0*/ CALL.REL.NOINC R8 0x0 ; /* 0xffffff3008007344 */ /* 0x000fea0003c3ffff */ /*00d0*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fd400000001ff */ /*00e0*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0003 */ /*00f0*/ STG.E [R2.64], R4 ; /* 0x0000000402007986 */ /* 0x000fe2000c101924 */ /*0100*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0110*/ BRA 0x110; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <cuda.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
.file "tmpxft_0002196b_00000000-6_main.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z13multiplyByTwoPfj .type _Z13multiplyByTwoPfj, @function _Z13multiplyByTwoPfj: .LFB2057: .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 .LFE2057: .size _Z13multiplyByTwoPfj, .-_Z13multiplyByTwoPfj .globl _Z11divideByTwoPfj .type _Z11divideByTwoPfj, @function _Z11divideByTwoPfj: .LFB2058: .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 .LFE2058: .size _Z11divideByTwoPfj, .-_Z11divideByTwoPfj .globl _Z22__device_stub__Z3fooPfPf .type _Z22__device_stub__Z3fooPfPf, @function _Z22__device_stub__Z3fooPfPf: .LFB2083: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z3fooPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z22__device_stub__Z3fooPfPf, .-_Z22__device_stub__Z3fooPfPf .globl _Z3fooPf .type _Z3fooPf, @function _Z3fooPf: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z22__device_stub__Z3fooPfPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z3fooPf, .-_Z3fooPf .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3fooPf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3fooPf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <cuda.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <hip/hip_runtime.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <hip/hip_runtime.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3fooPf .globl _Z3fooPf .p2align 8 .type _Z3fooPf,@function _Z3fooPf: s_mov_b64 s[10:11], s[4:5] s_clause 0x1 s_load_b32 s4, s[2:3], 0x14 s_load_b64 s[34:35], s[2:3], 0x0 v_and_b32_e32 v1, 0x3ff, v0 s_mov_b32 s12, s13 v_mov_b32_e32 v31, v0 s_mov_b32 s13, s14 s_mov_b32 s14, s15 s_mov_b32 s32, 0 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff v_mov_b32_e32 v0, s34 v_mad_u64_u32 v[40:41], null, s12, s4, v[1:2] v_mov_b32_e32 v1, s35 s_add_u32 s8, s2, 8 s_addc_u32 s9, s3, 0 s_mov_b64 s[4:5], s[0:1] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mov_b32_e32 v2, v40 s_swappc_b64 s[30:31], 3 v_mov_b32_e32 v41, 0 v_lshlrev_b64 v[1:2], 2, v[40:41] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v1, vcc_lo, s34, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s35, v2, vcc_lo global_store_b32 v[1:2], v0, off s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3fooPf .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 13 .amdhsa_user_sgpr_dispatch_ptr 1 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 1 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 1 .amdhsa_enable_private_segment 1 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 1 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 2 .amdhsa_next_free_vgpr 42 .amdhsa_next_free_sgpr 36 .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 _Z3fooPf, .Lfunc_end0-_Z3fooPf .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims - .offset: 88 .size: 8 .value_kind: hidden_hostcall_buffer - .offset: 96 .size: 8 .value_kind: hidden_multigrid_sync_arg - .offset: 104 .size: 8 .value_kind: hidden_heap_v1 - .offset: 112 .size: 8 .value_kind: hidden_default_queue - .offset: 208 .size: 8 .value_kind: hidden_queue_ptr .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3fooPf .private_segment_fixed_size: 0 .sgpr_count: 38 .sgpr_spill_count: 0 .symbol: _Z3fooPf.kd .uniform_work_group_size: 1 .uses_dynamic_stack: true .vgpr_count: 42 .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.
//xfail:BOOGIE_ERROR //--blockDim=1024 --gridDim=1 --no-inline //error: possible null pointer access #include <stdio.h> #include <assert.h> #include <hip/hip_runtime.h> #define N 2//8 #define tid (blockIdx.x * blockDim.x + threadIdx.x) __device__ float multiplyByTwo(float *v, unsigned int index) { return v[index] * 2.0f; } __device__ float divideByTwo(float *v, unsigned int index) { return v[index] * 0.5f; } typedef float(*funcType)(float*, unsigned int); __global__ void foo(float *v) { funcType f = (funcType)3; // it's a null pointer v[tid] = (*f)(v, tid); }
.text .file "main.hip" .globl _Z18__device_stub__fooPf # -- Begin function _Z18__device_stub__fooPf .p2align 4, 0x90 .type _Z18__device_stub__fooPf,@function _Z18__device_stub__fooPf: # @_Z18__device_stub__fooPf .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 $_Z3fooPf, %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 _Z18__device_stub__fooPf, .Lfunc_end0-_Z18__device_stub__fooPf .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 $_Z3fooPf, %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 _Z3fooPf,@object # @_Z3fooPf .section .rodata,"a",@progbits .globl _Z3fooPf .p2align 3, 0x0 _Z3fooPf: .quad _Z18__device_stub__fooPf .size _Z3fooPf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3fooPf" .size .L__unnamed_1, 9 .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__fooPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3fooPf .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 : _Z3fooPf .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 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x0][0x160] ; /* 0x00005800ff047624 */ /* 0x000fe200078e00ff */ /*0030*/ MOV R5, c[0x0][0x164] ; /* 0x0000590000057a02 */ /* 0x000fe40000000f00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fe400078e0203 */ /*0060*/ HFMA2.MMA R9, -RZ, RZ, 0, 0 ; /* 0x00000000ff097435 */ /* 0x000fe200000001ff */ /*0070*/ IMAD.MOV.U32 R8, RZ, RZ, 0x3 ; /* 0x00000003ff087424 */ /* 0x000fe200078e00ff */ /*0080*/ ULDC.64 UR36, c[0x0][0x118] ; /* 0x0000460000247ab9 */ /* 0x000fe20000000a00 */ /*0090*/ IMAD.MOV.U32 R6, RZ, RZ, R2 ; /* 0x000000ffff067224 */ /* 0x000fe400078e0002 */ /*00a0*/ MOV R20, 0xd0 ; /* 0x000000d000147802 */ /* 0x000fe40000000f00 */ /*00b0*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fc80000000f00 */ /*00c0*/ CALL.REL.NOINC R8 0x0 ; /* 0xffffff3008007344 */ /* 0x000fea0003c3ffff */ /*00d0*/ HFMA2.MMA R3, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff037435 */ /* 0x000fd400000001ff */ /*00e0*/ IMAD.WIDE.U32 R2, R2, R3, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0003 */ /*00f0*/ STG.E [R2.64], R4 ; /* 0x0000000402007986 */ /* 0x000fe2000c101924 */ /*0100*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0110*/ BRA 0x110; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3fooPf .globl _Z3fooPf .p2align 8 .type _Z3fooPf,@function _Z3fooPf: s_mov_b64 s[10:11], s[4:5] s_clause 0x1 s_load_b32 s4, s[2:3], 0x14 s_load_b64 s[34:35], s[2:3], 0x0 v_and_b32_e32 v1, 0x3ff, v0 s_mov_b32 s12, s13 v_mov_b32_e32 v31, v0 s_mov_b32 s13, s14 s_mov_b32 s14, s15 s_mov_b32 s32, 0 s_waitcnt lgkmcnt(0) s_and_b32 s4, s4, 0xffff v_mov_b32_e32 v0, s34 v_mad_u64_u32 v[40:41], null, s12, s4, v[1:2] v_mov_b32_e32 v1, s35 s_add_u32 s8, s2, 8 s_addc_u32 s9, s3, 0 s_mov_b64 s[4:5], s[0:1] s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mov_b32_e32 v2, v40 s_swappc_b64 s[30:31], 3 v_mov_b32_e32 v41, 0 v_lshlrev_b64 v[1:2], 2, v[40:41] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v1, vcc_lo, s34, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s35, v2, vcc_lo global_store_b32 v[1:2], v0, off s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3fooPf .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 264 .amdhsa_user_sgpr_count 13 .amdhsa_user_sgpr_dispatch_ptr 1 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 1 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 1 .amdhsa_enable_private_segment 1 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 1 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 2 .amdhsa_next_free_vgpr 42 .amdhsa_next_free_sgpr 36 .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 _Z3fooPf, .Lfunc_end0-_Z3fooPf .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: hidden_block_count_x - .offset: 12 .size: 4 .value_kind: hidden_block_count_y - .offset: 16 .size: 4 .value_kind: hidden_block_count_z - .offset: 20 .size: 2 .value_kind: hidden_group_size_x - .offset: 22 .size: 2 .value_kind: hidden_group_size_y - .offset: 24 .size: 2 .value_kind: hidden_group_size_z - .offset: 26 .size: 2 .value_kind: hidden_remainder_x - .offset: 28 .size: 2 .value_kind: hidden_remainder_y - .offset: 30 .size: 2 .value_kind: hidden_remainder_z - .offset: 48 .size: 8 .value_kind: hidden_global_offset_x - .offset: 56 .size: 8 .value_kind: hidden_global_offset_y - .offset: 64 .size: 8 .value_kind: hidden_global_offset_z - .offset: 72 .size: 2 .value_kind: hidden_grid_dims - .offset: 88 .size: 8 .value_kind: hidden_hostcall_buffer - .offset: 96 .size: 8 .value_kind: hidden_multigrid_sync_arg - .offset: 104 .size: 8 .value_kind: hidden_heap_v1 - .offset: 112 .size: 8 .value_kind: hidden_default_queue - .offset: 208 .size: 8 .value_kind: hidden_queue_ptr .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 264 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3fooPf .private_segment_fixed_size: 0 .sgpr_count: 38 .sgpr_spill_count: 0 .symbol: _Z3fooPf.kd .uniform_work_group_size: 1 .uses_dynamic_stack: true .vgpr_count: 42 .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_0002196b_00000000-6_main.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2061: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2061: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z13multiplyByTwoPfj .type _Z13multiplyByTwoPfj, @function _Z13multiplyByTwoPfj: .LFB2057: .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 .LFE2057: .size _Z13multiplyByTwoPfj, .-_Z13multiplyByTwoPfj .globl _Z11divideByTwoPfj .type _Z11divideByTwoPfj, @function _Z11divideByTwoPfj: .LFB2058: .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 .LFE2058: .size _Z11divideByTwoPfj, .-_Z11divideByTwoPfj .globl _Z22__device_stub__Z3fooPfPf .type _Z22__device_stub__Z3fooPfPf, @function _Z22__device_stub__Z3fooPfPf: .LFB2083: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 88(%rsp), %rax subq %fs:40, %rax jne .L12 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z3fooPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z22__device_stub__Z3fooPfPf, .-_Z22__device_stub__Z3fooPfPf .globl _Z3fooPf .type _Z3fooPf, @function _Z3fooPf: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z22__device_stub__Z3fooPfPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z3fooPf, .-_Z3fooPf .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3fooPf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3fooPf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "main.hip" .globl _Z18__device_stub__fooPf # -- Begin function _Z18__device_stub__fooPf .p2align 4, 0x90 .type _Z18__device_stub__fooPf,@function _Z18__device_stub__fooPf: # @_Z18__device_stub__fooPf .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 $_Z3fooPf, %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 _Z18__device_stub__fooPf, .Lfunc_end0-_Z18__device_stub__fooPf .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 $_Z3fooPf, %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 _Z3fooPf,@object # @_Z3fooPf .section .rodata,"a",@progbits .globl _Z3fooPf .p2align 3, 0x0 _Z3fooPf: .quad _Z18__device_stub__fooPf .size _Z3fooPf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3fooPf" .size .L__unnamed_1, 9 .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__fooPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3fooPf .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include "curand_kernel.h" #include <cuda.h> #include <curand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; curandGenerator_t gen; curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); curandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; cudaMalloc(&result, size); cudaMalloc(&x, size); cudaMalloc(&y, size); curandGenerateUniformDouble(gen, x, steps); curandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; cudaMemcpy(check, result, size, cudaMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; cudaFree(result); cudaFree(x); cudaFree(y); return 0; }
code for sm_80 Function : _Z14belongs_circlePdS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GT.AND P0, PT, R0, 0x1fffff, PT ; /* 0x001fffff0000780c */ /* 0x000fda0003f04270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R11, -RZ, RZ, 0, 4.76837158203125e-07 ; /* 0x00000008ff0b7435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R6, R0, R11, c[0x0][0x168] ; /* 0x00005a0000067625 */ /* 0x000fc800078e020b */ /*0090*/ IMAD.WIDE R2, R0, R11, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x000fe400078e020b */ /*00a0*/ LDG.E.64 R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea8000c1e1b00 */ /*00b0*/ LDG.E.64 R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ee2000c1e1b00 */ /*00c0*/ DADD R8, R6, -0.5 ; /* 0xbfe0000006087429 */ /* 0x004e080000000000 */ /*00d0*/ DADD R4, R2, -0.5 ; /* 0xbfe0000002047429 */ /* 0x008fc80000000000 */ /*00e0*/ DMUL R8, R8, R8 ; /* 0x0000000808087228 */ /* 0x001e0c0000000000 */ /*00f0*/ DFMA R8, R4, R4, R8 ; /* 0x000000040408722b */ /* 0x0010640000000008 */ /*0100*/ IMAD.WIDE R4, R0, R11, c[0x0][0x170] ; /* 0x00005c0000047625 */ /* 0x001fc800078e020b */ /*0110*/ DSETP.GTU.AND P0, PT, R8, 0.25, PT ; /* 0x3fd000000800742a */ /* 0x002e1c0003f0c000 */ /*0120*/ @!P0 MOV R8, 0x0 ; /* 0x0000000000088802 */ /* 0x001fe40000000f00 */ /*0130*/ @!P0 MOV R9, 0x3ff00000 ; /* 0x3ff0000000098802 */ /* 0x000fca0000000f00 */ /*0140*/ @!P0 STG.E.64 [R4.64], R8 ; /* 0x0000000804008986 */ /* 0x0001e2000c101b04 */ /*0150*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0160*/ STG.E.64 [R4.64], RZ ; /* 0x000000ff04007986 */ /* 0x000fe2000c101b04 */ /*0170*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0180*/ BRA 0x180; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ /*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.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include "curand_kernel.h" #include <cuda.h> #include <curand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; curandGenerator_t gen; curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); curandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; cudaMalloc(&result, size); cudaMalloc(&x, size); cudaMalloc(&y, size); curandGenerateUniformDouble(gen, x, steps); curandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; cudaMemcpy(check, result, size, cudaMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; cudaFree(result); cudaFree(x); cudaFree(y); return 0; }
.file "tmpxft_0005e9bb_00000000-6_Pi.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3925: .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 .LFE3925: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ .type _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_, @function _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_: .LFB3947: .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 .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 _Z14belongs_circlePdS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3947: .size _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_, .-_Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ .globl _Z14belongs_circlePdS_S_ .type _Z14belongs_circlePdS_S_, @function _Z14belongs_circlePdS_S_: .LFB3948: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3948: .size _Z14belongs_circlePdS_S_, .-_Z14belongs_circlePdS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Sum is " .LC2: .string "Pi is " .text .globl main .type main, @function main: .LFB3922: .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 leaq -16777216(%rsp), %r11 .cfi_def_cfa 11, 16777240 .LPSRL0: subq $4096, %rsp orq $0, (%rsp) cmpq %r11, %rsp jne .LPSRL0 .cfi_def_cfa_register 7 subq $88, %rsp .cfi_def_cfa_offset 16777328 movq %fs:40, %rax movq %rax, 16777288(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $100, %esi call curandCreateGenerator@PLT movl $1234, %esi movq 8(%rsp), %rdi call curandSetPseudoRandomGeneratorSeed@PLT leaq 16(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT movl $2097152, %edx movq 24(%rsp), %rsi movq 8(%rsp), %rdi call curandGenerateUniformDouble@PLT movl $2097152, %edx movq 32(%rsp), %rsi movq 8(%rsp), %rdi call curandGenerateUniformDouble@PLT movl $256, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $8192, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $0, %r9d movl $0, %r8d movq 52(%rsp), %rdx movl $1, %ecx movq 40(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L17 .L12: leaq 64(%rsp), %rbx movl $2, %ecx movl $16777216, %edx movq 16(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rax leaq 16777280(%rsp), %rdx movl $0x000000000, %ebx .L13: movq %rbx, %xmm1 addsd (%rax), %xmm1 movq %xmm1, %rbx addq $8, %rax cmpq %rdx, %rax jne .L13 leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rbp movq %rbp, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %xmm0 call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT leaq .LC2(%rip), %rsi movq %rbp, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %xmm0 mulsd .LC3(%rip), %xmm0 mulsd .LC4(%rip), %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 32(%rsp), %rdi call cudaFree@PLT movq 16777288(%rsp), %rax subq %fs:40, %rax jne .L18 movl $0, %eax addq $16777304, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state movq 16(%rsp), %rdx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ jmp .L12 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE3922: .size main, .-main .section .rodata.str1.1 .LC5: .string "_Z14belongs_circlePdS_S_" .LC6: .string "precalc_xorwow_matrix" .LC7: .string "precalc_xorwow_offset_matrix" .LC8: .string "mrg32k3aM1" .LC9: .string "mrg32k3aM2" .LC10: .string "mrg32k3aM1SubSeq" .LC11: .string "mrg32k3aM2SubSeq" .LC12: .string "mrg32k3aM1Seq" .LC13: .string "mrg32k3aM2Seq" .LC14: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3950: .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 .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z14belongs_circlePdS_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 movl $102400, %r9d movl $0, %r8d leaq .LC6(%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 .LC7(%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 .LC8(%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 .LC9(%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 .LC10(%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 .LC11(%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 .LC12(%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 .LC13(%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 .LC14(%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 .LFE3950: .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 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC3: .long 0 .long 1074790400 .align 8 .LC4: .long 0 .long 1050673152 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "cuda_runtime.h" #include "device_launch_parameters.h" #include "curand_kernel.h" #include <cuda.h> #include <curand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; curandGenerator_t gen; curandCreateGenerator(&gen, CURAND_RNG_PSEUDO_DEFAULT); curandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; cudaMalloc(&result, size); cudaMalloc(&x, size); cudaMalloc(&y, size); curandGenerateUniformDouble(gen, x, steps); curandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; cudaMemcpy(check, result, size, cudaMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; cudaFree(result); cudaFree(x); cudaFree(y); return 0; }
#include "hip/hip_runtime.h" #include "hiprand/hiprand_kernel.h" #include <hiprand/hiprand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_DEFAULT); hiprandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; hipMalloc(&result, size); hipMalloc(&x, size); hipMalloc(&y, size); hiprandGenerateUniformDouble(gen, x, steps); hiprandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; hipMemcpy(check, result, size, hipMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; hipFree(result); hipFree(x); hipFree(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 "hiprand/hiprand_kernel.h" #include <hiprand/hiprand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_DEFAULT); hiprandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; hipMalloc(&result, size); hipMalloc(&x, size); hipMalloc(&y, size); hiprandGenerateUniformDouble(gen, x, steps); hiprandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; hipMemcpy(check, result, size, hipMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; hipFree(result); hipFree(x); hipFree(y); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14belongs_circlePdS_S_ .globl _Z14belongs_circlePdS_S_ .p2align 8 .type _Z14belongs_circlePdS_S_,@function _Z14belongs_circlePdS_S_: s_load_b32 s2, s[0:1], 0x24 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_e32 0x200000, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s6, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s4, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v1, vcc_lo global_load_b64 v[2:3], v[2:3], off v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b64 v[4:5], v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(1) v_add_f64 v[2:3], v[2:3], -0.5 s_waitcnt vmcnt(0) v_add_f64 v[4:5], v[4:5], -0.5 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[2:3], v[2:3], v[2:3] v_fma_f64 v[2:3], v[4:5], v[4:5], v[2:3] v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_2) v_cmp_nge_f64_e32 vcc_lo, 0x3fd00000, v[2:3] v_cndmask_b32_e64 v5, 0x3ff00000, 0, vcc_lo global_store_b64 v[0:1], v[4:5], off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14belongs_circlePdS_S_ .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 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z14belongs_circlePdS_S_, .Lfunc_end0-_Z14belongs_circlePdS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14belongs_circlePdS_S_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z14belongs_circlePdS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include "hip/hip_runtime.h" #include "hiprand/hiprand_kernel.h" #include <hiprand/hiprand.h> #include <iostream> #include <numeric> using namespace std; const long steps = 1 << 21; __global__ void belongs_circle(double* x, double* y, double* result) { const int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid >= steps) return; if (((x[tid] - 0.5) * (x[tid] - 0.5)) + ((y[tid] - 0.5) * (y[tid] - 0.5)) <= (0.5 * 0.5)) { result[tid] = 1; } else { result[tid] = 0; } } int main() { const long size = steps * sizeof(double); long blockSize = 256; long numBlocks = (steps + blockSize - 1) / blockSize; hiprandGenerator_t gen; hiprandCreateGenerator(&gen, HIPRAND_RNG_PSEUDO_DEFAULT); hiprandSetPseudoRandomGeneratorSeed(gen, 1234ULL); double *result, *x, *y; hipMalloc(&result, size); hipMalloc(&x, size); hipMalloc(&y, size); hiprandGenerateUniformDouble(gen, x, steps); hiprandGenerateUniformDouble(gen, y, steps); belongs_circle <<<numBlocks, blockSize>>>(x, y, result); double check[steps]; hipMemcpy(check, result, size, hipMemcpyDeviceToHost); double sum = 0; for (long i = 0; i < steps; ++i) { sum += check[i]; } cout << "Sum is " << sum << endl; cout << "Pi is " << 4 * sum / steps << endl; hipFree(result); hipFree(x); hipFree(y); return 0; }
.text .file "Pi.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z29__device_stub__belongs_circlePdS_S_ # -- Begin function _Z29__device_stub__belongs_circlePdS_S_ .p2align 4, 0x90 .type _Z29__device_stub__belongs_circlePdS_S_,@function _Z29__device_stub__belongs_circlePdS_S_: # @_Z29__device_stub__belongs_circlePdS_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 $_Z14belongs_circlePdS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z29__device_stub__belongs_circlePdS_S_, .Lfunc_end0-_Z29__device_stub__belongs_circlePdS_S_ .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI1_0: .quad 0x4010000000000000 # double 4 .LCPI1_1: .quad 0x3ea0000000000000 # double 4.76837158203125E-7 .text .globl 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 $16777336, %rsp # imm = 0x1000078 .cfi_def_cfa_offset 16777360 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 24(%rsp), %rdi movl $400, %esi # imm = 0x190 callq hiprandCreateGenerator movq 24(%rsp), %rdi movl $1234, %esi # imm = 0x4D2 callq hiprandSetPseudoRandomGeneratorSeed leaq 16(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc leaq 8(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc movq %rsp, %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc movq 24(%rsp), %rdi movq 8(%rsp), %rsi movl $2097152, %edx # imm = 0x200000 callq hiprandGenerateUniformDouble movq 24(%rsp), %rdi movq (%rsp), %rsi movl $2097152, %edx # imm = 0x200000 callq hiprandGenerateUniformDouble movabsq $4294967552, %rdx # imm = 0x100000100 leaq 7936(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 8(%rsp), %rax movq (%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z14belongs_circlePdS_S_, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: movq 16(%rsp), %rsi leaq 112(%rsp), %rdi movl $16777216, %edx # imm = 0x1000000 movl $2, %ecx callq hipMemcpy xorpd %xmm0, %xmm0 xorl %eax, %eax .p2align 4, 0x90 .LBB1_3: # =>This Inner Loop Header: Depth=1 addsd 112(%rsp,%rax,8), %xmm0 incq %rax cmpq $2097152, %rax # imm = 0x200000 jne .LBB1_3 # %bb.4: movl $_ZSt4cout, %edi movl $.L.str, %esi movl $7, %edx movsd %xmm0, 32(%rsp) # 8-byte Spill callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movl $_ZSt4cout, %edi movsd 32(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB1_13 # %bb.5: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB1_7 # %bb.6: movzbl 67(%rbx), %ecx jmp .LBB1_8 .LBB1_7: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB1_8: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $6, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movsd 32(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero mulsd .LCPI1_0(%rip), %xmm0 mulsd .LCPI1_1(%rip), %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB1_13 # %bb.9: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i13 cmpb $0, 56(%rbx) je .LBB1_11 # %bb.10: movzbl 67(%rbx), %ecx jmp .LBB1_12 .LBB1_11: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB1_12: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit16 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 (%rsp), %rdi callq hipFree xorl %eax, %eax addq $16777336, %rsp # imm = 0x1000078 .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .LBB1_13: .cfi_def_cfa_offset 16777360 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 $_Z14belongs_circlePdS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z14belongs_circlePdS_S_,@object # @_Z14belongs_circlePdS_S_ .section .rodata,"a",@progbits .globl _Z14belongs_circlePdS_S_ .p2align 3, 0x0 _Z14belongs_circlePdS_S_: .quad _Z29__device_stub__belongs_circlePdS_S_ .size _Z14belongs_circlePdS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Sum is " .size .L.str, 8 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Pi is " .size .L.str.1, 7 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z14belongs_circlePdS_S_" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z29__device_stub__belongs_circlePdS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14belongs_circlePdS_S_ .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 : _Z14belongs_circlePdS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GT.AND P0, PT, R0, 0x1fffff, PT ; /* 0x001fffff0000780c */ /* 0x000fda0003f04270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R11, -RZ, RZ, 0, 4.76837158203125e-07 ; /* 0x00000008ff0b7435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R6, R0, R11, c[0x0][0x168] ; /* 0x00005a0000067625 */ /* 0x000fc800078e020b */ /*0090*/ IMAD.WIDE R2, R0, R11, c[0x0][0x160] ; /* 0x0000580000027625 */ /* 0x000fe400078e020b */ /*00a0*/ LDG.E.64 R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea8000c1e1b00 */ /*00b0*/ LDG.E.64 R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ee2000c1e1b00 */ /*00c0*/ DADD R8, R6, -0.5 ; /* 0xbfe0000006087429 */ /* 0x004e080000000000 */ /*00d0*/ DADD R4, R2, -0.5 ; /* 0xbfe0000002047429 */ /* 0x008fc80000000000 */ /*00e0*/ DMUL R8, R8, R8 ; /* 0x0000000808087228 */ /* 0x001e0c0000000000 */ /*00f0*/ DFMA R8, R4, R4, R8 ; /* 0x000000040408722b */ /* 0x0010640000000008 */ /*0100*/ IMAD.WIDE R4, R0, R11, c[0x0][0x170] ; /* 0x00005c0000047625 */ /* 0x001fc800078e020b */ /*0110*/ DSETP.GTU.AND P0, PT, R8, 0.25, PT ; /* 0x3fd000000800742a */ /* 0x002e1c0003f0c000 */ /*0120*/ @!P0 MOV R8, 0x0 ; /* 0x0000000000088802 */ /* 0x001fe40000000f00 */ /*0130*/ @!P0 MOV R9, 0x3ff00000 ; /* 0x3ff0000000098802 */ /* 0x000fca0000000f00 */ /*0140*/ @!P0 STG.E.64 [R4.64], R8 ; /* 0x0000000804008986 */ /* 0x0001e2000c101b04 */ /*0150*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0160*/ STG.E.64 [R4.64], RZ ; /* 0x000000ff04007986 */ /* 0x000fe2000c101b04 */ /*0170*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0180*/ BRA 0x180; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ /*0200*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0210*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0220*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0230*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0240*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0250*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0260*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0270*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z14belongs_circlePdS_S_ .globl _Z14belongs_circlePdS_S_ .p2align 8 .type _Z14belongs_circlePdS_S_,@function _Z14belongs_circlePdS_S_: s_load_b32 s2, s[0:1], 0x24 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_e32 0x200000, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s6, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s4, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v1, vcc_lo global_load_b64 v[2:3], v[2:3], off v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b64 v[4:5], v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(1) v_add_f64 v[2:3], v[2:3], -0.5 s_waitcnt vmcnt(0) v_add_f64 v[4:5], v[4:5], -0.5 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_f64 v[2:3], v[2:3], v[2:3] v_fma_f64 v[2:3], v[4:5], v[4:5], v[2:3] v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_2) v_cmp_nge_f64_e32 vcc_lo, 0x3fd00000, v[2:3] v_cndmask_b32_e64 v5, 0x3ff00000, 0, vcc_lo global_store_b64 v[0:1], v[4:5], off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z14belongs_circlePdS_S_ .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 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z14belongs_circlePdS_S_, .Lfunc_end0-_Z14belongs_circlePdS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z14belongs_circlePdS_S_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z14belongs_circlePdS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0005e9bb_00000000-6_Pi.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3925: .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 .LFE3925: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ .type _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_, @function _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_: .LFB3947: .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 .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 _Z14belongs_circlePdS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3947: .size _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_, .-_Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ .globl _Z14belongs_circlePdS_S_ .type _Z14belongs_circlePdS_S_, @function _Z14belongs_circlePdS_S_: .LFB3948: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3948: .size _Z14belongs_circlePdS_S_, .-_Z14belongs_circlePdS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Sum is " .LC2: .string "Pi is " .text .globl main .type main, @function main: .LFB3922: .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 leaq -16777216(%rsp), %r11 .cfi_def_cfa 11, 16777240 .LPSRL0: subq $4096, %rsp orq $0, (%rsp) cmpq %r11, %rsp jne .LPSRL0 .cfi_def_cfa_register 7 subq $88, %rsp .cfi_def_cfa_offset 16777328 movq %fs:40, %rax movq %rax, 16777288(%rsp) xorl %eax, %eax leaq 8(%rsp), %rdi movl $100, %esi call curandCreateGenerator@PLT movl $1234, %esi movq 8(%rsp), %rdi call curandSetPseudoRandomGeneratorSeed@PLT leaq 16(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT movl $2097152, %edx movq 24(%rsp), %rsi movq 8(%rsp), %rdi call curandGenerateUniformDouble@PLT movl $2097152, %edx movq 32(%rsp), %rsi movq 8(%rsp), %rdi call curandGenerateUniformDouble@PLT movl $256, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $8192, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $0, %r9d movl $0, %r8d movq 52(%rsp), %rdx movl $1, %ecx movq 40(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L17 .L12: leaq 64(%rsp), %rbx movl $2, %ecx movl $16777216, %edx movq 16(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rax leaq 16777280(%rsp), %rdx movl $0x000000000, %ebx .L13: movq %rbx, %xmm1 addsd (%rax), %xmm1 movq %xmm1, %rbx addq $8, %rax cmpq %rdx, %rax jne .L13 leaq .LC1(%rip), %rsi leaq _ZSt4cout(%rip), %rbp movq %rbp, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %xmm0 call _ZNSo9_M_insertIdEERSoT_@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT leaq .LC2(%rip), %rsi movq %rbp, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi movq %rbx, %xmm0 mulsd .LC3(%rip), %xmm0 mulsd .LC4(%rip), %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 32(%rsp), %rdi call cudaFree@PLT movq 16777288(%rsp), %rax subq %fs:40, %rax jne .L18 movl $0, %eax addq $16777304, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state movq 16(%rsp), %rdx movq 32(%rsp), %rsi movq 24(%rsp), %rdi call _Z38__device_stub__Z14belongs_circlePdS_S_PdS_S_ jmp .L12 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE3922: .size main, .-main .section .rodata.str1.1 .LC5: .string "_Z14belongs_circlePdS_S_" .LC6: .string "precalc_xorwow_matrix" .LC7: .string "precalc_xorwow_offset_matrix" .LC8: .string "mrg32k3aM1" .LC9: .string "mrg32k3aM2" .LC10: .string "mrg32k3aM1SubSeq" .LC11: .string "mrg32k3aM2SubSeq" .LC12: .string "mrg32k3aM1Seq" .LC13: .string "mrg32k3aM2Seq" .LC14: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3950: .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 .LC5(%rip), %rdx movq %rdx, %rcx leaq _Z14belongs_circlePdS_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 movl $102400, %r9d movl $0, %r8d leaq .LC6(%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 .LC7(%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 .LC8(%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 .LC9(%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 .LC10(%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 .LC11(%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 .LC12(%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 .LC13(%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 .LC14(%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 .LFE3950: .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 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC3: .long 0 .long 1074790400 .align 8 .LC4: .long 0 .long 1050673152 .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 "Pi.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z29__device_stub__belongs_circlePdS_S_ # -- Begin function _Z29__device_stub__belongs_circlePdS_S_ .p2align 4, 0x90 .type _Z29__device_stub__belongs_circlePdS_S_,@function _Z29__device_stub__belongs_circlePdS_S_: # @_Z29__device_stub__belongs_circlePdS_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 $_Z14belongs_circlePdS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z29__device_stub__belongs_circlePdS_S_, .Lfunc_end0-_Z29__device_stub__belongs_circlePdS_S_ .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI1_0: .quad 0x4010000000000000 # double 4 .LCPI1_1: .quad 0x3ea0000000000000 # double 4.76837158203125E-7 .text .globl 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 $16777336, %rsp # imm = 0x1000078 .cfi_def_cfa_offset 16777360 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 24(%rsp), %rdi movl $400, %esi # imm = 0x190 callq hiprandCreateGenerator movq 24(%rsp), %rdi movl $1234, %esi # imm = 0x4D2 callq hiprandSetPseudoRandomGeneratorSeed leaq 16(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc leaq 8(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc movq %rsp, %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc movq 24(%rsp), %rdi movq 8(%rsp), %rsi movl $2097152, %edx # imm = 0x200000 callq hiprandGenerateUniformDouble movq 24(%rsp), %rdi movq (%rsp), %rsi movl $2097152, %edx # imm = 0x200000 callq hiprandGenerateUniformDouble movabsq $4294967552, %rdx # imm = 0x100000100 leaq 7936(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movq 8(%rsp), %rax movq (%rsp), %rcx movq 16(%rsp), %rdx movq %rax, 104(%rsp) movq %rcx, 96(%rsp) movq %rdx, 88(%rsp) leaq 104(%rsp), %rax movq %rax, 112(%rsp) leaq 96(%rsp), %rax movq %rax, 120(%rsp) leaq 88(%rsp), %rax movq %rax, 128(%rsp) leaq 72(%rsp), %rdi leaq 56(%rsp), %rsi leaq 48(%rsp), %rdx leaq 40(%rsp), %rcx callq __hipPopCallConfiguration movq 72(%rsp), %rsi movl 80(%rsp), %edx movq 56(%rsp), %rcx movl 64(%rsp), %r8d leaq 112(%rsp), %r9 movl $_Z14belongs_circlePdS_S_, %edi pushq 40(%rsp) .cfi_adjust_cfa_offset 8 pushq 56(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: movq 16(%rsp), %rsi leaq 112(%rsp), %rdi movl $16777216, %edx # imm = 0x1000000 movl $2, %ecx callq hipMemcpy xorpd %xmm0, %xmm0 xorl %eax, %eax .p2align 4, 0x90 .LBB1_3: # =>This Inner Loop Header: Depth=1 addsd 112(%rsp,%rax,8), %xmm0 incq %rax cmpq $2097152, %rax # imm = 0x200000 jne .LBB1_3 # %bb.4: movl $_ZSt4cout, %edi movl $.L.str, %esi movl $7, %edx movsd %xmm0, 32(%rsp) # 8-byte Spill callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movl $_ZSt4cout, %edi movsd 32(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB1_13 # %bb.5: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i cmpb $0, 56(%rbx) je .LBB1_7 # %bb.6: movzbl 67(%rbx), %ecx jmp .LBB1_8 .LBB1_7: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB1_8: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv movl $_ZSt4cout, %edi movl $.L.str.1, %esi movl $6, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l movsd 32(%rsp), %xmm0 # 8-byte Reload # xmm0 = mem[0],zero mulsd .LCPI1_0(%rip), %xmm0 mulsd .LCPI1_1(%rip), %xmm0 movl $_ZSt4cout, %edi callq _ZNSo9_M_insertIdEERSoT_ movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %rbx testq %rbx, %rbx je .LBB1_13 # %bb.9: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i13 cmpb $0, 56(%rbx) je .LBB1_11 # %bb.10: movzbl 67(%rbx), %ecx jmp .LBB1_12 .LBB1_11: movq %rbx, %rdi movq %rax, %r14 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%rbx), %rax movq %rbx, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r14, %rax .LBB1_12: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit16 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 (%rsp), %rdi callq hipFree xorl %eax, %eax addq $16777336, %rsp # imm = 0x1000078 .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .LBB1_13: .cfi_def_cfa_offset 16777360 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 $_Z14belongs_circlePdS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z14belongs_circlePdS_S_,@object # @_Z14belongs_circlePdS_S_ .section .rodata,"a",@progbits .globl _Z14belongs_circlePdS_S_ .p2align 3, 0x0 _Z14belongs_circlePdS_S_: .quad _Z29__device_stub__belongs_circlePdS_S_ .size _Z14belongs_circlePdS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Sum is " .size .L.str, 8 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Pi is " .size .L.str.1, 7 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z14belongs_circlePdS_S_" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z29__device_stub__belongs_circlePdS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z14belongs_circlePdS_S_ .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> #include <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); 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 <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); printf("------------------------\n"); return(0); }
.file "tmpxft_000e367c_00000000-6_01hello_pthreads.cudafe1.cpp" .text #APP .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "HelloThread %d\n" #NO_APP .text .globl _Z5helloPv .type _Z5helloPv, @function _Z5helloPv: .LFB2080: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movq %rdi, %rbx movl $5, %edi call sleep@PLT movl (%rbx), %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %eax popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2080: .size _Z5helloPv, .-_Z5helloPv .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2084: .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 .LFE2084: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1 .LC1: .string "------------------------\n" .text .globl main .type main, @function main: .LFB2081: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movl $1, 12(%rsp) leaq 12(%rsp), %rcx leaq 16(%rsp), %rdi leaq _Z5helloPv(%rip), %rdx movl $0, %esi call pthread_create@PLT movl $0, %esi movq 16(%rsp), %rdi call pthread_join@PLT leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 24(%rsp), %rax subq %fs:40, %rax jne .L8 movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L8: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2107: .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 .LFE2107: .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 <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); printf("------------------------\n"); return(0); }
#include <hip/hip_runtime.h> #include <stdio.h> #include <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); 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 <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); 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 <pthread.h> #include <unistd.h> //sleep void *hello (void *arg) { int *ptrToThreadNumber = (int *)arg; sleep(5); printf("HelloThread %d\n",*ptrToThreadNumber); return 0; } int main(void) { pthread_t tid; int threadNum=1; //pthread_create creates a new thread and makes it executable. //This routine can be called any number of times from anywhere within your code. pthread_create(&tid,NULL,hello,&threadNum); //"Joining" is one way to accomplish synchronization between threads //The pthread_join() subroutine blocks the calling thread until the //specified threadid thread terminates. pthread_join(tid,NULL); printf("------------------------\n"); return(0); }
.text .file "01hello_pthreads.hip" .globl _Z5helloPv # -- Begin function _Z5helloPv .p2align 4, 0x90 .type _Z5helloPv,@function _Z5helloPv: # @_Z5helloPv .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 movq %rdi, %rbx movl $5, %edi callq sleep movl (%rbx), %esi movl $.L.str, %edi xorl %eax, %eax callq printf xorl %eax, %eax popq %rbx .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z5helloPv, .Lfunc_end0-_Z5helloPv .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 movl $1, 12(%rsp) leaq 16(%rsp), %rdi leaq 12(%rsp), %rcx movl $_Z5helloPv, %edx xorl %esi, %esi callq pthread_create movq 16(%rsp), %rdi xorl %esi, %esi callq pthread_join movl $.Lstr, %edi callq puts@PLT xorl %eax, %eax addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "HelloThread %d\n" .size .L.str, 16 .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 "------------------------" .size .Lstr, 25 .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 _Z5helloPv .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_000e367c_00000000-6_01hello_pthreads.cudafe1.cpp" .text #APP .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "HelloThread %d\n" #NO_APP .text .globl _Z5helloPv .type _Z5helloPv, @function _Z5helloPv: .LFB2080: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 movq %rdi, %rbx movl $5, %edi call sleep@PLT movl (%rbx), %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %eax popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2080: .size _Z5helloPv, .-_Z5helloPv .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2084: .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 .LFE2084: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1 .LC1: .string "------------------------\n" .text .globl main .type main, @function main: .LFB2081: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movl $1, 12(%rsp) leaq 12(%rsp), %rcx leaq 16(%rsp), %rdi leaq _Z5helloPv(%rip), %rdx movl $0, %esi call pthread_create@PLT movl $0, %esi movq 16(%rsp), %rdi call pthread_join@PLT leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 24(%rsp), %rax subq %fs:40, %rax jne .L8 movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L8: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2107: .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 .LFE2107: .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 "01hello_pthreads.hip" .globl _Z5helloPv # -- Begin function _Z5helloPv .p2align 4, 0x90 .type _Z5helloPv,@function _Z5helloPv: # @_Z5helloPv .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 movq %rdi, %rbx movl $5, %edi callq sleep movl (%rbx), %esi movl $.L.str, %edi xorl %eax, %eax callq printf xorl %eax, %eax popq %rbx .cfi_def_cfa_offset 8 retq .Lfunc_end0: .size _Z5helloPv, .Lfunc_end0-_Z5helloPv .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 movl $1, 12(%rsp) leaq 16(%rsp), %rdi leaq 12(%rsp), %rcx movl $_Z5helloPv, %edx xorl %esi, %esi callq pthread_create movq 16(%rsp), %rdi xorl %esi, %esi callq pthread_join movl $.Lstr, %edi callq puts@PLT xorl %eax, %eax addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "HelloThread %d\n" .size .L.str, 16 .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 "------------------------" .size .Lstr, 25 .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 _Z5helloPv .addrsig_sym __hip_cuid_