system
stringclasses
7 values
input
stringlengths
10
94.5k
output
stringlengths
14
55.5k
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> __global__ void matrix(float *A, int numElements, int n) { int i = blockDim.x * blockIdx.x + threadIdx.x; int j = blockDim.y * blockIdx.y + threadIdx.y; float a; if (i < n && j< n && i!= (n-1) && i%2==0) { a = A[j*n + i]; A[j*n + i] = A[j*n + i + 1]; A[j*n + i +1] = a; } if (i<n&& j<n && i<j){ A[i*n+j] = A[j*n + i]; } }
.text .file "kernel.hip" .globl _Z21__device_stub__matrixPfii # -- Begin function _Z21__device_stub__matrixPfii .p2align 4, 0x90 .type _Z21__device_stub__matrixPfii,@function _Z21__device_stub__matrixPfii: # @_Z21__device_stub__matrixPfii .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 $_Z6matrixPfii, %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 _Z21__device_stub__matrixPfii, .Lfunc_end0-_Z21__device_stub__matrixPfii .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 $_Z6matrixPfii, %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 _Z6matrixPfii,@object # @_Z6matrixPfii .section .rodata,"a",@progbits .globl _Z6matrixPfii .p2align 3, 0x0 _Z6matrixPfii: .quad _Z21__device_stub__matrixPfii .size _Z6matrixPfii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z6matrixPfii" .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 _Z21__device_stub__matrixPfii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6matrixPfii .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 : _Z6matrixPfii .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.Y ; /* 0x0000000000057919 */ /* 0x000e220000002600 */ /*0020*/ ULDC UR4, c[0x0][0x16c] ; /* 0x00005b0000047ab9 */ /* 0x000fe20000000800 */ /*0030*/ HFMA2.MMA R4, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff047435 */ /* 0x000fe200000001ff */ /*0040*/ UIADD3 UR4, UR4, -0x1, URZ ; /* 0xffffffff04047890 */ /* 0x000fe2000fffe03f */ /*0050*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e220000002200 */ /*0060*/ BSSY B0, 0x1a0 ; /* 0x0000013000007945 */ /* 0x000fe60003800000 */ /*0070*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e680000002500 */ /*0080*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e620000002100 */ /*0090*/ IMAD R5, R5, c[0x0][0x4], R2 ; /* 0x0000010005057a24 */ /* 0x001fca00078e0202 */ /*00a0*/ ISETP.GE.AND P0, PT, R5, c[0x0][0x16c], PT ; /* 0x00005b0005007a0c */ /* 0x000fe20003f06270 */ /*00b0*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x002fca00078e0203 */ /*00c0*/ ISETP.LT.AND P0, PT, R0.reuse, c[0x0][0x16c], !P0 ; /* 0x00005b0000007a0c */ /* 0x040fe40004701270 */ /*00d0*/ LOP3.LUT R2, R0.reuse, 0x1, RZ, 0xc0, !PT ; /* 0x0000000100027812 */ /* 0x040fe400078ec0ff */ /*00e0*/ ISETP.EQ.OR P1, PT, R0.reuse, UR4, !P0 ; /* 0x0000000400007c0c */ /* 0x040fe2000c722670 */ /*00f0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0100*/ ISETP.GE.OR P0, PT, R0, R5, !P0 ; /* 0x000000050000720c */ /* 0x000fe40004706670 */ /*0110*/ ISETP.EQ.U32.OR P1, PT, R2, 0x1, P1 ; /* 0x000000010200780c */ /* 0x000fe20000f22470 */ /*0120*/ IMAD R2, R5, c[0x0][0x16c], R0 ; /* 0x00005b0005027a24 */ /* 0x000fc800078e0200 */ /*0130*/ IMAD.WIDE R2, R2, R4, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fd000078e0204 */ /*0140*/ @P1 BRA 0x190 ; /* 0x0000004000001947 */ /* 0x000fea0003800000 */ /*0150*/ LDG.E R9, [R2.64+0x4] ; /* 0x0000040402097981 */ /* 0x000ea8000c1e1900 */ /*0160*/ LDG.E R7, [R2.64] ; /* 0x0000000402077981 */ /* 0x000ee8000c1e1900 */ /*0170*/ STG.E [R2.64], R9 ; /* 0x0000000902007986 */ /* 0x0041e8000c101904 */ /*0180*/ STG.E [R2.64+0x4], R7 ; /* 0x0000040702007986 */ /* 0x0081e4000c101904 */ /*0190*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*01a0*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*01b0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x001ea2000c1e1900 */ /*01c0*/ IMAD R5, R0, c[0x0][0x16c], R5 ; /* 0x00005b0000057a24 */ /* 0x000fc800078e0205 */ /*01d0*/ IMAD.WIDE R4, R5, R4, c[0x0][0x160] ; /* 0x0000580005047625 */ /* 0x000fca00078e0204 */ /*01e0*/ STG.E [R4.64], R3 ; /* 0x0000000304007986 */ /* 0x004fe2000c101904 */ /*01f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0200*/ BRA 0x200; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ /*0280*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0290*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*02f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z6matrixPfii .globl _Z6matrixPfii .p2align 8 .type _Z6matrixPfii,@function _Z6matrixPfii: s_clause 0x1 s_load_b32 s2, s[0:1], 0x1c s_load_b32 s4, s[0:1], 0xc v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s3, s2, 0xffff s_lshr_b32 s2, s2, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s14, s3, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s2, v[3:4] s_load_b64 s[2:3], s[0:1], 0x0 s_add_i32 s0, s4, -1 v_and_b32_e32 v3, 1, v0 v_cmp_ne_u32_e64 s0, s0, v0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_max_i32_e32 v2, v0, v1 v_cmp_eq_u32_e64 s1, 0, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_4) v_cmp_gt_i32_e32 vcc_lo, s4, v2 s_and_b32 s0, vcc_lo, s0 s_delay_alu instid0(VALU_DEP_2) | instid1(SALU_CYCLE_1) s_and_b32 s0, s1, s0 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s1, s0 s_cbranch_execz .LBB0_2 v_mad_u64_u32 v[2:3], null, v1, s4, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[2:3], 2, v[2:3] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v2, s0, s2, v2 v_add_co_ci_u32_e64 v3, s0, s3, v3, s0 global_load_b64 v[4:5], v[2:3], off s_waitcnt vmcnt(0) v_mov_b32_e32 v6, v4 global_store_b64 v[2:3], v[5:6], off .LBB0_2: s_or_b32 exec_lo, exec_lo, s1 v_cmp_lt_i32_e64 s0, v0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s0, s0, vcc_lo s_and_saveexec_b32 s1, s0 s_cbranch_execz .LBB0_4 v_mad_u64_u32 v[2:3], null, v1, s4, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[2:3], 2, v[2:3] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v2, vcc_lo, s2, v2 v_add_co_ci_u32_e32 v3, vcc_lo, s3, v3, vcc_lo global_load_b32 v4, v[2:3], off v_mad_u64_u32 v[2:3], null, v0, s4, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[0:1], 2, v[2:3] s_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], v4, off .LBB0_4: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z6matrixPfii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .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 7 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z6matrixPfii, .Lfunc_end0-_Z6matrixPfii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z6matrixPfii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z6matrixPfii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 7 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000d7537_00000000-6_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z6matrixPfiiPfii .type _Z27__device_stub__Z6matrixPfiiPfii, @function _Z27__device_stub__Z6matrixPfiiPfii: .LFB2051: .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 _Z6matrixPfii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z27__device_stub__Z6matrixPfiiPfii, .-_Z27__device_stub__Z6matrixPfiiPfii .globl _Z6matrixPfii .type _Z6matrixPfii, @function _Z6matrixPfii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z6matrixPfiiPfii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z6matrixPfii, .-_Z6matrixPfii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z6matrixPfii" .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 _Z6matrixPfii(%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 "kernel.hip" .globl _Z21__device_stub__matrixPfii # -- Begin function _Z21__device_stub__matrixPfii .p2align 4, 0x90 .type _Z21__device_stub__matrixPfii,@function _Z21__device_stub__matrixPfii: # @_Z21__device_stub__matrixPfii .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 $_Z6matrixPfii, %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 _Z21__device_stub__matrixPfii, .Lfunc_end0-_Z21__device_stub__matrixPfii .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 $_Z6matrixPfii, %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 _Z6matrixPfii,@object # @_Z6matrixPfii .section .rodata,"a",@progbits .globl _Z6matrixPfii .p2align 3, 0x0 _Z6matrixPfii: .quad _Z21__device_stub__matrixPfii .size _Z6matrixPfii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z6matrixPfii" .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 _Z21__device_stub__matrixPfii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6matrixPfii .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.
// Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
// Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
.file "tmpxft_00114ab8_00000000-6_cuda-types.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
// Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
#include <hip/hip_runtime.h> // Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> // Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
.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> // Check that types, widths, __GCC_ATOMIC* macros, etc. match on the host and // device sides of CUDA compilations. Note that we filter out long double, as // this is intentionally different on host and device. // // FIXME: We really should make __GCC_HAVE_SYNC_COMPARE_AND_SWAP identical on // host and device, but architecturally this is difficult at the moment. // RUN: mkdir -p %t // RUN: %clang --cuda-host-only -nocudainc -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-device-defines-filtered // RUN: diff %t/i386-host-defines-filtered %t/i386-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-device-defines-filtered // RUN: diff %t/x86_64-host-defines-filtered %t/x86_64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target powerpc64-unknown-linux-gnu -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/powerpc64-device-defines-filtered // RUN: diff %t/powerpc64-host-defines-filtered %t/powerpc64-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target i386-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/i386-msvc-device-defines-filtered // RUN: diff %t/i386-msvc-host-defines-filtered %t/i386-msvc-device-defines-filtered // RUN: %clang --cuda-host-only -nocudainc -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-host-defines-filtered // RUN: %clang --cuda-device-only -nocudainc -nocudalib -target x86_64-windows-msvc -x cuda -E -dM -o - /dev/null \ // RUN: | grep -E 'define __[^ ]*(TYPE|MAX|SIZEOF|WIDTH)|define __GCC_ATOMIC' \ // RUN: | grep -Ev '__LDBL|_LONG_DOUBLE' > %t/x86_64-msvc-device-defines-filtered // RUN: diff %t/x86_64-msvc-host-defines-filtered %t/x86_64-msvc-device-defines-filtered
.text .file "cuda-types.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00114ab8_00000000-6_cuda-types.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "cuda-types.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
__global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight){ int ia = 16 * blockIdx.x + threadIdx.x; int ib = 16 * blockIdx.y + threadIdx.y; if (ia >= na || ib >= nb) return; proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; } // __global__ void kernel_add(cudaArray *proj1, cudaArray *proj, int iv, int na, int nb, float weight){ // int ia = 16 * blockIdx.x + threadIdx.x; // int ib = 16 * blockIdx.y + threadIdx.y; // if (ia >= na || ib >= nb) // return; // proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; // }
code for sm_80 Function : _Z10kernel_addPfS_iiif .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e280000002600 */ /*0020*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e280000002200 */ /*0030*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e680000002500 */ /*0040*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e620000002100 */ /*0050*/ LEA R3, R3, R2, 0x4 ; /* 0x0000000203037211 */ /* 0x001fc800078e20ff */ /*0060*/ ISETP.GE.AND P0, PT, R3, c[0x0][0x178], PT ; /* 0x00005e0003007a0c */ /* 0x000fe40003f06270 */ /*0070*/ LEA R0, R0, R5, 0x4 ; /* 0x0000000500007211 */ /* 0x002fc800078e20ff */ /*0080*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x174], P0 ; /* 0x00005d0000007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ MOV R2, c[0x0][0x174] ; /* 0x00005d0000027a02 */ /* 0x000fe20000000f00 */ /*00b0*/ IMAD R0, R3, c[0x0][0x174], R0 ; /* 0x00005d0003007a24 */ /* 0x000fe200078e0200 */ /*00c0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*00d0*/ IMAD R3, R2, c[0x0][0x170], RZ ; /* 0x00005c0002037a24 */ /* 0x000fe200078e02ff */ /*00e0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fc600000001ff */ /*00f0*/ IMAD R3, R3, c[0x0][0x178], R0 ; /* 0x00005e0003037a24 */ /* 0x000fce00078e0200 */ /*0100*/ IMAD.WIDE R4, R0, R2, c[0x0][0x160] ; /* 0x0000580000047625 */ /* 0x000fc800078e0202 */ /*0110*/ IMAD.WIDE R2, R3, R2, c[0x0][0x168] ; /* 0x00005a0003027625 */ /* 0x000fe200078e0202 */ /*0120*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000eaa000c1e1900 */ /*0130*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea4000c1e1900 */ /*0140*/ FFMA R7, R2, c[0x0][0x17c], R7 ; /* 0x00005f0002077a23 */ /* 0x004fca0000000007 */ /*0150*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
__global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight){ int ia = 16 * blockIdx.x + threadIdx.x; int ib = 16 * blockIdx.y + threadIdx.y; if (ia >= na || ib >= nb) return; proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; } // __global__ void kernel_add(cudaArray *proj1, cudaArray *proj, int iv, int na, int nb, float weight){ // int ia = 16 * blockIdx.x + threadIdx.x; // int ib = 16 * blockIdx.y + threadIdx.y; // if (ia >= na || ib >= nb) // return; // proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; // }
.file "tmpxft_0009d49f_00000000-6_kernel_add.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 _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif .type _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif, @function _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif: .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) movss %xmm0, (%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) 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 _Z10kernel_addPfS_iiif(%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 _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif, .-_Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif .globl _Z10kernel_addPfS_iiif .type _Z10kernel_addPfS_iiif, @function _Z10kernel_addPfS_iiif: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z10kernel_addPfS_iiif, .-_Z10kernel_addPfS_iiif .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10kernel_addPfS_iiif" .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 _Z10kernel_addPfS_iiif(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
__global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight){ int ia = 16 * blockIdx.x + threadIdx.x; int ib = 16 * blockIdx.y + threadIdx.y; if (ia >= na || ib >= nb) return; proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; } // __global__ void kernel_add(cudaArray *proj1, cudaArray *proj, int iv, int na, int nb, float weight){ // int ia = 16 * blockIdx.x + threadIdx.x; // int ib = 16 * blockIdx.y + threadIdx.y; // if (ia >= na || ib >= nb) // return; // proj1[ia + ib * na] += proj[ia + ib * na + iv * na * nb] * weight; // }
#include <hip/hip_runtime.h> __global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight);
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> __global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight);
.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> __global__ void kernel_add(float *proj1, float *proj, int iv, int na, int nb, float weight);
.text .file "kernel_add.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z10kernel_addPfS_iiif .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e280000002600 */ /*0020*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e280000002200 */ /*0030*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e680000002500 */ /*0040*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e620000002100 */ /*0050*/ LEA R3, R3, R2, 0x4 ; /* 0x0000000203037211 */ /* 0x001fc800078e20ff */ /*0060*/ ISETP.GE.AND P0, PT, R3, c[0x0][0x178], PT ; /* 0x00005e0003007a0c */ /* 0x000fe40003f06270 */ /*0070*/ LEA R0, R0, R5, 0x4 ; /* 0x0000000500007211 */ /* 0x002fc800078e20ff */ /*0080*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x174], P0 ; /* 0x00005d0000007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ MOV R2, c[0x0][0x174] ; /* 0x00005d0000027a02 */ /* 0x000fe20000000f00 */ /*00b0*/ IMAD R0, R3, c[0x0][0x174], R0 ; /* 0x00005d0003007a24 */ /* 0x000fe200078e0200 */ /*00c0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*00d0*/ IMAD R3, R2, c[0x0][0x170], RZ ; /* 0x00005c0002037a24 */ /* 0x000fe200078e02ff */ /*00e0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fc600000001ff */ /*00f0*/ IMAD R3, R3, c[0x0][0x178], R0 ; /* 0x00005e0003037a24 */ /* 0x000fce00078e0200 */ /*0100*/ IMAD.WIDE R4, R0, R2, c[0x0][0x160] ; /* 0x0000580000047625 */ /* 0x000fc800078e0202 */ /*0110*/ IMAD.WIDE R2, R3, R2, c[0x0][0x168] ; /* 0x00005a0003027625 */ /* 0x000fe200078e0202 */ /*0120*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000eaa000c1e1900 */ /*0130*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea4000c1e1900 */ /*0140*/ FFMA R7, R2, c[0x0][0x17c], R7 ; /* 0x00005f0002077a23 */ /* 0x004fca0000000007 */ /*0150*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*0160*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0170*/ BRA 0x170; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .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_0009d49f_00000000-6_kernel_add.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 _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif .type _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif, @function _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif: .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) movss %xmm0, (%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) 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 _Z10kernel_addPfS_iiif(%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 _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif, .-_Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif .globl _Z10kernel_addPfS_iiif .type _Z10kernel_addPfS_iiif, @function _Z10kernel_addPfS_iiif: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z36__device_stub__Z10kernel_addPfS_iiifPfS_iiif addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z10kernel_addPfS_iiif, .-_Z10kernel_addPfS_iiif .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10kernel_addPfS_iiif" .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 _Z10kernel_addPfS_iiif(%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 "kernel_add.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void ConvolutionRowGPU(double *d_Dst, double *d_Src, double *d_Filter, int imageW, int imageH, int filterR){ int k; double sum=0; int row=blockDim.y*blockIdx.y+threadIdx.y; int col=blockDim.x*blockIdx.x+threadIdx.x; for (k = -filterR; k <= filterR; k++) { int d = col+ k; if (d >= 0 && d < imageW) { sum += d_Src[row * imageW + d] * d_Filter[filterR - k]; } d_Dst[row * imageW + col] = sum; } }
.file "tmpxft_000963a0_00000000-6_ConvolutionRowGPU.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 _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii .type _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii, @function _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_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 _Z17ConvolutionRowGPUPdS_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 _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii, .-_Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii .globl _Z17ConvolutionRowGPUPdS_S_iii .type _Z17ConvolutionRowGPUPdS_S_iii, @function _Z17ConvolutionRowGPUPdS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z17ConvolutionRowGPUPdS_S_iii, .-_Z17ConvolutionRowGPUPdS_S_iii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z17ConvolutionRowGPUPdS_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 _Z17ConvolutionRowGPUPdS_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 ConvolutionRowGPU(double *d_Dst, double *d_Src, double *d_Filter, int imageW, int imageH, int filterR){ int k; double sum=0; int row=blockDim.y*blockIdx.y+threadIdx.y; int col=blockDim.x*blockIdx.x+threadIdx.x; for (k = -filterR; k <= filterR; k++) { int d = col+ k; if (d >= 0 && d < imageW) { sum += d_Src[row * imageW + d] * d_Filter[filterR - k]; } d_Dst[row * imageW + col] = sum; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void ConvolutionRowGPU(double *d_Dst, double *d_Src, double *d_Filter, int imageW, int imageH, int filterR){ int k; double sum=0; int row=blockDim.y*blockIdx.y+threadIdx.y; int col=blockDim.x*blockIdx.x+threadIdx.x; for (k = -filterR; k <= filterR; k++) { int d = col+ k; if (d >= 0 && d < imageW) { sum += d_Src[row * imageW + d] * d_Filter[filterR - k]; } d_Dst[row * imageW + col] = sum; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void ConvolutionRowGPU(double *d_Dst, double *d_Src, double *d_Filter, int imageW, int imageH, int filterR){ int k; double sum=0; int row=blockDim.y*blockIdx.y+threadIdx.y; int col=blockDim.x*blockIdx.x+threadIdx.x; for (k = -filterR; k <= filterR; k++) { int d = col+ k; if (d >= 0 && d < imageW) { sum += d_Src[row * imageW + d] * d_Filter[filterR - k]; } d_Dst[row * imageW + col] = sum; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z17ConvolutionRowGPUPdS_S_iii .globl _Z17ConvolutionRowGPUPdS_S_iii .p2align 8 .type _Z17ConvolutionRowGPUPdS_S_iii,@function _Z17ConvolutionRowGPUPdS_S_iii: s_load_b32 s9, s[0:1], 0x20 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s9, 0 s_cbranch_scc1 .LBB0_5 s_clause 0x1 s_load_b32 s2, s[0:1], 0x34 s_load_b32 s8, s[0:1], 0x18 v_bfe_u32 v1, v0, 10, 10 v_and_b32_e32 v0, 0x3ff, v0 s_load_b128 s[4:7], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_lshr_b32 s3, s2, 16 s_and_b32 s2, s2, 0xffff v_mad_u64_u32 v[2:3], null, s15, s3, v[1:2] v_mad_u64_u32 v[3:4], null, s14, s2, v[0:1] s_load_b64 s[2:3], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mul_lo_u32 v4, v2, s8 v_subrev_nc_u32_e32 v5, s9, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v0, v4, v3 v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_lshlrev_b64 v[6:7], 3, v[0:1] v_mov_b32_e32 v0, 0 v_mov_b32_e32 v1, 0 v_add_co_u32 v2, vcc_lo, s4, v6 v_mov_b32_e32 v6, 0 v_add_co_ci_u32_e32 v3, vcc_lo, s5, v7, vcc_lo s_lshl_b32 s4, s9, 1 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_or_b32 exec_lo, exec_lo, s0 v_add_nc_u32_e32 v5, 1, v5 s_add_i32 s4, s4, -1 global_store_b64 v[2:3], v[0:1], off s_cmp_lg_u32 s4, -1 s_cbranch_scc0 .LBB0_5 .LBB0_3: v_cmp_lt_i32_e32 vcc_lo, -1, v5 v_cmp_gt_i32_e64 s0, s8, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s1, vcc_lo, s0 s_and_saveexec_b32 s0, s1 s_cbranch_execz .LBB0_2 v_add_nc_u32_e32 v7, v4, v5 s_ashr_i32 s5, s4, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) s_lshl_b64 s[10:11], s[4:5], 3 s_waitcnt lgkmcnt(0) s_add_u32 s10, s2, s10 v_ashrrev_i32_e32 v8, 31, v7 s_addc_u32 s11, s3, s11 v_lshlrev_b64 v[7:8], 3, v[7:8] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v7, vcc_lo, s6, v7 v_add_co_ci_u32_e32 v8, vcc_lo, s7, v8, vcc_lo global_load_b64 v[7:8], v[7:8], off global_load_b64 v[9:10], v6, s[10:11] s_waitcnt vmcnt(0) v_fma_f64 v[0:1], v[7:8], v[9:10], v[0:1] s_branch .LBB0_2 .LBB0_5: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z17ConvolutionRowGPUPdS_S_iii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .amdhsa_user_sgpr_count 14 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 11 .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 _Z17ConvolutionRowGPUPdS_S_iii, .Lfunc_end0-_Z17ConvolutionRowGPUPdS_S_iii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 296 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z17ConvolutionRowGPUPdS_S_iii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z17ConvolutionRowGPUPdS_S_iii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 11 .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 ConvolutionRowGPU(double *d_Dst, double *d_Src, double *d_Filter, int imageW, int imageH, int filterR){ int k; double sum=0; int row=blockDim.y*blockIdx.y+threadIdx.y; int col=blockDim.x*blockIdx.x+threadIdx.x; for (k = -filterR; k <= filterR; k++) { int d = col+ k; if (d >= 0 && d < imageW) { sum += d_Src[row * imageW + d] * d_Filter[filterR - k]; } d_Dst[row * imageW + col] = sum; } }
.text .file "ConvolutionRowGPU.hip" .globl _Z32__device_stub__ConvolutionRowGPUPdS_S_iii # -- Begin function _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .p2align 4, 0x90 .type _Z32__device_stub__ConvolutionRowGPUPdS_S_iii,@function _Z32__device_stub__ConvolutionRowGPUPdS_S_iii: # @_Z32__device_stub__ConvolutionRowGPUPdS_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 $_Z17ConvolutionRowGPUPdS_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 _Z32__device_stub__ConvolutionRowGPUPdS_S_iii, .Lfunc_end0-_Z32__device_stub__ConvolutionRowGPUPdS_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 $_Z17ConvolutionRowGPUPdS_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 _Z17ConvolutionRowGPUPdS_S_iii,@object # @_Z17ConvolutionRowGPUPdS_S_iii .section .rodata,"a",@progbits .globl _Z17ConvolutionRowGPUPdS_S_iii .p2align 3, 0x0 _Z17ConvolutionRowGPUPdS_S_iii: .quad _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .size _Z17ConvolutionRowGPUPdS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z17ConvolutionRowGPUPdS_S_iii" .size .L__unnamed_1, 31 .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 _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17ConvolutionRowGPUPdS_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 host assembly to AMD host assembly.
.file "tmpxft_000963a0_00000000-6_ConvolutionRowGPU.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 _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii .type _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii, @function _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_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 _Z17ConvolutionRowGPUPdS_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 _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii, .-_Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii .globl _Z17ConvolutionRowGPUPdS_S_iii .type _Z17ConvolutionRowGPUPdS_S_iii, @function _Z17ConvolutionRowGPUPdS_S_iii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z44__device_stub__Z17ConvolutionRowGPUPdS_S_iiiPdS_S_iii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z17ConvolutionRowGPUPdS_S_iii, .-_Z17ConvolutionRowGPUPdS_S_iii .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z17ConvolutionRowGPUPdS_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 _Z17ConvolutionRowGPUPdS_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 "ConvolutionRowGPU.hip" .globl _Z32__device_stub__ConvolutionRowGPUPdS_S_iii # -- Begin function _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .p2align 4, 0x90 .type _Z32__device_stub__ConvolutionRowGPUPdS_S_iii,@function _Z32__device_stub__ConvolutionRowGPUPdS_S_iii: # @_Z32__device_stub__ConvolutionRowGPUPdS_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 $_Z17ConvolutionRowGPUPdS_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 _Z32__device_stub__ConvolutionRowGPUPdS_S_iii, .Lfunc_end0-_Z32__device_stub__ConvolutionRowGPUPdS_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 $_Z17ConvolutionRowGPUPdS_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 _Z17ConvolutionRowGPUPdS_S_iii,@object # @_Z17ConvolutionRowGPUPdS_S_iii .section .rodata,"a",@progbits .globl _Z17ConvolutionRowGPUPdS_S_iii .p2align 3, 0x0 _Z17ConvolutionRowGPUPdS_S_iii: .quad _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .size _Z17ConvolutionRowGPUPdS_S_iii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z17ConvolutionRowGPUPdS_S_iii" .size .L__unnamed_1, 31 .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 _Z32__device_stub__ConvolutionRowGPUPdS_S_iii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17ConvolutionRowGPUPdS_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.
__global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
code for sm_80 Function : _Z3addPiS_S_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*/ MOV R2, c[0x0][0x178] ; /* 0x00005e0000027a02 */ /* 0x000fe20000000f00 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ MOV R3, c[0x0][0x17c] ; /* 0x00005f0000037a02 */ /* 0x000fca0000000f00 */ /*0040*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0050*/ S2R R7, SR_CTAID.X ; /* 0x0000000000077919 */ /* 0x000ea40000002500 */ /*0060*/ ISETP.GE.AND P0, PT, R7, R2, PT ; /* 0x000000020700720c */ /* 0x004fda0003f06270 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fd400000001ff */ /*0090*/ IMAD.WIDE R4, R7, R6, c[0x0][0x168] ; /* 0x00005a0007047625 */ /* 0x000fc800078e0206 */ /*00a0*/ IMAD.WIDE R2, R7.reuse, R6.reuse, c[0x0][0x160] ; /* 0x0000580007027625 */ /* 0x0c0fe400078e0206 */ /*00b0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*00d0*/ IMAD.WIDE R6, R7, R6, c[0x0][0x170] ; /* 0x00005c0007067625 */ /* 0x000fe200078e0206 */ /*00e0*/ IADD3 R9, R4, R3, RZ ; /* 0x0000000304097210 */ /* 0x004fca0007ffe0ff */ /*00f0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*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.
__global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
.file "tmpxft_000ee78c_00000000-6_kernel_add_vector.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 _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ .type _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_, @function _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z3addPiS_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_, .-_Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ .globl _Z3addPiS_S_S_ .type _Z3addPiS_S_S_, @function _Z3addPiS_S_S_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z3addPiS_S_S_, .-_Z3addPiS_S_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3addPiS_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3addPiS_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
__global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
#include <hip/hip_runtime.h> __global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> __global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addPiS_S_S_ .globl _Z3addPiS_S_S_ .p2align 8 .type _Z3addPiS_S_S_,@function _Z3addPiS_S_S_: s_load_b64 s[4:5], s[0:1], 0x18 s_waitcnt lgkmcnt(0) s_load_b32 s3, s[4:5], 0x0 s_waitcnt lgkmcnt(0) s_cmp_ge_i32 s15, s3 s_cbranch_scc1 .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x0 s_mov_b32 s2, s15 s_ashr_i32 s3, s15, 31 s_load_b64 s[0:1], s[0:1], 0x10 s_lshl_b64 s[2:3], s[2:3], 2 s_waitcnt lgkmcnt(0) s_add_u32 s4, s4, s2 s_addc_u32 s5, s5, s3 s_add_u32 s6, s6, s2 s_addc_u32 s7, s7, s3 s_load_b32 s4, s[4:5], 0x0 s_load_b32 s5, s[6:7], 0x0 s_waitcnt lgkmcnt(0) s_add_i32 s4, s5, s4 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s4 s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 global_store_b32 v0, v1, s[0:1] .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3addPiS_S_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 2 .amdhsa_next_free_sgpr 16 .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 _Z3addPiS_S_S_, .Lfunc_end0-_Z3addPiS_S_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addPiS_S_S_ .private_segment_fixed_size: 0 .sgpr_count: 16 .sgpr_spill_count: 0 .symbol: _Z3addPiS_S_S_.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> __global__ void add(int *a, int *b, int *c, int *n) { int index = blockIdx.x; if( index < *n) { c[index] = a[index] + b[index] ; } }
.text .file "kernel_add_vector.hip" .globl _Z18__device_stub__addPiS_S_S_ # -- Begin function _Z18__device_stub__addPiS_S_S_ .p2align 4, 0x90 .type _Z18__device_stub__addPiS_S_S_,@function _Z18__device_stub__addPiS_S_S_: # @_Z18__device_stub__addPiS_S_S_ .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 $_Z3addPiS_S_S_, %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 _Z18__device_stub__addPiS_S_S_, .Lfunc_end0-_Z18__device_stub__addPiS_S_S_ .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z3addPiS_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z3addPiS_S_S_,@object # @_Z3addPiS_S_S_ .section .rodata,"a",@progbits .globl _Z3addPiS_S_S_ .p2align 3, 0x0 _Z3addPiS_S_S_: .quad _Z18__device_stub__addPiS_S_S_ .size _Z3addPiS_S_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3addPiS_S_S_" .size .L__unnamed_1, 15 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z18__device_stub__addPiS_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addPiS_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z3addPiS_S_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*/ MOV R2, c[0x0][0x178] ; /* 0x00005e0000027a02 */ /* 0x000fe20000000f00 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0030*/ MOV R3, c[0x0][0x17c] ; /* 0x00005f0000037a02 */ /* 0x000fca0000000f00 */ /*0040*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0050*/ S2R R7, SR_CTAID.X ; /* 0x0000000000077919 */ /* 0x000ea40000002500 */ /*0060*/ ISETP.GE.AND P0, PT, R7, R2, PT ; /* 0x000000020700720c */ /* 0x004fda0003f06270 */ /*0070*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0080*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fd400000001ff */ /*0090*/ IMAD.WIDE R4, R7, R6, c[0x0][0x168] ; /* 0x00005a0007047625 */ /* 0x000fc800078e0206 */ /*00a0*/ IMAD.WIDE R2, R7.reuse, R6.reuse, c[0x0][0x160] ; /* 0x0000580007027625 */ /* 0x0c0fe400078e0206 */ /*00b0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*00c0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*00d0*/ IMAD.WIDE R6, R7, R6, c[0x0][0x170] ; /* 0x00005c0007067625 */ /* 0x000fe200078e0206 */ /*00e0*/ IADD3 R9, R4, R3, RZ ; /* 0x0000000304097210 */ /* 0x004fca0007ffe0ff */ /*00f0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*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 _Z3addPiS_S_S_ .globl _Z3addPiS_S_S_ .p2align 8 .type _Z3addPiS_S_S_,@function _Z3addPiS_S_S_: s_load_b64 s[4:5], s[0:1], 0x18 s_waitcnt lgkmcnt(0) s_load_b32 s3, s[4:5], 0x0 s_waitcnt lgkmcnt(0) s_cmp_ge_i32 s15, s3 s_cbranch_scc1 .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x0 s_mov_b32 s2, s15 s_ashr_i32 s3, s15, 31 s_load_b64 s[0:1], s[0:1], 0x10 s_lshl_b64 s[2:3], s[2:3], 2 s_waitcnt lgkmcnt(0) s_add_u32 s4, s4, s2 s_addc_u32 s5, s5, s3 s_add_u32 s6, s6, s2 s_addc_u32 s7, s7, s3 s_load_b32 s4, s[4:5], 0x0 s_load_b32 s5, s[6:7], 0x0 s_waitcnt lgkmcnt(0) s_add_i32 s4, s5, s4 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s4 s_add_u32 s0, s0, s2 s_addc_u32 s1, s1, s3 global_store_b32 v0, v1, s[0:1] .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3addPiS_S_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 2 .amdhsa_next_free_sgpr 16 .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 _Z3addPiS_S_S_, .Lfunc_end0-_Z3addPiS_S_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 24 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addPiS_S_S_ .private_segment_fixed_size: 0 .sgpr_count: 16 .sgpr_spill_count: 0 .symbol: _Z3addPiS_S_S_.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_000ee78c_00000000-6_kernel_add_vector.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 _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ .type _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_, @function _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %rcx, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z3addPiS_S_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_, .-_Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ .globl _Z3addPiS_S_S_ .type _Z3addPiS_S_S_, @function _Z3addPiS_S_S_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z3addPiS_S_S_PiS_S_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z3addPiS_S_S_, .-_Z3addPiS_S_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3addPiS_S_S_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3addPiS_S_S_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "kernel_add_vector.hip" .globl _Z18__device_stub__addPiS_S_S_ # -- Begin function _Z18__device_stub__addPiS_S_S_ .p2align 4, 0x90 .type _Z18__device_stub__addPiS_S_S_,@function _Z18__device_stub__addPiS_S_S_: # @_Z18__device_stub__addPiS_S_S_ .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 $_Z3addPiS_S_S_, %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 _Z18__device_stub__addPiS_S_S_, .Lfunc_end0-_Z18__device_stub__addPiS_S_S_ .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z3addPiS_S_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z3addPiS_S_S_,@object # @_Z3addPiS_S_S_ .section .rodata,"a",@progbits .globl _Z3addPiS_S_S_ .p2align 3, 0x0 _Z3addPiS_S_S_: .quad _Z18__device_stub__addPiS_S_S_ .size _Z3addPiS_S_S_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3addPiS_S_S_" .size .L__unnamed_1, 15 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z18__device_stub__addPiS_S_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addPiS_S_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
code for sm_80 Function : testKernel .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 */ /* 0x000e280000002100 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0030*/ IMAD R4, R3, c[0x0][0x0], R4 ; /* 0x0000000003047a24 */ /* 0x001fca00078e0204 */ /*0040*/ ISETP.GE.AND P0, PT, R4, c[0x0][0x170], PT ; /* 0x00005c0004007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R5, -RZ, RZ, 0, 4.76837158203125e-07 ; /* 0x00000008ff057435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R2, R4, R5, c[0x0][0x160] ; /* 0x0000580004027625 */ /* 0x000fcc00078e0205 */ /*0090*/ LDG.E.64 R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1b00 */ /*00a0*/ IMAD.WIDE R4, R4, R5, c[0x0][0x168] ; /* 0x00005a0004047625 */ /* 0x000fca00078e0205 */ /*00b0*/ STG.E.64 [R4.64], R2 ; /* 0x0000000204007986 */ /* 0x004fe2000c101b04 */ /*00c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00d0*/ BRA 0xd0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
.file "tmpxft_00052256_00000000-6_JCudaKernelParamsTestKernel.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__Z10testKernelPPfS0_iPPfS0_i .type _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i, @function _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq testKernel(%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 _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i, .-_Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i .globl testKernel .type testKernel, @function testKernel: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size testKernel, .-testKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "testKernel" .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 testKernel(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
#include <hip/hip_runtime.h> extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected testKernel .globl testKernel .p2align 8 .type testKernel,@function testKernel: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 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[0:3], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s1, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo global_load_b64 v[2:3], v[2:3], off s_waitcnt vmcnt(0) global_store_b64 v[0:1], v[2:3], off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel testKernel .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 280 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 4 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size testKernel, .Lfunc_end0-testKernel .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: 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: testKernel .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: testKernel.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> extern "C" __global__ void testKernel( float **inputPointers, float **outputPointers, int numPointers) { int tid = threadIdx.x + blockDim.x * blockIdx.x; if (tid < numPointers) { outputPointers[tid] = inputPointers[tid]; } }
.text .file "JCudaKernelParamsTestKernel.hip" .globl __device_stub__testKernel # -- Begin function __device_stub__testKernel .p2align 4, 0x90 .type __device_stub__testKernel,@function __device_stub__testKernel: # @__device_stub__testKernel .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $testKernel, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size __device_stub__testKernel, .Lfunc_end0-__device_stub__testKernel .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 $testKernel, %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 testKernel,@object # @testKernel .section .rodata,"a",@progbits .globl testKernel .p2align 3, 0x0 testKernel: .quad __device_stub__testKernel .size testKernel, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "testKernel" .size .L__unnamed_1, 11 .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__testKernel .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym testKernel .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 : testKernel .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 */ /* 0x000e280000002100 */ /*0020*/ S2R R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0030*/ IMAD R4, R3, c[0x0][0x0], R4 ; /* 0x0000000003047a24 */ /* 0x001fca00078e0204 */ /*0040*/ ISETP.GE.AND P0, PT, R4, c[0x0][0x170], PT ; /* 0x00005c0004007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R5, -RZ, RZ, 0, 4.76837158203125e-07 ; /* 0x00000008ff057435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R2, R4, R5, c[0x0][0x160] ; /* 0x0000580004027625 */ /* 0x000fcc00078e0205 */ /*0090*/ LDG.E.64 R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea2000c1e1b00 */ /*00a0*/ IMAD.WIDE R4, R4, R5, c[0x0][0x168] ; /* 0x00005a0004047625 */ /* 0x000fca00078e0205 */ /*00b0*/ STG.E.64 [R4.64], R2 ; /* 0x0000000204007986 */ /* 0x004fe2000c101b04 */ /*00c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00d0*/ BRA 0xd0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected testKernel .globl testKernel .p2align 8 .type testKernel,@function testKernel: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 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[0:3], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 3, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s1, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo global_load_b64 v[2:3], v[2:3], off s_waitcnt vmcnt(0) global_store_b64 v[0:1], v[2:3], off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel testKernel .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 280 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 4 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size testKernel, .Lfunc_end0-testKernel .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: 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: testKernel .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: testKernel.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_00052256_00000000-6_JCudaKernelParamsTestKernel.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__Z10testKernelPPfS0_iPPfS0_i .type _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i, @function _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq testKernel(%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 _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i, .-_Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i .globl testKernel .type testKernel, @function testKernel: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z10testKernelPPfS0_iPPfS0_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size testKernel, .-testKernel .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "testKernel" .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 testKernel(%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 "JCudaKernelParamsTestKernel.hip" .globl __device_stub__testKernel # -- Begin function __device_stub__testKernel .p2align 4, 0x90 .type __device_stub__testKernel,@function __device_stub__testKernel: # @__device_stub__testKernel .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $testKernel, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size __device_stub__testKernel, .Lfunc_end0-__device_stub__testKernel .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 $testKernel, %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 testKernel,@object # @testKernel .section .rodata,"a",@progbits .globl testKernel .p2align 3, 0x0 testKernel: .quad __device_stub__testKernel .size testKernel, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "testKernel" .size .L__unnamed_1, 11 .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__testKernel .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym testKernel .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> __global__ void holaCUDA(float e){ printf("Hola, soy el hilo %d del bloque %d con valor pi -> %f\n", threadIdx.x, blockIdx.x,e); } int main (int argc, char **argv) { holaCUDA<<<3,4>>>(3.1416); cudaDeviceReset(); //Reinicializa el device return 0; }
code for sm_80 Function : _Z8holaCUDAf .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 */ /* 0x000fc600078e00ff */ /*0010*/ S2R R11, SR_CTAID.X ; /* 0x00000000000b7919 */ /* 0x000e220000002500 */ /*0020*/ F2F.F64.F32 R2, c[0x0][0x160] ; /* 0x0000580000027b10 */ /* 0x000e620000201800 */ /*0030*/ IADD3 R1, R1, -0x10, RZ ; /* 0xfffffff001017810 */ /* 0x000fe20007ffe0ff */ /*0040*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff047624 */ /* 0x000fe200078e00ff */ /*0050*/ S2R R10, SR_TID.X ; /* 0x00000000000a7919 */ /* 0x000e220000002100 */ /*0060*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe20000000f00 */ /*0070*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff057624 */ /* 0x000fe200078e00ff */ /*0080*/ IADD3 R6, P0, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fe40007f1e0ff */ /*0090*/ LDC.64 R8, c[0x4][R0] ; /* 0x0100000000087b82 */ /* 0x0004e60000000a00 */ /*00a0*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P0 ; /* 0x00000900ff077624 */ /* 0x000fe200000e06ff */ /*00b0*/ STL.64 [R1+0x8], R2 ; /* 0x0000080201007387 */ /* 0x0025e80000100a00 */ /*00c0*/ STL.64 [R1], R10 ; /* 0x0000000a01007387 */ /* 0x0015e60000100a00 */ /*00d0*/ LEPC R2 ; /* 0x000000000002734e */ /* 0x004fe40000000000 */ /*00e0*/ MOV R11, 0x150 ; /* 0x00000150000b7802 */ /* 0x000fe40000000f00 */ /*00f0*/ MOV R20, 0xd0 ; /* 0x000000d000147802 */ /* 0x000fc40000000f00 */ /*0100*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fe40000000f00 */ /*0110*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe40000000f00 */ /*0120*/ IADD3 R20, P0, P1, -R20, R11, R2 ; /* 0x0000000b14147210 */ /* 0x000fc8000791e102 */ /*0130*/ IADD3.X R21, ~R0, R21, R3, P0, P1 ; /* 0x0000001500157210 */ /* 0x000fc800007e2503 */ /*0140*/ CALL.ABS.NOINC R8 ; /* 0x0000000008007343 */ /* 0x008fea0003c00000 */ /*0150*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0160*/ BRA 0x160; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> __global__ void holaCUDA(float e){ printf("Hola, soy el hilo %d del bloque %d con valor pi -> %f\n", threadIdx.x, blockIdx.x,e); } int main (int argc, char **argv) { holaCUDA<<<3,4>>>(3.1416); cudaDeviceReset(); //Reinicializa el device return 0; }
.file "tmpxft_000129da_00000000-6_hello_world.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z8holaCUDAff .type _Z26__device_stub__Z8holaCUDAff, @function _Z26__device_stub__Z8holaCUDAff: .LFB2082: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z8holaCUDAf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z26__device_stub__Z8holaCUDAff, .-_Z26__device_stub__Z8holaCUDAff .globl _Z8holaCUDAf .type _Z8holaCUDAf, @function _Z8holaCUDAf: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z8holaCUDAff addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z8holaCUDAf, .-_Z8holaCUDAf .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $4, 20(%rsp) movl $1, 24(%rsp) movl $3, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceReset@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state movss .LC0(%rip), %xmm0 call _Z26__device_stub__Z8holaCUDAff jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "_Z8holaCUDAf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z8holaCUDAf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1078530041 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> __global__ void holaCUDA(float e){ printf("Hola, soy el hilo %d del bloque %d con valor pi -> %f\n", threadIdx.x, blockIdx.x,e); } int main (int argc, char **argv) { holaCUDA<<<3,4>>>(3.1416); cudaDeviceReset(); //Reinicializa el device return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void holaCUDA(float e){ printf("Hola, soy el hilo %d del bloque %d con valor pi -> %f\n", threadIdx.x, blockIdx.x,e); } int main (int argc, char **argv) { holaCUDA<<<3,4>>>(3.1416); hipDeviceReset(); //Reinicializa el device return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void holaCUDA(float e){ printf("Hola, soy el hilo %d del bloque %d con valor pi -> %f\n", threadIdx.x, blockIdx.x,e); } int main (int argc, char **argv) { holaCUDA<<<3,4>>>(3.1416); hipDeviceReset(); //Reinicializa el device return 0; }
.text .file "hello_world.hip" .globl _Z23__device_stub__holaCUDAf # -- Begin function _Z23__device_stub__holaCUDAf .p2align 4, 0x90 .type _Z23__device_stub__holaCUDAf,@function _Z23__device_stub__holaCUDAf: # @_Z23__device_stub__holaCUDAf .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movss %xmm0, 12(%rsp) leaq 12(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z8holaCUDAf, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size _Z23__device_stub__holaCUDAf, .Lfunc_end0-_Z23__device_stub__holaCUDAf .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 movabsq $4294967299, %rdi # imm = 0x100000003 leaq 1(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movl $1078530041, 12(%rsp) # imm = 0x40490FF9 leaq 12(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z8holaCUDAf, %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 hipDeviceReset 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: 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 $_Z8holaCUDAf, %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 _Z8holaCUDAf,@object # @_Z8holaCUDAf .section .rodata,"a",@progbits .globl _Z8holaCUDAf .p2align 3, 0x0 _Z8holaCUDAf: .quad _Z23__device_stub__holaCUDAf .size _Z8holaCUDAf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8holaCUDAf" .size .L__unnamed_1, 13 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__holaCUDAf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8holaCUDAf .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_000129da_00000000-6_hello_world.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z26__device_stub__Z8holaCUDAff .type _Z26__device_stub__Z8holaCUDAff, @function _Z26__device_stub__Z8holaCUDAff: .LFB2082: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z8holaCUDAf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z26__device_stub__Z8holaCUDAff, .-_Z26__device_stub__Z8holaCUDAff .globl _Z8holaCUDAf .type _Z8holaCUDAf, @function _Z8holaCUDAf: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z26__device_stub__Z8holaCUDAff addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z8holaCUDAf, .-_Z8holaCUDAf .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $4, 20(%rsp) movl $1, 24(%rsp) movl $3, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceReset@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state movss .LC0(%rip), %xmm0 call _Z26__device_stub__Z8holaCUDAff jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "_Z8holaCUDAf" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z8holaCUDAf(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1078530041 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello_world.hip" .globl _Z23__device_stub__holaCUDAf # -- Begin function _Z23__device_stub__holaCUDAf .p2align 4, 0x90 .type _Z23__device_stub__holaCUDAf,@function _Z23__device_stub__holaCUDAf: # @_Z23__device_stub__holaCUDAf .cfi_startproc # %bb.0: subq $72, %rsp .cfi_def_cfa_offset 80 movss %xmm0, 12(%rsp) leaq 12(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z8holaCUDAf, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $88, %rsp .cfi_adjust_cfa_offset -88 retq .Lfunc_end0: .size _Z23__device_stub__holaCUDAf, .Lfunc_end0-_Z23__device_stub__holaCUDAf .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 movabsq $4294967299, %rdi # imm = 0x100000003 leaq 1(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: movl $1078530041, 12(%rsp) # imm = 0x40490FF9 leaq 12(%rsp), %rax movq %rax, 16(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z8holaCUDAf, %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 hipDeviceReset 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: 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 $_Z8holaCUDAf, %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 _Z8holaCUDAf,@object # @_Z8holaCUDAf .section .rodata,"a",@progbits .globl _Z8holaCUDAf .p2align 3, 0x0 _Z8holaCUDAf: .quad _Z23__device_stub__holaCUDAf .size _Z8holaCUDAf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8holaCUDAf" .size .L__unnamed_1, 13 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__holaCUDAf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8holaCUDAf .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 <assert.h> #include <stdio.h> #include <cuda.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); cudaMalloc((void **) &d_a, memSize); cudaMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } cudaMemcpy( d_a, h_a, memSize, cudaMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced cudaThreadSynchronize(); checkCudaError("kernel invocation"); cudaMemcpy(h_a, d_b, memSize, cudaMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } cudaFree(d_a); cudaFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
code for sm_80 Function : _Z17reverseArrayBlockPiS_ .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 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R0, c[0x0][0x0], R5 ; /* 0x0000000000027a24 */ /* 0x001fca00078e0205 */ /*0060*/ IMAD.WIDE R2, R2, R6, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fcc00078e0206 */ /*0070*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0080*/ LOP3.LUT R0, RZ, R0, RZ, 0x33, !PT ; /* 0x00000000ff007212 */ /* 0x000fe400078e33ff */ /*0090*/ LOP3.LUT R4, RZ, R5, RZ, 0x33, !PT ; /* 0x00000005ff047212 */ /* 0x000fe400078e33ff */ /*00a0*/ IADD3 R0, R0, c[0x0][0xc], RZ ; /* 0x0000030000007a10 */ /* 0x000fe40007ffe0ff */ /*00b0*/ IADD3 R5, R4, c[0x0][0x0], RZ ; /* 0x0000000004057a10 */ /* 0x000fca0007ffe0ff */ /*00c0*/ IMAD R5, R0, c[0x0][0x0], R5 ; /* 0x0000000000057a24 */ /* 0x000fc800078e0205 */ /*00d0*/ IMAD.WIDE R4, R5, R6, c[0x0][0x160] ; /* 0x0000580005047625 */ /* 0x000fca00078e0206 */ /*00e0*/ STG.E [R4.64], R3 ; /* 0x0000000304007986 */ /* 0x004fe2000c101904 */ /*00f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0100*/ BRA 0x100; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <assert.h> #include <stdio.h> #include <cuda.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); cudaMalloc((void **) &d_a, memSize); cudaMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } cudaMemcpy( d_a, h_a, memSize, cudaMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced cudaThreadSynchronize(); checkCudaError("kernel invocation"); cudaMemcpy(h_a, d_b, memSize, cudaMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } cudaFree(d_a); cudaFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
.file "tmpxft_00161aeb_00000000-6_reverseArray.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 _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ .type _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_, @function _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_: .LFB2083: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 104(%rsp), %rax subq %fs:40, %rax jne .L12 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .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 _Z17reverseArrayBlockPiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_, .-_Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ .globl _Z17reverseArrayBlockPiS_ .type _Z17reverseArrayBlockPiS_, @function _Z17reverseArrayBlockPiS_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z17reverseArrayBlockPiS_, .-_Z17reverseArrayBlockPiS_ .section .rodata.str1.1 .LC1: .string "kernel invocation" .LC2: .string "memcpy" .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 $1048576, %edi call malloc@PLT movq %rax, %rbx leaq 8(%rsp), %rdi movl $1048576, %esi call cudaMalloc@PLT movq %rsp, %rdi movl $1048576, %esi call cudaMalloc@PLT movl $0, %eax .L16: movl %eax, (%rbx,%rax,4) addq $1, %rax cmpq $262144, %rax jne .L16 movl $1, %ecx movl $1048576, %edx movq %rbx, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 16(%rsp) movl $1, 20(%rsp) movl $256, 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 .L23 .L17: call cudaThreadSynchronize@PLT leaq .LC1(%rip), %rdi call _Z14checkCudaErrorPKc movl $2, %ecx movl $1048576, %edx movq (%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rdi call _Z14checkCudaErrorPKc movl $262144, %eax .L18: subl $1, %eax jne .L18 movq 8(%rsp), %rdi call cudaFree@PLT movq (%rsp), %rdi call cudaFree@PLT movq %rbx, %rdi call free@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L24 movl $0, %eax addq $48, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ jmp .L17 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z17reverseArrayBlockPiS_" .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z17reverseArrayBlockPiS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <assert.h> #include <stdio.h> #include <cuda.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); cudaMalloc((void **) &d_a, memSize); cudaMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } cudaMemcpy( d_a, h_a, memSize, cudaMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced cudaThreadSynchronize(); checkCudaError("kernel invocation"); cudaMemcpy(h_a, d_b, memSize, cudaMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } cudaFree(d_a); cudaFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
#include <assert.h> #include <stdio.h> #include <hip/hip_runtime.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); hipMalloc((void **) &d_a, memSize); hipMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } hipMemcpy( d_a, h_a, memSize, hipMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced hipDeviceSynchronize(); checkCudaError("kernel invocation"); hipMemcpy(h_a, d_b, memSize, hipMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } hipFree(d_a); hipFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <assert.h> #include <stdio.h> #include <hip/hip_runtime.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); hipMalloc((void **) &d_a, memSize); hipMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } hipMemcpy( d_a, h_a, memSize, hipMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced hipDeviceSynchronize(); checkCudaError("kernel invocation"); hipMemcpy(h_a, d_b, memSize, hipMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } hipFree(d_a); hipFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z17reverseArrayBlockPiS_ .globl _Z17reverseArrayBlockPiS_ .p2align 8 .type _Z17reverseArrayBlockPiS_,@function _Z17reverseArrayBlockPiS_: s_clause 0x2 s_load_b32 s2, s[0:1], 0x1c s_load_b128 s[4:7], s[0:1], 0x0 s_load_b32 s0, s[0:1], 0x10 s_not_b32 s1, s15 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_add_i32 s0, s0, s1 v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_not_b32_e32 v0, v0 s_mul_i32 s0, s0, s2 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_add3_u32 v0, s0, s2, v0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 v_lshlrev_b64 v[1:2], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v1, vcc_lo, s6, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v2, vcc_lo global_load_b32 v2, v[1:2], off v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] 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_waitcnt vmcnt(0) 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 _Z17reverseArrayBlockPiS_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z17reverseArrayBlockPiS_, .Lfunc_end0-_Z17reverseArrayBlockPiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z17reverseArrayBlockPiS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z17reverseArrayBlockPiS_.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 <assert.h> #include <stdio.h> #include <hip/hip_runtime.h> // declaration of function to check for errors void checkCudaError(const char* msg); // kernel __global__ void reverseArrayBlock(int *d_out, int *d_in) { int inOffset = blockDim.x * blockIdx.x; int outOffset = blockDim.x * (gridDim.x - 1 - blockIdx.x); int in = inOffset + threadIdx.x; int out = outOffset + (blockDim.x - 1 - threadIdx.x); d_out[out] = d_in[in]; } // main thread int main(int argc, char** argv) { int *h_a; int dimA = 256 * 1024; int *d_b, *d_a; int numThreadsPerBlock = 256; // compute number of blocks needed int numBlocks = dimA / numThreadsPerBlock; // allocate memory size_t memSize = numBlocks * numThreadsPerBlock*sizeof(int); h_a = (int *)malloc(memSize); hipMalloc((void **) &d_a, memSize); hipMalloc((void **) &d_b, memSize); // init array on host for (int i=0; i<dimA; i++) { h_a[i] = i; } hipMemcpy( d_a, h_a, memSize, hipMemcpyHostToDevice); dim3 dimGrid(numBlocks); dim3 dimBlock(numThreadsPerBlock); reverseArrayBlock <<< dimGrid, dimBlock >>> (d_b, d_a); // block until synced hipDeviceSynchronize(); checkCudaError("kernel invocation"); hipMemcpy(h_a, d_b, memSize, hipMemcpyDeviceToHost); checkCudaError("memcpy"); for (int i=0; i<dimA; i++) { assert( h_a[i] == dimA - 1 - i); } hipFree(d_a); hipFree(d_b); free(h_a); 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(EXIT_FAILURE); } }
.text .file "reverseArray.hip" .globl _Z32__device_stub__reverseArrayBlockPiS_ # -- Begin function _Z32__device_stub__reverseArrayBlockPiS_ .p2align 4, 0x90 .type _Z32__device_stub__reverseArrayBlockPiS_,@function _Z32__device_stub__reverseArrayBlockPiS_: # @_Z32__device_stub__reverseArrayBlockPiS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z17reverseArrayBlockPiS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z32__device_stub__reverseArrayBlockPiS_, .Lfunc_end0-_Z32__device_stub__reverseArrayBlockPiS_ .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 $96, %rsp .cfi_def_cfa_offset 112 .cfi_offset %rbx, -16 movl $1048576, %edi # imm = 0x100000 callq malloc movq %rax, %rbx movq %rsp, %rdi movl $1048576, %esi # imm = 0x100000 callq hipMalloc leaq 8(%rsp), %rdi movl $1048576, %esi # imm = 0x100000 callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $262144, %rax # imm = 0x40000 jne .LBB1_1 # %bb.2: movq (%rsp), %rdi movl $1048576, %edx # imm = 0x100000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967552, %rdx # imm = 0x100000100 leaq 768(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 8(%rsp), %rax movq (%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z17reverseArrayBlockPiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: callq hipDeviceSynchronize callq hipGetLastError testl %eax, %eax jne .LBB1_5 # %bb.7: # %_Z14checkCudaErrorPKc.exit movq 8(%rsp), %rsi movl $1048576, %edx # imm = 0x100000 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy callq hipGetLastError testl %eax, %eax jne .LBB1_8 # %bb.9: # %_Z14checkCudaErrorPKc.exit30.preheader movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free xorl %eax, %eax addq $96, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_5: .cfi_def_cfa_offset 112 movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str, %edx jmp .LBB1_6 .LBB1_8: movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str.1, %edx .LBB1_6: 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.2, %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 $_Z17reverseArrayBlockPiS_, %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 _Z17reverseArrayBlockPiS_,@object # @_Z17reverseArrayBlockPiS_ .section .rodata,"a",@progbits .globl _Z17reverseArrayBlockPiS_ .p2align 3, 0x0 _Z17reverseArrayBlockPiS_: .quad _Z32__device_stub__reverseArrayBlockPiS_ .size _Z17reverseArrayBlockPiS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "kernel invocation" .size .L.str, 18 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "memcpy" .size .L.str.1, 7 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Cuda error: %s : %s .\n" .size .L.str.2, 23 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z17reverseArrayBlockPiS_" .size .L__unnamed_1, 26 .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 _Z32__device_stub__reverseArrayBlockPiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17reverseArrayBlockPiS_ .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 : _Z17reverseArrayBlockPiS_ .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 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R6, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff067435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R0, c[0x0][0x0], R5 ; /* 0x0000000000027a24 */ /* 0x001fca00078e0205 */ /*0060*/ IMAD.WIDE R2, R2, R6, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fcc00078e0206 */ /*0070*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*0080*/ LOP3.LUT R0, RZ, R0, RZ, 0x33, !PT ; /* 0x00000000ff007212 */ /* 0x000fe400078e33ff */ /*0090*/ LOP3.LUT R4, RZ, R5, RZ, 0x33, !PT ; /* 0x00000005ff047212 */ /* 0x000fe400078e33ff */ /*00a0*/ IADD3 R0, R0, c[0x0][0xc], RZ ; /* 0x0000030000007a10 */ /* 0x000fe40007ffe0ff */ /*00b0*/ IADD3 R5, R4, c[0x0][0x0], RZ ; /* 0x0000000004057a10 */ /* 0x000fca0007ffe0ff */ /*00c0*/ IMAD R5, R0, c[0x0][0x0], R5 ; /* 0x0000000000057a24 */ /* 0x000fc800078e0205 */ /*00d0*/ IMAD.WIDE R4, R5, R6, c[0x0][0x160] ; /* 0x0000580005047625 */ /* 0x000fca00078e0206 */ /*00e0*/ STG.E [R4.64], R3 ; /* 0x0000000304007986 */ /* 0x004fe2000c101904 */ /*00f0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0100*/ BRA 0x100; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z17reverseArrayBlockPiS_ .globl _Z17reverseArrayBlockPiS_ .p2align 8 .type _Z17reverseArrayBlockPiS_,@function _Z17reverseArrayBlockPiS_: s_clause 0x2 s_load_b32 s2, s[0:1], 0x1c s_load_b128 s[4:7], s[0:1], 0x0 s_load_b32 s0, s[0:1], 0x10 s_not_b32 s1, s15 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_add_i32 s0, s0, s1 v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_not_b32_e32 v0, v0 s_mul_i32 s0, s0, s2 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_add3_u32 v0, s0, s2, v0 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 v_lshlrev_b64 v[1:2], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v1, vcc_lo, s6, v1 v_add_co_ci_u32_e32 v2, vcc_lo, s7, v2, vcc_lo global_load_b32 v2, v[1:2], off v_ashrrev_i32_e32 v1, 31, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] 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_waitcnt vmcnt(0) 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 _Z17reverseArrayBlockPiS_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 272 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z17reverseArrayBlockPiS_, .Lfunc_end0-_Z17reverseArrayBlockPiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: hidden_block_count_x - .offset: 20 .size: 4 .value_kind: hidden_block_count_y - .offset: 24 .size: 4 .value_kind: hidden_block_count_z - .offset: 28 .size: 2 .value_kind: hidden_group_size_x - .offset: 30 .size: 2 .value_kind: hidden_group_size_y - .offset: 32 .size: 2 .value_kind: hidden_group_size_z - .offset: 34 .size: 2 .value_kind: hidden_remainder_x - .offset: 36 .size: 2 .value_kind: hidden_remainder_y - .offset: 38 .size: 2 .value_kind: hidden_remainder_z - .offset: 56 .size: 8 .value_kind: hidden_global_offset_x - .offset: 64 .size: 8 .value_kind: hidden_global_offset_y - .offset: 72 .size: 8 .value_kind: hidden_global_offset_z - .offset: 80 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 272 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z17reverseArrayBlockPiS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z17reverseArrayBlockPiS_.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_00161aeb_00000000-6_reverseArray.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 _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ .type _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_, @function _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_: .LFB2083: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L11 .L7: movq 104(%rsp), %rax subq %fs:40, %rax jne .L12 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L11: .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 _Z17reverseArrayBlockPiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L7 .L12: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_, .-_Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ .globl _Z17reverseArrayBlockPiS_ .type _Z17reverseArrayBlockPiS_, @function _Z17reverseArrayBlockPiS_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z17reverseArrayBlockPiS_, .-_Z17reverseArrayBlockPiS_ .section .rodata.str1.1 .LC1: .string "kernel invocation" .LC2: .string "memcpy" .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 $1048576, %edi call malloc@PLT movq %rax, %rbx leaq 8(%rsp), %rdi movl $1048576, %esi call cudaMalloc@PLT movq %rsp, %rdi movl $1048576, %esi call cudaMalloc@PLT movl $0, %eax .L16: movl %eax, (%rbx,%rax,4) addq $1, %rax cmpq $262144, %rax jne .L16 movl $1, %ecx movl $1048576, %edx movq %rbx, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1024, 16(%rsp) movl $1, 20(%rsp) movl $256, 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 .L23 .L17: call cudaThreadSynchronize@PLT leaq .LC1(%rip), %rdi call _Z14checkCudaErrorPKc movl $2, %ecx movl $1048576, %edx movq (%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT leaq .LC2(%rip), %rdi call _Z14checkCudaErrorPKc movl $262144, %eax .L18: subl $1, %eax jne .L18 movq 8(%rsp), %rdi call cudaFree@PLT movq (%rsp), %rdi call cudaFree@PLT movq %rbx, %rdi call free@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L24 movl $0, %eax addq $48, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z39__device_stub__Z17reverseArrayBlockPiS_PiS_ jmp .L17 .L24: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z17reverseArrayBlockPiS_" .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 .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z17reverseArrayBlockPiS_(%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 "reverseArray.hip" .globl _Z32__device_stub__reverseArrayBlockPiS_ # -- Begin function _Z32__device_stub__reverseArrayBlockPiS_ .p2align 4, 0x90 .type _Z32__device_stub__reverseArrayBlockPiS_,@function _Z32__device_stub__reverseArrayBlockPiS_: # @_Z32__device_stub__reverseArrayBlockPiS_ .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movq %rdi, 56(%rsp) movq %rsi, 48(%rsp) leaq 56(%rsp), %rax movq %rax, 64(%rsp) leaq 48(%rsp), %rax movq %rax, 72(%rsp) leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z17reverseArrayBlockPiS_, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $104, %rsp .cfi_adjust_cfa_offset -104 retq .Lfunc_end0: .size _Z32__device_stub__reverseArrayBlockPiS_, .Lfunc_end0-_Z32__device_stub__reverseArrayBlockPiS_ .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 $96, %rsp .cfi_def_cfa_offset 112 .cfi_offset %rbx, -16 movl $1048576, %edi # imm = 0x100000 callq malloc movq %rax, %rbx movq %rsp, %rdi movl $1048576, %esi # imm = 0x100000 callq hipMalloc leaq 8(%rsp), %rdi movl $1048576, %esi # imm = 0x100000 callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $262144, %rax # imm = 0x40000 jne .LBB1_1 # %bb.2: movq (%rsp), %rdi movl $1048576, %edx # imm = 0x100000 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967552, %rdx # imm = 0x100000100 leaq 768(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 8(%rsp), %rax movq (%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z17reverseArrayBlockPiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: callq hipDeviceSynchronize callq hipGetLastError testl %eax, %eax jne .LBB1_5 # %bb.7: # %_Z14checkCudaErrorPKc.exit movq 8(%rsp), %rsi movl $1048576, %edx # imm = 0x100000 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy callq hipGetLastError testl %eax, %eax jne .LBB1_8 # %bb.9: # %_Z14checkCudaErrorPKc.exit30.preheader movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq %rbx, %rdi callq free xorl %eax, %eax addq $96, %rsp .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB1_5: .cfi_def_cfa_offset 112 movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str, %edx jmp .LBB1_6 .LBB1_8: movq stderr(%rip), %rbx movl %eax, %edi callq hipGetErrorString movl $.L.str.2, %esi movl $.L.str.1, %edx .LBB1_6: 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.2, %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 $_Z17reverseArrayBlockPiS_, %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 _Z17reverseArrayBlockPiS_,@object # @_Z17reverseArrayBlockPiS_ .section .rodata,"a",@progbits .globl _Z17reverseArrayBlockPiS_ .p2align 3, 0x0 _Z17reverseArrayBlockPiS_: .quad _Z32__device_stub__reverseArrayBlockPiS_ .size _Z17reverseArrayBlockPiS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "kernel invocation" .size .L.str, 18 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "memcpy" .size .L.str.1, 7 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Cuda error: %s : %s .\n" .size .L.str.2, 23 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z17reverseArrayBlockPiS_" .size .L__unnamed_1, 26 .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 _Z32__device_stub__reverseArrayBlockPiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17reverseArrayBlockPiS_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> __global__ void hello(){ printf("Hello from block: %u, thread: %u\n", blockIdx.x, threadIdx.x); } int main(){ hello<<<2,2>>>(); cudaDeviceSynchronize(); }
code for sm_80 Function : _Z5hellov .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fc800078e00ff */ /*0010*/ S2R R9, SR_TID.X ; /* 0x0000000000097919 */ /* 0x000e220000002100 */ /*0020*/ IADD3 R1, R1, -0x8, RZ ; /* 0xfffffff801017810 */ /* 0x000fe20007ffe0ff */ /*0030*/ IMAD.MOV.U32 R4, RZ, RZ, c[0x4][0x8] ; /* 0x01000200ff047624 */ /* 0x000fe200078e00ff */ /*0040*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe20000000f00 */ /*0050*/ S2R R8, SR_CTAID.X ; /* 0x0000000000087919 */ /* 0x000e220000002500 */ /*0060*/ IADD3 R6, P0, R1, c[0x0][0x20], RZ ; /* 0x0000080001067a10 */ /* 0x000fe20007f1e0ff */ /*0070*/ IMAD.MOV.U32 R5, RZ, RZ, c[0x4][0xc] ; /* 0x01000300ff057624 */ /* 0x000fe200078e00ff */ /*0080*/ LDC.64 R2, c[0x4][R0] ; /* 0x0100000000027b82 */ /* 0x0002a60000000a00 */ /*0090*/ IMAD.X R7, RZ, RZ, c[0x0][0x24], P0 ; /* 0x00000900ff077624 */ /* 0x000fe200000e06ff */ /*00a0*/ STL.64 [R1], R8 ; /* 0x0000000801007387 */ /* 0x0013e80000100a00 */ /*00b0*/ LEPC R8 ; /* 0x000000000008734e */ /* 0x002fc60000000000 */ /*00c0*/ MOV R11, 0x130 ; /* 0x00000130000b7802 */ /* 0x000fe40000000f00 */ /*00d0*/ MOV R20, 0xb0 ; /* 0x000000b000147802 */ /* 0x000fc40000000f00 */ /*00e0*/ MOV R21, 0x0 ; /* 0x0000000000157802 */ /* 0x000fe40000000f00 */ /*00f0*/ MOV R0, 0x0 ; /* 0x0000000000007802 */ /* 0x000fe40000000f00 */ /*0100*/ IADD3 R20, P0, P1, -R20, R11, R8 ; /* 0x0000000b14147210 */ /* 0x000fc8000791e108 */ /*0110*/ IADD3.X R21, ~R0, R21, R9, P0, P1 ; /* 0x0000001500157210 */ /* 0x000fc800007e2509 */ /*0120*/ CALL.ABS.NOINC R2 ; /* 0x0000000002007343 */ /* 0x004fea0003c00000 */ /*0130*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0140*/ BRA 0x140; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*01f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> __global__ void hello(){ printf("Hello from block: %u, thread: %u\n", blockIdx.x, threadIdx.x); } int main(){ hello<<<2,2>>>(); cudaDeviceSynchronize(); }
.file "tmpxft_00125a7d_00000000-6_hello_solution.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z23__device_stub__Z5hellovv .type _Z23__device_stub__Z5hellovv, @function _Z23__device_stub__Z5hellovv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z5hellov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z23__device_stub__Z5hellovv, .-_Z23__device_stub__Z5hellovv .globl _Z5hellov .type _Z5hellov, @function _Z5hellov: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z23__device_stub__Z5hellovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5hellov, .-_Z5hellov .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $2, 20(%rsp) movl $1, 24(%rsp) movl $2, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z23__device_stub__Z5hellovv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z5hellov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z5hellov(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> __global__ void hello(){ printf("Hello from block: %u, thread: %u\n", blockIdx.x, threadIdx.x); } int main(){ hello<<<2,2>>>(); cudaDeviceSynchronize(); }
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void hello(){ printf("Hello from block: %u, thread: %u\n", blockIdx.x, threadIdx.x); } int main(){ hello<<<2,2>>>(); hipDeviceSynchronize(); }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void hello(){ printf("Hello from block: %u, thread: %u\n", blockIdx.x, threadIdx.x); } int main(){ hello<<<2,2>>>(); hipDeviceSynchronize(); }
.text .file "hello_solution.hip" .globl _Z20__device_stub__hellov # -- Begin function _Z20__device_stub__hellov .p2align 4, 0x90 .type _Z20__device_stub__hellov,@function _Z20__device_stub__hellov: # @_Z20__device_stub__hellov .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z5hellov, %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 _Z20__device_stub__hellov, .Lfunc_end0-_Z20__device_stub__hellov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movabsq $4294967298, %rdi # imm = 0x100000002 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z5hellov, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5hellov, %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 _Z5hellov,@object # @_Z5hellov .section .rodata,"a",@progbits .globl _Z5hellov .p2align 3, 0x0 _Z5hellov: .quad _Z20__device_stub__hellov .size _Z5hellov, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5hellov" .size .L__unnamed_1, 10 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__hellov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5hellov .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_00125a7d_00000000-6_hello_solution.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z23__device_stub__Z5hellovv .type _Z23__device_stub__Z5hellovv, @function _Z23__device_stub__Z5hellovv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z5hellov(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z23__device_stub__Z5hellovv, .-_Z23__device_stub__Z5hellovv .globl _Z5hellov .type _Z5hellov, @function _Z5hellov: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z23__device_stub__Z5hellovv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z5hellov, .-_Z5hellov .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movl $2, 20(%rsp) movl $1, 24(%rsp) movl $2, 8(%rsp) movl $1, 12(%rsp) movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movl $1, %ecx movq 8(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L14 .L12: call cudaDeviceSynchronize@PLT movl $0, %eax addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L14: .cfi_restore_state call _Z23__device_stub__Z5hellovv jmp .L12 .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z5hellov" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z5hellov(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "hello_solution.hip" .globl _Z20__device_stub__hellov # -- Begin function _Z20__device_stub__hellov .p2align 4, 0x90 .type _Z20__device_stub__hellov,@function _Z20__device_stub__hellov: # @_Z20__device_stub__hellov .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z5hellov, %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 _Z20__device_stub__hellov, .Lfunc_end0-_Z20__device_stub__hellov .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $56, %rsp .cfi_def_cfa_offset 64 movabsq $4294967298, %rdi # imm = 0x100000002 movl $1, %esi movq %rdi, %rdx movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 32(%rsp), %rdi leaq 16(%rsp), %rsi leaq 8(%rsp), %rdx movq %rsp, %rcx callq __hipPopCallConfiguration movq 32(%rsp), %rsi movl 40(%rsp), %edx movq 16(%rsp), %rcx movl 24(%rsp), %r8d leaq 48(%rsp), %r9 movl $_Z5hellov, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_2: callq hipDeviceSynchronize xorl %eax, %eax addq $56, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5hellov, %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 _Z5hellov,@object # @_Z5hellov .section .rodata,"a",@progbits .globl _Z5hellov .p2align 3, 0x0 _Z5hellov: .quad _Z20__device_stub__hellov .size _Z5hellov, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z5hellov" .size .L__unnamed_1, 10 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__hellov .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5hellov .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.
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(d_data, h_data, sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(h_data, d_data, sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); //DTH cudaMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); //HTD cudaMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); cudaMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); cudaMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { cudaDeviceProp props; cudaGetDeviceProperties(&props, gpu); cudaSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; cudaMallocHost((void**)&h_data_benchmark, tam * sizeof(char)); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; cudaMalloc((void **)&d_data_benchmark, tam * sizeof(char)); cudaMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); cudaFreeHost(h_data_benchmark); cudaFree(d_data_benchmark); }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(d_data, h_data, sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(h_data, d_data, sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); //DTH cudaMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); //HTD cudaMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); cudaMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); cudaMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { cudaDeviceProp props; cudaGetDeviceProperties(&props, gpu); cudaSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; cudaMallocHost((void**)&h_data_benchmark, tam * sizeof(char)); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; cudaMalloc((void **)&d_data_benchmark, tam * sizeof(char)); cudaMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); cudaFreeHost(h_data_benchmark); cudaFree(d_data_benchmark); }
.file "tmpxft_0000eb66_00000000-6_microbenchmarking_transfers.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2071: .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 .LFE2071: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z4swapRfS_ .type _Z4swapRfS_, @function _Z4swapRfS_: .LFB2057: .cfi_startproc endbr64 movss (%rdi), %xmm0 movss (%rsi), %xmm1 movss %xmm1, (%rdi) movss %xmm0, (%rsi) ret .cfi_endproc .LFE2057: .size _Z4swapRfS_, .-_Z4swapRfS_ .globl _Z9partitionPfii .type _Z9partitionPfii, @function _Z9partitionPfii: .LFB2058: .cfi_startproc endbr64 movq %rdi, %rcx movl %esi, %edi movl %edx, %eax leal (%rdx,%rsi), %esi movl %esi, %edx shrl $31, %edx addl %esi, %edx sarl %edx movslq %edx, %rdx leaq (%rcx,%rdx,4), %rdx movslq %eax, %rsi leaq (%rcx,%rsi,4), %r9 movss (%r9), %xmm0 movss (%rdx), %xmm1 movss %xmm1, (%r9) movss %xmm0, (%rdx) subl $1, %eax movslq %edi, %rdx leaq (%rcx,%rdx,4), %rsi jmp .L5 .L10: cmpl %eax, %edx jg .L7 movslq %eax, %rdx leaq (%rcx,%rdx,4), %rdx movss (%rdx), %xmm0 movss %xmm0, (%r8) movss %xmm2, (%rdx) subl $1, %eax .L6: addq $4, %rsi addl $1, %edi .L5: movq %rsi, %r8 movss (%rsi), %xmm2 movss (%r9), %xmm1 comiss %xmm2, %xmm1 ja .L6 movl %edi, %edx cmpl %eax, %edi jg .L7 cltq .L8: movss (%rcx,%rax,4), %xmm0 comiss %xmm1, %xmm0 jb .L10 subq $1, %rax cmpl %eax, %edx jle .L8 .L7: movss %xmm2, (%r9) movss %xmm1, (%r8) movl %edx, %eax ret .cfi_endproc .LFE2058: .size _Z9partitionPfii, .-_Z9partitionPfii .globl _Z2qsPfii .type _Z2qsPfii, @function _Z2qsPfii: .LFB2059: .cfi_startproc endbr64 cmpl %edx, %esi jl .L17 ret .L17: 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 $8, %rsp .cfi_def_cfa_offset 48 movq %rdi, %r12 movl %esi, %ebx movl %edx, %ebp call _Z9partitionPfii movl %eax, %r13d leal 1(%rax), %esi movl %ebp, %edx movq %r12, %rdi call _Z2qsPfii leal -1(%r13), %edx movl %ebx, %esi movq %r12, %rdi call _Z2qsPfii addq $8, %rsp .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 .cfi_endproc .LFE2059: .size _Z2qsPfii, .-_Z2qsPfii .globl _Z15serialQuickSortPfi .type _Z15serialQuickSortPfi, @function _Z15serialQuickSortPfi: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leal -1(%rsi), %edx movl $0, %esi call _Z2qsPfii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _Z15serialQuickSortPfi, .-_Z15serialQuickSortPfi .globl _Z13getMedianTimePfi .type _Z13getMedianTimePfi, @function _Z13getMedianTimePfi: .LFB2061: .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 $8, %rsp .cfi_def_cfa_offset 48 movq %rdi, %r12 movl %esi, %r13d movslq %esi, %rbp salq $2, %rbp movq %rbp, %rdi call malloc@PLT movq %rax, %rbx testl %r13d, %r13d jle .L21 movl $0, %eax .L22: movss (%r12,%rax), %xmm0 movss %xmm0, (%rbx,%rax) addq $4, %rax cmpq %rax, %rbp jne .L22 .L21: movl %r13d, %esi movq %rbx, %rdi call _Z15serialQuickSortPfi testb $1, %r13b jne .L23 movl %r13d, %eax shrl $31, %eax addl %r13d, %eax sarl %eax cltq movss 4(%rbx,%rax,4), %xmm0 addss (%rbx,%rax,4), %xmm0 mulss .LC0(%rip), %xmm0 movd %xmm0, %ebp .L24: movq %rbx, %rdi call free@PLT movd %ebp, %xmm0 addq $8, %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 .L23: .cfi_restore_state movl %r13d, %eax shrl $31, %eax addl %r13d, %eax sarl %eax cltq movl (%rbx,%rax,4), %ebp jmp .L24 .cfi_endproc .LFE2061: .size _Z13getMedianTimePfi, .-_Z13getMedianTimePfi .globl _Z8getLoHTDPcS_i .type _Z8getLoHTDPcS_i, @function _Z8getLoHTDPcS_i: .LFB2062: .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 $56, %rsp .cfi_def_cfa_offset 112 movq %rdi, %r12 movq %rsi, %r13 movl %edx, %ebx movl %edx, 12(%rsp) movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax leaq 24(%rsp), %rdi call cudaEventCreate@PLT leaq 32(%rsp), %rdi call cudaEventCreate@PLT movl $0x00000000, 20(%rsp) movslq %ebx, %rbp salq $2, %rbp movq %rbp, %rdi call malloc@PLT movq %rax, %r15 movq %rbp, %rcx movq %rbp, %rdx movl $0, %esi movq %rax, %rdi call __memset_chk@PLT testl %ebx, %ebx jle .L28 movq %r15, %rbx addq %r15, %rbp leaq 20(%rsp), %r14 .L29: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movl $1, %ecx movl $1, %edx movq %r13, %rsi movq %r12, %rdi call cudaMemcpy@PLT movl $0, %esi movq 32(%rsp), %rdi call cudaEventRecord@PLT movq 32(%rsp), %rdi call cudaEventSynchronize@PLT movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq %r14, %rdi call cudaEventElapsedTime@PLT movss 20(%rsp), %xmm0 movss %xmm0, (%rbx) addq $4, %rbx cmpq %rbp, %rbx jne .L29 .L28: movl 12(%rsp), %esi movq %r15, %rdi call _Z13getMedianTimePfi movss %xmm0, 20(%rsp) movq %r15, %rdi call free@PLT movq 24(%rsp), %rdi call cudaEventDestroy@PLT movq 32(%rsp), %rdi call cudaEventDestroy@PLT movss 20(%rsp), %xmm0 movq 40(%rsp), %rax subq %fs:40, %rax jne .L33 addq $56, %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 .L33: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2062: .size _Z8getLoHTDPcS_i, .-_Z8getLoHTDPcS_i .globl _Z8getLoDTHPcS_i .type _Z8getLoDTHPcS_i, @function _Z8getLoDTHPcS_i: .LFB2063: .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 $56, %rsp .cfi_def_cfa_offset 112 movq %rdi, %r12 movq %rsi, %r13 movl %edx, %ebx movl %edx, 12(%rsp) movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax leaq 24(%rsp), %rdi call cudaEventCreate@PLT leaq 32(%rsp), %rdi call cudaEventCreate@PLT movl $0x00000000, 20(%rsp) movslq %ebx, %rbp salq $2, %rbp movq %rbp, %rdi call malloc@PLT movq %rax, %r15 movq %rbp, %rcx movq %rbp, %rdx movl $0, %esi movq %rax, %rdi call __memset_chk@PLT testl %ebx, %ebx jle .L35 movq %r15, %rbx addq %r15, %rbp leaq 20(%rsp), %r14 .L36: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movl $2, %ecx movl $1, %edx movq %r12, %rsi movq %r13, %rdi call cudaMemcpy@PLT movl $0, %esi movq 32(%rsp), %rdi call cudaEventRecord@PLT movq 32(%rsp), %rdi call cudaEventSynchronize@PLT movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq %r14, %rdi call cudaEventElapsedTime@PLT movss 20(%rsp), %xmm0 movss %xmm0, (%rbx) addq $4, %rbx cmpq %rbp, %rbx jne .L36 .L35: movl 12(%rsp), %esi movq %r15, %rdi call _Z13getMedianTimePfi movss %xmm0, 20(%rsp) movq %r15, %rdi call free@PLT movq 24(%rsp), %rdi call cudaEventDestroy@PLT movq 32(%rsp), %rdi call cudaEventDestroy@PLT movss 20(%rsp), %xmm0 movq 40(%rsp), %rax subq %fs:40, %rax jne .L40 addq $56, %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 .L40: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2063: .size _Z8getLoDTHPcS_i, .-_Z8getLoDTHPcS_i .globl _Z7getGDTHPcS_fi .type _Z7getGDTHPcS_fi, @function _Z7getGDTHPcS_fi: .LFB2064: .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 $56, %rsp .cfi_def_cfa_offset 112 movq %rdi, (%rsp) movq %rsi, %r13 movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0x00000000, 20(%rsp) leaq 24(%rsp), %rdi call cudaEventCreate@PLT leaq 32(%rsp), %rdi call cudaEventCreate@PLT movl $16, %ebp movl $0, %r14d movl $0, %ebx movl $0x00000000, 8(%rsp) leaq 20(%rsp), %r15 .L42: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movl %ebp, %r12d sall $20, %r12d movslq %r12d, %rdx movl $2, %ecx movq (%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT movl $0, %esi movq 32(%rsp), %rdi call cudaEventRecord@PLT movq 32(%rsp), %rdi call cudaEventSynchronize@PLT movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq %r15, %rdi call cudaEventElapsedTime@PLT movss 8(%rsp), %xmm2 addss 20(%rsp), %xmm2 movss %xmm2, 8(%rsp) addl %r12d, %r14d addl $1, %ebx addl %ebp, %ebp cmpl $6, %ebx jne .L42 movss 12(%rsp), %xmm0 mulss .LC2(%rip), %xmm0 movaps %xmm2, %xmm1 subss %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtsi2ssl %r14d, %xmm0 divss %xmm0, %xmm1 movd %xmm1, %r13d movq 24(%rsp), %rdi call cudaEventDestroy@PLT movq 32(%rsp), %rdi call cudaEventDestroy@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L46 movd %r13d, %xmm0 addq $56, %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 .L46: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2064: .size _Z7getGDTHPcS_fi, .-_Z7getGDTHPcS_fi .globl _Z7getGHTDPcS_fi .type _Z7getGHTDPcS_fi, @function _Z7getGHTDPcS_fi: .LFB2065: .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 $56, %rsp .cfi_def_cfa_offset 112 movq %rdi, (%rsp) movq %rsi, %r13 movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $0x00000000, 20(%rsp) leaq 24(%rsp), %rdi call cudaEventCreate@PLT leaq 32(%rsp), %rdi call cudaEventCreate@PLT movl $16, %ebp movl $0, %r14d movl $0, %ebx movl $0x00000000, 8(%rsp) leaq 20(%rsp), %r15 .L48: movl $0, %esi movq 24(%rsp), %rdi call cudaEventRecord@PLT movl %ebp, %r12d sall $20, %r12d movslq %r12d, %rdx movl $1, %ecx movq %r13, %rsi movq (%rsp), %rdi call cudaMemcpy@PLT movl $0, %esi movq 32(%rsp), %rdi call cudaEventRecord@PLT movq 32(%rsp), %rdi call cudaEventSynchronize@PLT movq 32(%rsp), %rdx movq 24(%rsp), %rsi movq %r15, %rdi call cudaEventElapsedTime@PLT movss 8(%rsp), %xmm2 addss 20(%rsp), %xmm2 movss %xmm2, 8(%rsp) addl %r12d, %r14d addl $1, %ebx addl %ebp, %ebp cmpl $6, %ebx jne .L48 movss 12(%rsp), %xmm0 mulss .LC2(%rip), %xmm0 movaps %xmm2, %xmm1 subss %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtsi2ssl %r14d, %xmm0 divss %xmm0, %xmm1 movd %xmm1, %r13d movq 24(%rsp), %rdi call cudaEventDestroy@PLT movq 32(%rsp), %rdi call cudaEventDestroy@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L52 movd %r13d, %xmm0 addq $56, %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 .L52: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2065: .size _Z7getGHTDPcS_fi, .-_Z7getGHTDPcS_fi .globl _Z17getOverlappedGDTHPcS_fi .type _Z17getOverlappedGDTHPcS_fi, @function _Z17getOverlappedGDTHPcS_fi: .LFB2066: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %rdi, (%rsp) movq %rsi, 8(%rsp) movss %xmm0, 28(%rsp) movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $0x00000000, 36(%rsp) movl $16, %edi call malloc@PLT movq %rax, %rbx movq %rax, %rdi call cudaStreamCreate@PLT leaq 8(%rbx), %rdi call cudaStreamCreate@PLT leaq 40(%rsp), %rdi call cudaEventCreate@PLT leaq 48(%rsp), %rdi call cudaEventCreate@PLT movl $16, %r13d movl $0, %r15d movl $0, %r12d movl $0x00000000, 24(%rsp) leaq 36(%rsp), %rax movq %rax, 16(%rsp) .L54: movq (%rbx), %rsi movq 40(%rsp), %rdi call cudaEventRecord@PLT movl %r13d, %r14d sall $20, %r14d movslq %r14d, %rbp movq (%rbx), %r8 movl $2, %ecx movq %rbp, %rdx movq (%rsp), %rsi movq 8(%rsp), %rdi call cudaMemcpyAsync@PLT movq (%rbx), %rsi movq 48(%rsp), %rdi call cudaEventRecord@PLT movq 8(%rsp), %rax leaq (%rax,%rbp), %rsi movq (%rsp), %rax leaq (%rax,%rbp), %rdi movq 8(%rbx), %r8 movl $1, %ecx movq %rbp, %rdx call cudaMemcpyAsync@PLT call cudaDeviceSynchronize@PLT movq 48(%rsp), %rdx movq 40(%rsp), %rsi movq 16(%rsp), %rdi call cudaEventElapsedTime@PLT movss 24(%rsp), %xmm2 addss 36(%rsp), %xmm2 movss %xmm2, 24(%rsp) addl %r14d, %r15d addl $1, %r12d addl %r13d, %r13d cmpl $6, %r12d jne .L54 movss 28(%rsp), %xmm0 mulss .LC2(%rip), %xmm0 movaps %xmm2, %xmm1 subss %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtsi2ssl %r15d, %xmm0 divss %xmm0, %xmm1 movd %xmm1, %r15d movq (%rbx), %rdi call cudaStreamDestroy@PLT movq 8(%rbx), %rdi call cudaStreamDestroy@PLT movq %rbx, %rdi call free@PLT movq 40(%rsp), %rdi call cudaEventDestroy@PLT movq 48(%rsp), %rdi call cudaEventDestroy@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L58 movd %r15d, %xmm0 addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L58: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2066: .size _Z17getOverlappedGDTHPcS_fi, .-_Z17getOverlappedGDTHPcS_fi .globl _Z17getOverlappedGHTDPcS_fi .type _Z17getOverlappedGHTDPcS_fi, @function _Z17getOverlappedGHTDPcS_fi: .LFB2067: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $72, %rsp .cfi_def_cfa_offset 128 movq %rdi, (%rsp) movq %rsi, 8(%rsp) movss %xmm0, 28(%rsp) movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax movl $0x00000000, 36(%rsp) movl $16, %edi call malloc@PLT movq %rax, %rbx movq %rax, %rdi call cudaStreamCreate@PLT leaq 8(%rbx), %rdi call cudaStreamCreate@PLT leaq 40(%rsp), %rdi call cudaEventCreate@PLT leaq 48(%rsp), %rdi call cudaEventCreate@PLT movl $16, %r13d movl $0, %r15d movl $0, %r12d movl $0x00000000, 24(%rsp) leaq 36(%rsp), %rax movq %rax, 16(%rsp) .L60: movq (%rbx), %rsi movq 40(%rsp), %rdi call cudaEventRecord@PLT movl %r13d, %r14d sall $20, %r14d movslq %r14d, %rbp movq (%rbx), %r8 movl $1, %ecx movq %rbp, %rdx movq 8(%rsp), %rsi movq (%rsp), %rdi call cudaMemcpyAsync@PLT movq (%rbx), %rsi movq 48(%rsp), %rdi call cudaEventRecord@PLT movq (%rsp), %rax leaq (%rax,%rbp), %rsi movq 8(%rsp), %rax leaq (%rax,%rbp), %rdi movq 8(%rbx), %r8 movl $2, %ecx movq %rbp, %rdx call cudaMemcpyAsync@PLT call cudaDeviceSynchronize@PLT movq 48(%rsp), %rdx movq 40(%rsp), %rsi movq 16(%rsp), %rdi call cudaEventElapsedTime@PLT movss 24(%rsp), %xmm2 addss 36(%rsp), %xmm2 movss %xmm2, 24(%rsp) addl %r14d, %r15d addl $1, %r12d addl %r13d, %r13d cmpl $6, %r12d jne .L60 movss 28(%rsp), %xmm0 mulss .LC2(%rip), %xmm0 movaps %xmm2, %xmm1 subss %xmm0, %xmm1 pxor %xmm0, %xmm0 cvtsi2ssl %r15d, %xmm0 divss %xmm0, %xmm1 movd %xmm1, %r15d movq (%rbx), %rdi call cudaStreamDestroy@PLT movq 8(%rbx), %rdi call cudaStreamDestroy@PLT movq %rbx, %rdi call free@PLT movq 40(%rsp), %rdi call cudaEventDestroy@PLT movq 48(%rsp), %rdi call cudaEventDestroy@PLT movq 56(%rsp), %rax subq %fs:40, %rax jne .L64 movd %r15d, %xmm0 addq $72, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L64: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2067: .size _Z17getOverlappedGHTDPcS_fi, .-_Z17getOverlappedGHTDPcS_fi .globl _Z20microbenchmarkingPCIiPfS_S_S_S_S_i .type _Z20microbenchmarkingPCIiPfS_S_S_S_S_i, @function _Z20microbenchmarkingPCIiPfS_S_S_S_S_i: .LFB2068: .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 $1080, %rsp .cfi_def_cfa_offset 1136 movl %edi, %ebp movq %rsi, %r13 movq %rdx, %r12 movq %rcx, %r14 movq %r8, 8(%rsp) movq %r9, (%rsp) movq 1136(%rsp), %r15 movl 1144(%rsp), %ebx movq %fs:40, %rax movq %rax, 1064(%rsp) xorl %eax, %eax leaq 32(%rsp), %rdi movl %ebp, %esi call cudaGetDeviceProperties_v2@PLT movl %ebp, %edi call cudaSetDevice@PLT leaq 16(%rsp), %rdi movl $1073741824, %esi call cudaMallocHost@PLT movl $1073741824, %edx movl $0, %esi movq 16(%rsp), %rdi call memset@PLT leaq 24(%rsp), %rdi movl $1073741824, %esi call cudaMalloc@PLT movl $1073741824, %edx movl $0, %esi movq 24(%rsp), %rdi call cudaMemset@PLT movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z8getLoHTDPcS_i movss %xmm0, 0(%r13) movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z7getGHTDPcS_fi movss %xmm0, (%r14) cmpl $2, 632(%rsp) je .L70 .L66: movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z8getLoDTHPcS_i movss %xmm0, (%r12) movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z7getGDTHPcS_fi movq (%rsp), %rax movss %xmm0, (%rax) cmpl $2, 632(%rsp) je .L71 .L67: movq 16(%rsp), %rdi call cudaFreeHost@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 1064(%rsp), %rax subq %fs:40, %rax jne .L72 addq $1080, %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 .L70: .cfi_restore_state movss 0(%r13), %xmm0 movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z17getOverlappedGHTDPcS_fi movq 8(%rsp), %rax movss %xmm0, (%rax) jmp .L66 .L71: movss (%r12), %xmm0 movl %ebx, %edx movq 16(%rsp), %rsi movq 24(%rsp), %rdi call _Z17getOverlappedGDTHPcS_fi movss %xmm0, (%r15) jmp .L67 .L72: call __stack_chk_fail@PLT .cfi_endproc .LFE2068: .size _Z20microbenchmarkingPCIiPfS_S_S_S_S_i, .-_Z20microbenchmarkingPCIiPfS_S_S_S_S_i .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2094: .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 .LFE2094: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC0: .long 1056964608 .align 4 .LC2: .long 1086324736 .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.
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(d_data, h_data, sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { cudaEventRecord(start_event, 0); //we only transfer 1 byte. cudaMemcpy(h_data, d_data, sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, 0); cudaMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice); cudaEventRecord(stop_event, 0); cudaEventSynchronize(stop_event); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); //DTH cudaMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); //HTD cudaMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; cudaStream_t *stream_benchmark = (cudaStream_t *)malloc(2 * sizeof(cudaStream_t)); for(int i = 0; i < 2; i++) cudaStreamCreate(&(stream_benchmark[i])); cudaEvent_t start_event, stop_event; cudaEventCreate(&start_event); cudaEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { cudaEventRecord(start_event, stream_benchmark[0]); cudaMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), cudaMemcpyHostToDevice, stream_benchmark[0]); cudaEventRecord(stop_event, stream_benchmark[0]); cudaMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), cudaMemcpyDeviceToHost, stream_benchmark[1]); cudaDeviceSynchronize(); cudaEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) cudaStreamDestroy(stream_benchmark[i]); free(stream_benchmark); cudaEventDestroy(start_event); cudaEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { cudaDeviceProp props; cudaGetDeviceProperties(&props, gpu); cudaSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; cudaMallocHost((void**)&h_data_benchmark, tam * sizeof(char)); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; cudaMalloc((void **)&d_data_benchmark, tam * sizeof(char)); cudaMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); cudaFreeHost(h_data_benchmark); cudaFree(d_data_benchmark); }
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <hip/hip_runtime.h> #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(d_data, h_data, sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(h_data, d_data, sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); //DTH hipMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); //HTD hipMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); hipMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); hipMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { hipDeviceProp_t props; hipGetDeviceProperties(&props, gpu); hipSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; hipHostMalloc((void**)&h_data_benchmark, tam * sizeof(char), hipHostMallocDefault); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; hipMalloc((void **)&d_data_benchmark, tam * sizeof(char)); hipMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); hipHostFree(h_data_benchmark); hipFree(d_data_benchmark); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <hip/hip_runtime.h> #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(d_data, h_data, sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(h_data, d_data, sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); //DTH hipMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); //HTD hipMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); hipMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); hipMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { hipDeviceProp_t props; hipGetDeviceProperties(&props, gpu); hipSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; hipHostMalloc((void**)&h_data_benchmark, tam * sizeof(char), hipHostMallocDefault); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; hipMalloc((void **)&d_data_benchmark, tam * sizeof(char)); hipMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); hipHostFree(h_data_benchmark); hipFree(d_data_benchmark); }
.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.
/** * @file microbenchmarking_transfers.cu * @detail This file describes the implementation of the functions involved in the * PCIe bandwidth and latency. * @author Antonio Jose Lazaro Munoz. * @date 20/02/2016 */ #include <hip/hip_runtime.h> #include <stdio.h> void swap(float & v1, float & v2){ float tmp = v1; v1 = v2; v2 = tmp; } int partition(float *array, int left, int right){ int part = right; swap(array[part],array[(right+left) / 2]); --right; while(true){ while(array[left] < array[part]){ ++left; } while(right >= left && array[part] <= array[right]){ --right; } if(right < left) break; swap(array[left],array[right]); ++left; --right; } swap(array[part],array[left]); return left; } void qs(float * array, const int left, const int right){ if(left < right){ const int part = partition(array, left, right); qs(array, part + 1,right); qs(array, left,part - 1); } } /** * @brief Quicksort * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details Quicksort * * @param array elements array * @param size array size */ void serialQuickSort(float *array, const int size){ qs(array, 0,size-1); } /** * @brief Calculate times median * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the median from a set of times. * * @param h_times times array. * @param N array size. * * @return time median. */ float getMedianTime(float *h_times, int N) { float median = 0; float * h_sorted_times = (float *)malloc(N * sizeof(float)); for(int n = 0; n < N; n++) h_sorted_times[n] = h_times[n]; //Sort execution times serialQuickSort(h_sorted_times, N); //Calculate median if(N%2 == 0) { median = (h_sorted_times[N/2] + h_sorted_times[(N/2)+1])/2; } else { int p = N/2; median = h_sorted_times[p]; } free(h_sorted_times); return median; } /** * @brief Calculate PCIe HTD latency. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for HTD memory transfers. * * @param d_data device data. * @param h_data host data. * @param nreps iterations. * @return HTD PCIe latency (ms). */ float getLoHTD(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoHTD = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(d_data, h_data, sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoHTD, start_event, stop_event); h_times[k] = LoHTD; } LoHTD = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoHTD; } /** * @brief Calculate PCIe DTH latency * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the latency of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param nreps iterations * @return DTH PCIe latency (ms). */ float getLoDTH(char *d_data, char *h_data, int nreps) { hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); float LoDTH = 0.0; float *h_times = (float *)malloc(nreps * sizeof(float)); memset(h_times, 0, nreps * sizeof(float)); for(int k = 0; k < nreps; k++) { hipEventRecord(start_event, 0); //we only transfer 1 byte. hipMemcpy(h_data, d_data, sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&LoDTH, start_event, stop_event); h_times[k] = LoDTH; } LoDTH = getMedianTime(h_times, nreps); free(h_times); hipEventDestroy(start_event); hipEventDestroy(stop_event); return LoDTH; } /** * @brief DTH PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return DTH PCIe bandwidth (ms/byte). */ float getGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief HTD PCIe bandwidth (ms/byte). * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return HTD PCIe bandwidth (ms/byte). */ float getGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, 0); hipMemcpy(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice); hipEventRecord(stop_event, 0); hipEventSynchronize(stop_event); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief Overlap DTH PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for HTD memory transfers, when a DTH memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoDTH DTH PCIe latency. * @param nreps Iterations. * @return Overlap DTH PCIe bandwidth (byte/ms). */ float getOverlappedGDTH(char *d_data, char *h_data, float LoDTH, int nreps) { float time = 0; float GDTH = 0.0; float timeSumGDTH = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); //DTH hipMemcpyAsync(h_data, d_data, size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); //HTD hipMemcpyAsync(d_data + (size * 1024 * 1024), h_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGDTH += time; total_bytes += size * 1024 * 1024; n++; } GDTH = (timeSumGDTH - n*LoDTH)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GDTH; } /** * @brief Overlap HTD PCIe bandwidth. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function returns the bandwidth of the PCIe for DTH memory transfers, when a HTD memory transfer * is concurrently executed. * * @param d_data Device data. * @param h_data Host data. * @param LoHTD HTD PCIe latency. * @param nreps Iterations. * @return Overlap HTD PCIe bandwidth (byte/ms). */ float getOverlappedGHTD(char *d_data, char *h_data, float LoHTD, int nreps) { float time = 0; float GHTD = 0.0; float timeSumGHTD = 0.0; hipStream_t *stream_benchmark = (hipStream_t *)malloc(2 * sizeof(hipStream_t)); for(int i = 0; i < 2; i++) hipStreamCreate(&(stream_benchmark[i])); hipEvent_t start_event, stop_event; hipEventCreate(&start_event); hipEventCreate(&stop_event); int n = 0; int total_bytes = 0; //From 16 MB to 512 MB for(int size = 16; size <= 512; size=size*2) { hipEventRecord(start_event, stream_benchmark[0]); hipMemcpyAsync(d_data, h_data, size * 1024 * 1024* sizeof(char), hipMemcpyHostToDevice, stream_benchmark[0]); hipEventRecord(stop_event, stream_benchmark[0]); hipMemcpyAsync(h_data + (size * 1024 * 1024), d_data + (size * 1024 * 1024), size * 1024 * 1024* sizeof(char), hipMemcpyDeviceToHost, stream_benchmark[1]); hipDeviceSynchronize(); hipEventElapsedTime(&time, start_event, stop_event); timeSumGHTD += time; total_bytes += size * 1024 * 1024; n++; } GHTD = (timeSumGHTD - n*LoHTD)/total_bytes; for(int i = 0; i < 2; i++) hipStreamDestroy(stream_benchmark[i]); free(stream_benchmark); hipEventDestroy(start_event); hipEventDestroy(stop_event); return GHTD; } /** * @brief PCIe microbenchmarking. * @author Antonio Jose Lazaro Munoz * @date 17/02/2016 * @details This function calculates the values of PCIe latency and bandwidth * for HTD and DTH memory transfers. * * @param gpu GPU id. * @param LoHTD PCIe HTD latency (pointer). * @param LoDTH PCIe DTH latency (pointer). * @param GHTD HTD PCIe bandwidth (pointer). * @param overlappedGHTD Overlap HTD PCIe bandwidth (pointer). * @param GDTH DTH PCIe bandwidth (pointer). * @param overlappedGDTH Overlap DTH PCIe bandwidth (pointer). * @param nIter Iterations. */ void microbenchmarkingPCI(int gpu, float *LoHTD, float *LoDTH, float *GHTD, float *overlappedGHTD, float *GDTH, float *overlappedGDTH, int nIter) { hipDeviceProp_t props; hipGetDeviceProperties(&props, gpu); hipSetDevice(gpu); int tam = 1024*1024*1024; // 1GB char *h_data_benchmark; hipHostMalloc((void**)&h_data_benchmark, tam * sizeof(char), hipHostMallocDefault); memset(h_data_benchmark, 0, tam * sizeof(char)); char *d_data_benchmark; hipMalloc((void **)&d_data_benchmark, tam * sizeof(char)); hipMemset(d_data_benchmark, 0, tam * sizeof(char)); *LoHTD = getLoHTD(d_data_benchmark, h_data_benchmark, nIter); *GHTD = getGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); if(props.asyncEngineCount == 2) *overlappedGHTD = getOverlappedGHTD(d_data_benchmark, h_data_benchmark, *LoHTD, nIter); *LoDTH = getLoDTH(d_data_benchmark, h_data_benchmark, nIter); *GDTH = getGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); if(props.asyncEngineCount == 2) *overlappedGDTH = getOverlappedGDTH(d_data_benchmark, h_data_benchmark, *LoDTH, nIter); hipHostFree(h_data_benchmark); hipFree(d_data_benchmark); }
.text .file "microbenchmarking_transfers.hip" .globl _Z4swapRfS_ # -- Begin function _Z4swapRfS_ .p2align 4, 0x90 .type _Z4swapRfS_,@function _Z4swapRfS_: # @_Z4swapRfS_ .cfi_startproc # %bb.0: movss (%rdi), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%rsi), %xmm1 # xmm1 = mem[0],zero,zero,zero movss %xmm1, (%rdi) movss %xmm0, (%rsi) retq .Lfunc_end0: .size _Z4swapRfS_, .Lfunc_end0-_Z4swapRfS_ .cfi_endproc # -- End function .globl _Z9partitionPfii # -- Begin function _Z9partitionPfii .p2align 4, 0x90 .type _Z9partitionPfii,@function _Z9partitionPfii: # @_Z9partitionPfii .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 # kill: def $edx killed $edx def $rdx movl %esi, %eax movabsq $-4294967296, %rsi # imm = 0xFFFFFFFF00000000 movslq %edx, %rcx leal (%rcx,%rax), %r8d movl %r8d, %r9d shrl $31, %r9d addl %r8d, %r9d sarl %r9d movslq %r9d, %r8 movss (%rdi,%rcx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%rdi,%r8,4), %xmm1 # xmm1 = mem[0],zero,zero,zero movss %xmm1, (%rdi,%rcx,4) movss %xmm0, (%rdi,%r8,4) leaq -4(%rdi), %r8 movabsq $4294967296, %r9 # imm = 0x100000000 .LBB1_1: # =>This Loop Header: Depth=1 # Child Loop BB1_2 Depth 2 # Child Loop BB1_4 Depth 2 movss (%rdi,%rcx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movslq %eax, %r11 leaq (%r8,%r11,4), %r10 shlq $32, %r11 addq %rsi, %r11 .p2align 4, 0x90 .LBB1_2: # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 addq %r9, %r11 movss 4(%r10), %xmm1 # xmm1 = mem[0],zero,zero,zero addq $4, %r10 incl %eax ucomiss %xmm1, %xmm0 ja .LBB1_2 # %bb.3: # %.preheader # in Loop: Header=BB1_1 Depth=1 movslq %edx, %rdx sarq $32, %r11 movq %rdx, %rbx shlq $32, %rbx .p2align 4, 0x90 .LBB1_4: # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 cmpq %r11, %rdx jle .LBB1_8 # %bb.5: # in Loop: Header=BB1_4 Depth=2 movq %rdx, %r14 decq %rdx movss (%r8,%r14,4), %xmm2 # xmm2 = mem[0],zero,zero,zero addq %rsi, %rbx ucomiss %xmm0, %xmm2 jae .LBB1_4 # %bb.6: # %.critedge # in Loop: Header=BB1_1 Depth=1 cmpq %r11, %r14 jle .LBB1_8 # %bb.7: # in Loop: Header=BB1_1 Depth=1 sarq $30, %rbx movss (%rdi,%rbx), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, (%r10) movss %xmm1, (%rdi,%rbx) jmp .LBB1_1 .LBB1_8: # %.critedge33 movss %xmm1, (%rdi,%rcx,4) movss %xmm0, (%r10) decl %eax # kill: def $eax killed $eax killed $rax popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z9partitionPfii, .Lfunc_end1-_Z9partitionPfii .cfi_endproc # -- End function .globl _Z2qsPfii # -- Begin function _Z2qsPfii .p2align 4, 0x90 .type _Z2qsPfii,@function _Z2qsPfii: # @_Z2qsPfii .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 pushq %rax .cfi_def_cfa_offset 64 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 # kill: def $edx killed $edx def $rdx movl %esi, %ebx movq %rdi, %r14 movabsq $-4294967296, %r15 # imm = 0xFFFFFFFF00000000 leaq -4(%rdi), %r12 movabsq $4294967296, %r13 # imm = 0x100000000 jmp .LBB2_1 .p2align 4, 0x90 .LBB2_9: # %_Z9partitionPfii.exit # in Loop: Header=BB2_1 Depth=1 movss %xmm1, (%r14,%rax,4) movss %xmm0, (%rsi) movq %r14, %rdi movl %ebp, %esi # kill: def $edx killed $edx killed $rdx callq _Z2qsPfii addl $-2, %ebp movl %ebp, %edx .LBB2_1: # %tailrecurse # =>This Loop Header: Depth=1 # Child Loop BB2_3 Depth 2 # Child Loop BB2_4 Depth 3 # Child Loop BB2_6 Depth 3 cmpl %edx, %ebx jge .LBB2_10 # %bb.2: # in Loop: Header=BB2_1 Depth=1 movslq %edx, %rax leal (%rdx,%rbx), %ecx movl %ecx, %esi shrl $31, %esi addl %ecx, %esi sarl %esi movslq %esi, %rcx movss (%r14,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss (%r14,%rcx,4), %xmm1 # xmm1 = mem[0],zero,zero,zero movss %xmm1, (%r14,%rax,4) movss %xmm0, (%r14,%rcx,4) movl %edx, %ecx movl %ebx, %ebp .LBB2_3: # Parent Loop BB2_1 Depth=1 # => This Loop Header: Depth=2 # Child Loop BB2_4 Depth 3 # Child Loop BB2_6 Depth 3 movss (%r14,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movslq %ebp, %rdi leaq (%r12,%rdi,4), %rsi shlq $32, %rdi addq %r15, %rdi .p2align 4, 0x90 .LBB2_4: # Parent Loop BB2_1 Depth=1 # Parent Loop BB2_3 Depth=2 # => This Inner Loop Header: Depth=3 addq %r13, %rdi movss 4(%rsi), %xmm1 # xmm1 = mem[0],zero,zero,zero addq $4, %rsi incl %ebp ucomiss %xmm1, %xmm0 ja .LBB2_4 # %bb.5: # %.preheader.i # in Loop: Header=BB2_3 Depth=2 movslq %ecx, %rcx sarq $32, %rdi movq %rcx, %r8 shlq $32, %r8 .p2align 4, 0x90 .LBB2_6: # Parent Loop BB2_1 Depth=1 # Parent Loop BB2_3 Depth=2 # => This Inner Loop Header: Depth=3 cmpq %rdi, %rcx jle .LBB2_9 # %bb.7: # in Loop: Header=BB2_6 Depth=3 movss -4(%r14,%rcx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero decq %rcx addq %r15, %r8 ucomiss %xmm0, %xmm2 jae .LBB2_6 # %bb.8: # %.critedge.i # in Loop: Header=BB2_3 Depth=2 sarq $30, %r8 movss (%r14,%r8), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, (%rsi) movss %xmm1, (%r14,%r8) jmp .LBB2_3 .LBB2_10: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z2qsPfii, .Lfunc_end2-_Z2qsPfii .cfi_endproc # -- End function .globl _Z15serialQuickSortPfi # -- Begin function _Z15serialQuickSortPfi .p2align 4, 0x90 .type _Z15serialQuickSortPfi,@function _Z15serialQuickSortPfi: # @_Z15serialQuickSortPfi .cfi_startproc # %bb.0: # kill: def $esi killed $esi def $rsi leal -1(%rsi), %edx xorl %esi, %esi jmp _Z2qsPfii # TAILCALL .Lfunc_end3: .size _Z15serialQuickSortPfi, .Lfunc_end3-_Z15serialQuickSortPfi .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z13getMedianTimePfi .LCPI4_0: .long 0x3f000000 # float 0.5 .text .globl _Z13getMedianTimePfi .p2align 4, 0x90 .type _Z13getMedianTimePfi,@function _Z13getMedianTimePfi: # @_Z13getMedianTimePfi .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 pushq %rax .cfi_def_cfa_offset 48 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl %esi, %r14d movq %rdi, %r15 movslq %esi, %r12 leaq (,%r12,4), %rdi callq malloc movq %rax, %rbx testl %r12d, %r12d jle .LBB4_2 # %bb.1: # %.lr.ph.preheader movl %r14d, %edx shlq $2, %rdx movq %rbx, %rdi movq %r15, %rsi callq memcpy@PLT .LBB4_2: # %._crit_edge leal -1(%r14), %edx movq %rbx, %rdi xorl %esi, %esi callq _Z2qsPfii testb $1, %r14b jne .LBB4_4 # %bb.3: sarl %r14d movslq %r14d, %rax movss (%rbx,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss 4(%rbx,%rax,4), %xmm0 mulss .LCPI4_0(%rip), %xmm0 jmp .LBB4_5 .LBB4_4: movl %r14d, %eax shrl $31, %eax addl %r14d, %eax sarl %eax cltq movss (%rbx,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero .LBB4_5: movss %xmm0, 4(%rsp) # 4-byte Spill movq %rbx, %rdi callq free movss 4(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addq $8, %rsp .cfi_def_cfa_offset 40 popq %rbx .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end4: .size _Z13getMedianTimePfi, .Lfunc_end4-_Z13getMedianTimePfi .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z8getLoHTDPcS_i .LCPI5_0: .long 0x3f000000 # float 0.5 .text .globl _Z8getLoHTDPcS_i .p2align 4, 0x90 .type _Z8getLoHTDPcS_i,@function _Z8getLoHTDPcS_i: # @_Z8getLoHTDPcS_i .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $40, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl %edx, %r13d movq %rsi, %r15 movq %rdi, %r12 leaq 24(%rsp), %rdi callq hipEventCreate leaq 16(%rsp), %rdi callq hipEventCreate movl $0, 4(%rsp) movslq %r13d, %r14 leaq (,%r14,4), %rbp movq %rbp, %rdi callq malloc movq %rax, %rbx movq %rax, %rdi xorl %esi, %esi movq %rbp, 32(%rsp) # 8-byte Spill movq %rbp, %rdx callq memset@PLT movq %r13, 8(%rsp) # 8-byte Spill movl %r13d, %ebp testl %r14d, %r14d jle .LBB5_3 # %bb.1: # %.lr.ph.preheader xorl %r14d, %r14d leaq 4(%rsp), %r13 .p2align 4, 0x90 .LBB5_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movq 24(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movl $1, %edx movq %r12, %rdi movq %r15, %rsi movl $1, %ecx callq hipMemcpy movq 16(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 16(%rsp), %rdi callq hipEventSynchronize movq 24(%rsp), %rsi movq 16(%rsp), %rdx movq %r13, %rdi callq hipEventElapsedTime movss 4(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, (%rbx,%r14,4) incq %r14 cmpq %r14, %rbp jne .LBB5_2 .LBB5_3: # %._crit_edge movq 32(%rsp), %rdi # 8-byte Reload callq malloc movq %rax, %r15 movq 8(%rsp), %r14 # 8-byte Reload testl %r14d, %r14d jle .LBB5_5 # %bb.4: # %.lr.ph.preheader.i shlq $2, %rbp movq %r15, %rdi movq %rbx, %rsi movq %rbp, %rdx callq memcpy@PLT .LBB5_5: # %._crit_edge.i leal -1(%r14), %edx movq %r15, %rdi xorl %esi, %esi callq _Z2qsPfii testb $1, %r14b jne .LBB5_7 # %bb.6: sarl %r14d movslq %r14d, %rax movss (%r15,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss 4(%r15,%rax,4), %xmm0 mulss .LCPI5_0(%rip), %xmm0 jmp .LBB5_8 .LBB5_7: movl %r14d, %eax shrl $31, %eax addl %r14d, %eax sarl %eax cltq movss (%r15,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero .LBB5_8: # %_Z13getMedianTimePfi.exit movss %xmm0, 8(%rsp) # 4-byte Spill movq %r15, %rdi callq free movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero movss %xmm0, 4(%rsp) movq %rbx, %rdi callq free movq 24(%rsp), %rdi callq hipEventDestroy movq 16(%rsp), %rdi callq hipEventDestroy movss 4(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end5: .size _Z8getLoHTDPcS_i, .Lfunc_end5-_Z8getLoHTDPcS_i .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z8getLoDTHPcS_i .LCPI6_0: .long 0x3f000000 # float 0.5 .text .globl _Z8getLoDTHPcS_i .p2align 4, 0x90 .type _Z8getLoDTHPcS_i,@function _Z8getLoDTHPcS_i: # @_Z8getLoDTHPcS_i .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $40, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movl %edx, %r13d movq %rsi, %r15 movq %rdi, %r12 leaq 24(%rsp), %rdi callq hipEventCreate leaq 16(%rsp), %rdi callq hipEventCreate movl $0, 4(%rsp) movslq %r13d, %r14 leaq (,%r14,4), %rbp movq %rbp, %rdi callq malloc movq %rax, %rbx movq %rax, %rdi xorl %esi, %esi movq %rbp, 32(%rsp) # 8-byte Spill movq %rbp, %rdx callq memset@PLT movq %r13, 8(%rsp) # 8-byte Spill movl %r13d, %ebp testl %r14d, %r14d jle .LBB6_3 # %bb.1: # %.lr.ph.preheader xorl %r14d, %r14d leaq 4(%rsp), %r13 .p2align 4, 0x90 .LBB6_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 movq 24(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movl $1, %edx movq %r15, %rdi movq %r12, %rsi movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 16(%rsp), %rdi callq hipEventSynchronize movq 24(%rsp), %rsi movq 16(%rsp), %rdx movq %r13, %rdi callq hipEventElapsedTime movss 4(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, (%rbx,%r14,4) incq %r14 cmpq %r14, %rbp jne .LBB6_2 .LBB6_3: # %._crit_edge movq 32(%rsp), %rdi # 8-byte Reload callq malloc movq %rax, %r15 movq 8(%rsp), %r14 # 8-byte Reload testl %r14d, %r14d jle .LBB6_5 # %bb.4: # %.lr.ph.preheader.i shlq $2, %rbp movq %r15, %rdi movq %rbx, %rsi movq %rbp, %rdx callq memcpy@PLT .LBB6_5: # %._crit_edge.i leal -1(%r14), %edx movq %r15, %rdi xorl %esi, %esi callq _Z2qsPfii testb $1, %r14b jne .LBB6_7 # %bb.6: sarl %r14d movslq %r14d, %rax movss (%r15,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero addss 4(%r15,%rax,4), %xmm0 mulss .LCPI6_0(%rip), %xmm0 jmp .LBB6_8 .LBB6_7: movl %r14d, %eax shrl $31, %eax addl %r14d, %eax sarl %eax cltq movss (%r15,%rax,4), %xmm0 # xmm0 = mem[0],zero,zero,zero .LBB6_8: # %_Z13getMedianTimePfi.exit movss %xmm0, 8(%rsp) # 4-byte Spill movq %r15, %rdi callq free movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero movss %xmm0, 4(%rsp) movq %rbx, %rdi callq free movq 24(%rsp), %rdi callq hipEventDestroy movq 16(%rsp), %rdi callq hipEventDestroy movss 4(%rsp), %xmm0 # xmm0 = mem[0],zero,zero,zero addq $40, %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_end6: .size _Z8getLoDTHPcS_i, .Lfunc_end6-_Z8getLoDTHPcS_i .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z7getGDTHPcS_fi .LCPI7_0: .long 0xc0c00000 # float -6 .LCPI7_1: .long 0x4e7c0000 # float 1.05696461E+9 .text .globl _Z7getGDTHPcS_fi .p2align 4, 0x90 .type _Z7getGDTHPcS_fi,@function _Z7getGDTHPcS_fi: # @_Z7getGDTHPcS_fi .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $32, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movss %xmm0, 28(%rsp) # 4-byte Spill movq %rsi, %rbx movq %rdi, %r14 movl $0, 4(%rsp) leaq 16(%rsp), %rdi callq hipEventCreate leaq 8(%rsp), %rdi callq hipEventCreate xorps %xmm1, %xmm1 movl $16, %ebp movl $6, %r12d leaq 4(%rsp), %r15 .p2align 4, 0x90 .LBB7_1: # =>This Inner Loop Header: Depth=1 movss %xmm1, (%rsp) # 4-byte Spill movq 16(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movl %ebp, %edx shll $20, %edx movq %rbx, %rdi movq %r14, %rsi movl $2, %ecx callq hipMemcpy movq 8(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 8(%rsp), %rdi callq hipEventSynchronize movq 16(%rsp), %rsi movq 8(%rsp), %rdx movq %r15, %rdi callq hipEventElapsedTime movss (%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero addss 4(%rsp), %xmm1 addl %ebp, %ebp decl %r12d jne .LBB7_1 # %bb.2: movss 28(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero mulss .LCPI7_0(%rip), %xmm0 addss %xmm0, %xmm1 divss .LCPI7_1(%rip), %xmm1 movss %xmm1, (%rsp) # 4-byte Spill movq 16(%rsp), %rdi callq hipEventDestroy movq 8(%rsp), %rdi callq hipEventDestroy movss (%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addq $32, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end7: .size _Z7getGDTHPcS_fi, .Lfunc_end7-_Z7getGDTHPcS_fi .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z7getGHTDPcS_fi .LCPI8_0: .long 0xc0c00000 # float -6 .LCPI8_1: .long 0x4e7c0000 # float 1.05696461E+9 .text .globl _Z7getGHTDPcS_fi .p2align 4, 0x90 .type _Z7getGHTDPcS_fi,@function _Z7getGHTDPcS_fi: # @_Z7getGHTDPcS_fi .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $32, %rsp .cfi_def_cfa_offset 80 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movss %xmm0, 28(%rsp) # 4-byte Spill movq %rsi, %rbx movq %rdi, %r14 movl $0, 4(%rsp) leaq 16(%rsp), %rdi callq hipEventCreate leaq 8(%rsp), %rdi callq hipEventCreate xorps %xmm1, %xmm1 movl $16, %ebp movl $6, %r12d leaq 4(%rsp), %r15 .p2align 4, 0x90 .LBB8_1: # =>This Inner Loop Header: Depth=1 movss %xmm1, (%rsp) # 4-byte Spill movq 16(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movl %ebp, %edx shll $20, %edx movq %r14, %rdi movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi xorl %esi, %esi callq hipEventRecord movq 8(%rsp), %rdi callq hipEventSynchronize movq 16(%rsp), %rsi movq 8(%rsp), %rdx movq %r15, %rdi callq hipEventElapsedTime movss (%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero addss 4(%rsp), %xmm1 addl %ebp, %ebp decl %r12d jne .LBB8_1 # %bb.2: movss 28(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero mulss .LCPI8_0(%rip), %xmm0 addss %xmm0, %xmm1 divss .LCPI8_1(%rip), %xmm1 movss %xmm1, (%rsp) # 4-byte Spill movq 16(%rsp), %rdi callq hipEventDestroy movq 8(%rsp), %rdi callq hipEventDestroy movss (%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addq $32, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end8: .size _Z7getGHTDPcS_fi, .Lfunc_end8-_Z7getGHTDPcS_fi .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z17getOverlappedGDTHPcS_fi .LCPI9_0: .long 0xc0c00000 # float -6 .LCPI9_1: .long 0x4e7c0000 # float 1.05696461E+9 .text .globl _Z17getOverlappedGDTHPcS_fi .p2align 4, 0x90 .type _Z17getOverlappedGDTHPcS_fi,@function _Z17getOverlappedGDTHPcS_fi: # @_Z17getOverlappedGDTHPcS_fi .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 $40, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movss %xmm0, 36(%rsp) # 4-byte Spill movq %rsi, %r14 movq %rdi, %r15 movl $0, 12(%rsp) movl $16, %edi callq malloc movq %rax, %rbx xorl %r12d, %r12d .p2align 4, 0x90 .LBB9_1: # =>This Inner Loop Header: Depth=1 leaq (%rbx,%r12), %rdi callq hipStreamCreate addq $8, %r12 cmpq $8, %r12 je .LBB9_1 # %bb.2: leaq 24(%rsp), %rdi callq hipEventCreate leaq 16(%rsp), %rdi callq hipEventCreate xorps %xmm0, %xmm0 movl $16, %ebp movl $6, %r12d .p2align 4, 0x90 .LBB9_3: # =>This Inner Loop Header: Depth=1 movss %xmm0, 8(%rsp) # 4-byte Spill movq 24(%rsp), %rdi movq (%rbx), %rsi callq hipEventRecord movl %ebp, %r13d shll $20, %r13d movq (%rbx), %r8 movq %r14, %rdi movq %r15, %rsi movq %r13, %rdx movl $2, %ecx callq hipMemcpyAsync movq 16(%rsp), %rdi movq (%rbx), %rsi callq hipEventRecord leaq (%r15,%r13), %rdi leaq (%r14,%r13), %rsi movq 8(%rbx), %r8 movq %r13, %rdx movl $1, %ecx callq hipMemcpyAsync callq hipDeviceSynchronize movq 24(%rsp), %rsi movq 16(%rsp), %rdx leaq 12(%rsp), %rdi callq hipEventElapsedTime movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addss 12(%rsp), %xmm0 addl %ebp, %ebp decl %r12d jne .LBB9_3 # %bb.4: # %.preheader.preheader movss %xmm0, 8(%rsp) # 4-byte Spill xorl %r14d, %r14d .p2align 4, 0x90 .LBB9_5: # %.preheader # =>This Inner Loop Header: Depth=1 movq (%rbx,%r14,8), %rdi callq hipStreamDestroy incq %r14 cmpq $1, %r14 je .LBB9_5 # %bb.6: movss 36(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero mulss .LCPI9_0(%rip), %xmm0 movss 8(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero addss %xmm0, %xmm1 divss .LCPI9_1(%rip), %xmm1 movss %xmm1, 8(%rsp) # 4-byte Spill movq %rbx, %rdi callq free movq 24(%rsp), %rdi callq hipEventDestroy movq 16(%rsp), %rdi callq hipEventDestroy movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addq $40, %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_end9: .size _Z17getOverlappedGDTHPcS_fi, .Lfunc_end9-_Z17getOverlappedGDTHPcS_fi .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z17getOverlappedGHTDPcS_fi .LCPI10_0: .long 0xc0c00000 # float -6 .LCPI10_1: .long 0x4e7c0000 # float 1.05696461E+9 .text .globl _Z17getOverlappedGHTDPcS_fi .p2align 4, 0x90 .type _Z17getOverlappedGHTDPcS_fi,@function _Z17getOverlappedGHTDPcS_fi: # @_Z17getOverlappedGHTDPcS_fi .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 $40, %rsp .cfi_def_cfa_offset 96 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movss %xmm0, 36(%rsp) # 4-byte Spill movq %rsi, %r14 movq %rdi, %r15 movl $0, 12(%rsp) movl $16, %edi callq malloc movq %rax, %rbx xorl %r12d, %r12d .p2align 4, 0x90 .LBB10_1: # =>This Inner Loop Header: Depth=1 leaq (%rbx,%r12), %rdi callq hipStreamCreate addq $8, %r12 cmpq $8, %r12 je .LBB10_1 # %bb.2: leaq 24(%rsp), %rdi callq hipEventCreate leaq 16(%rsp), %rdi callq hipEventCreate xorps %xmm0, %xmm0 movl $16, %ebp movl $6, %r12d .p2align 4, 0x90 .LBB10_3: # =>This Inner Loop Header: Depth=1 movss %xmm0, 8(%rsp) # 4-byte Spill movq 24(%rsp), %rdi movq (%rbx), %rsi callq hipEventRecord movl %ebp, %r13d shll $20, %r13d movq (%rbx), %r8 movq %r15, %rdi movq %r14, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpyAsync movq 16(%rsp), %rdi movq (%rbx), %rsi callq hipEventRecord leaq (%r14,%r13), %rdi leaq (%r15,%r13), %rsi movq 8(%rbx), %r8 movq %r13, %rdx movl $2, %ecx callq hipMemcpyAsync callq hipDeviceSynchronize movq 24(%rsp), %rsi movq 16(%rsp), %rdx leaq 12(%rsp), %rdi callq hipEventElapsedTime movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addss 12(%rsp), %xmm0 addl %ebp, %ebp decl %r12d jne .LBB10_3 # %bb.4: # %.preheader.preheader movss %xmm0, 8(%rsp) # 4-byte Spill xorl %r14d, %r14d .p2align 4, 0x90 .LBB10_5: # %.preheader # =>This Inner Loop Header: Depth=1 movq (%rbx,%r14,8), %rdi callq hipStreamDestroy incq %r14 cmpq $1, %r14 je .LBB10_5 # %bb.6: movss 36(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero mulss .LCPI10_0(%rip), %xmm0 movss 8(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero addss %xmm0, %xmm1 divss .LCPI10_1(%rip), %xmm1 movss %xmm1, 8(%rsp) # 4-byte Spill movq %rbx, %rdi callq free movq 24(%rsp), %rdi callq hipEventDestroy movq 16(%rsp), %rdi callq hipEventDestroy movss 8(%rsp), %xmm0 # 4-byte Reload # xmm0 = mem[0],zero,zero,zero addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end10: .size _Z17getOverlappedGHTDPcS_fi, .Lfunc_end10-_Z17getOverlappedGHTDPcS_fi .cfi_endproc # -- End function .globl _Z20microbenchmarkingPCIiPfS_S_S_S_S_i # -- Begin function _Z20microbenchmarkingPCIiPfS_S_S_S_S_i .p2align 4, 0x90 .type _Z20microbenchmarkingPCIiPfS_S_S_S_S_i,@function _Z20microbenchmarkingPCIiPfS_S_S_S_S_i: # @_Z20microbenchmarkingPCIiPfS_S_S_S_S_i .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $1496, %rsp # imm = 0x5D8 .cfi_def_cfa_offset 1552 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 movq %r9, %r14 movq %r8, 16(%rsp) # 8-byte Spill movq %rcx, %r13 movq %rdx, %rbx movq %rsi, %r12 movl %edi, %ebp movl 1560(%rsp), %r15d leaq 24(%rsp), %rdi movl %ebp, %esi callq hipGetDevicePropertiesR0600 movl %ebp, %edi callq hipSetDevice leaq 8(%rsp), %rdi movl $1073741824, %esi # imm = 0x40000000 xorl %edx, %edx callq hipHostMalloc movq 8(%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 xorl %esi, %esi callq memset@PLT movq %rsp, %rdi movl $1073741824, %esi # imm = 0x40000000 callq hipMalloc movq (%rsp), %rdi movl $1073741824, %edx # imm = 0x40000000 xorl %esi, %esi callq hipMemset movq (%rsp), %rdi movq 8(%rsp), %rsi movl %r15d, %edx callq _Z8getLoHTDPcS_i movss %xmm0, (%r12) movq (%rsp), %rdi movq 8(%rsp), %rsi callq _Z7getGHTDPcS_fi movss %xmm0, (%r13) cmpl $2, 624(%rsp) jne .LBB11_2 # %bb.1: movq (%rsp), %rdi movq 8(%rsp), %rsi movss (%r12), %xmm0 # xmm0 = mem[0],zero,zero,zero callq _Z17getOverlappedGHTDPcS_fi movq 16(%rsp), %rax # 8-byte Reload movss %xmm0, (%rax) .LBB11_2: movq (%rsp), %rdi movq 8(%rsp), %rsi movl %r15d, %edx callq _Z8getLoDTHPcS_i movss %xmm0, (%rbx) movq (%rsp), %rdi movq 8(%rsp), %rsi callq _Z7getGDTHPcS_fi movss %xmm0, (%r14) cmpl $2, 624(%rsp) jne .LBB11_4 # %bb.3: movq 1552(%rsp), %r14 movq (%rsp), %rdi movq 8(%rsp), %rsi movss (%rbx), %xmm0 # xmm0 = mem[0],zero,zero,zero callq _Z17getOverlappedGDTHPcS_fi movss %xmm0, (%r14) .LBB11_4: movq 8(%rsp), %rdi callq hipHostFree movq (%rsp), %rdi callq hipFree addq $1496, %rsp # imm = 0x5D8 .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_end11: .size _Z20microbenchmarkingPCIiPfS_S_S_S_S_i, .Lfunc_end11-_Z20microbenchmarkingPCIiPfS_S_S_S_S_i .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 CUDA source code to CUDA device assembly.
#define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
.file "tmpxft_0010f6a9_00000000-6_Sum.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
#include <hip/hip_runtime.h> #define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #define TILE_DIM 1024 template<typename T> __device__ void sum(const T* matrix, T* result, const int rows, const int cols) { __shared__ T tile[TILE_DIM]; int index = threadIdx.x; int length = rows * cols; int partLength = (length + TILE_DIM - 1) / TILE_DIM; T sum = 0; for (int i = 0; i < partLength; i++) { int valueIndex = i * TILE_DIM + index; if (valueIndex < length) { T value = matrix[valueIndex]; sum += value; } } tile[index] = sum; for (int d = 1; d < TILE_DIM && d < length; d <<= 1) { __syncthreads(); if (index % (d << 1) == 0) { int valueIndex = index + d; if (valueIndex < TILE_DIM) { T value = tile[valueIndex]; sum += value; tile[index] = sum; } } } if (index == 0) { result[0] = sum; } }
.text .file "Sum.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80
.text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .amdgpu_metadata --- amdhsa.kernels: [] amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0010f6a9_00000000-6_Sum.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2030: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2030: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "Sum.hip" .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
code for sm_80 Function : _Z10SineKernelPf .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fc800078e0203 */ /*0060*/ IMAD.WIDE.U32 R2, R2, R5, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0005 */ /*0070*/ LDG.E R0, [R2.64] ; /* 0x0000000602007981 */ /* 0x000ea2000c1e1900 */ /*0080*/ BSSY B0, 0x7f0 ; /* 0x0000076000007945 */ /* 0x000fe20003800000 */ /*0090*/ FMUL R4, R0.reuse, 0.63661974668502807617 ; /* 0x3f22f98300047820 */ /* 0x044fe20000400000 */ /*00a0*/ FSETP.GE.AND P0, PT, |R0|, 105615, PT ; /* 0x47ce47800000780b */ /* 0x000fca0003f06200 */ /*00b0*/ F2I.NTZ R4, R4 ; /* 0x0000000400047305 */ /* 0x000e300000203100 */ /*00c0*/ I2F R5, R4 ; /* 0x0000000400057306 */ /* 0x001e240000201400 */ /*00d0*/ FFMA R6, R5, -1.5707962512969970703, R0 ; /* 0xbfc90fda05067823 */ /* 0x001fc80000000000 */ /*00e0*/ FFMA R6, R5, -7.5497894158615963534e-08, R6 ; /* 0xb3a2216805067823 */ /* 0x000fc80000000006 */ /*00f0*/ FFMA R5, R5, -5.3903029534742383927e-15, R6 ; /* 0xa7c234c505057823 */ /* 0x000fe20000000006 */ /*0100*/ @!P0 BRA 0x7e0 ; /* 0x000006d000008947 */ /* 0x000fea0003800000 */ /*0110*/ FSETP.NEU.AND P0, PT, |R0|, +INF , PT ; /* 0x7f8000000000780b */ /* 0x000fda0003f0d200 */ /*0120*/ @!P0 BRA 0x7c0 ; /* 0x0000069000008947 */ /* 0x000fea0003800000 */ /*0130*/ SHF.R.U32.HI R4, RZ, 0x17, R0 ; /* 0x00000017ff047819 */ /* 0x000fe20000011600 */ /*0140*/ IMAD.SHL.U32 R6, R0, 0x100, RZ ; /* 0x0000010000067824 */ /* 0x000fe200078e00ff */ /*0150*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0160*/ IMAD.MOV.U32 R13, RZ, RZ, RZ ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e00ff */ /*0170*/ LOP3.LUT R5, R4, 0xff, RZ, 0xc0, !PT ; /* 0x000000ff04057812 */ /* 0x000fe200078ec0ff */ /*0180*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe200078e00ff */ /*0190*/ LOP3.LUT R9, R6, 0x80000000, RZ, 0xfc, !PT ; /* 0x8000000006097812 */ /* 0x000fe200078efcff */ /*01a0*/ IMAD.MOV.U32 R19, RZ, RZ, RZ ; /* 0x000000ffff137224 */ /* 0x000fe200078e00ff */ /*01b0*/ IADD3 R5, R5, -0x80, RZ ; /* 0xffffff8005057810 */ /* 0x000fe20007ffe0ff */ /*01c0*/ ULDC.64 UR8, c[0x4][0x0] ; /* 0x0100000000087ab9 */ /* 0x000fc60000000a00 */ /*01d0*/ SHF.R.U32.HI R12, RZ, 0x5, R5 ; /* 0x00000005ff0c7819 */ /* 0x000fe40000011605 */ /*01e0*/ IMAD.U32 R10, RZ, RZ, UR8 ; /* 0x00000008ff0a7e24 */ /* 0x000fe4000f8e00ff */ /*01f0*/ IMAD.U32 R11, RZ, RZ, UR9 ; /* 0x00000009ff0b7e24 */ /* 0x000fca000f8e00ff */ /*0200*/ LDG.E.CONSTANT R6, [R10.64] ; /* 0x000000060a067981 */ /* 0x000ea2000c1e9900 */ /*0210*/ IADD3 R13, R13, 0x1, RZ ; /* 0x000000010d0d7810 */ /* 0x000fe20007ffe0ff */ /*0220*/ UIADD3 UR8, UP0, UR8, 0x4, URZ ; /* 0x0000000408087890 */ /* 0x000fe2000ff1e03f */ /*0230*/ ISETP.EQ.AND P0, PT, R19.reuse, RZ, PT ; /* 0x000000ff1300720c */ /* 0x040fe40003f02270 */ /*0240*/ ISETP.EQ.AND P5, PT, R19.reuse, 0x4, PT ; /* 0x000000041300780c */ /* 0x040fe20003fa2270 */ /*0250*/ UIADD3.X UR9, URZ, UR9, URZ, UP0, !UPT ; /* 0x000000093f097290 */ /* 0x000fe200087fe43f */ /*0260*/ ISETP.EQ.AND P4, PT, R19.reuse, 0x8, PT ; /* 0x000000081300780c */ /* 0x040fe40003f82270 */ /*0270*/ ISETP.EQ.AND P3, PT, R19.reuse, 0xc, PT ; /* 0x0000000c1300780c */ /* 0x040fe40003f62270 */ /*0280*/ ISETP.EQ.AND P2, PT, R19, 0x10, PT ; /* 0x000000101300780c */ /* 0x000fc40003f42270 */ /*0290*/ ISETP.EQ.AND P1, PT, R19.reuse, 0x14, PT ; /* 0x000000141300780c */ /* 0x040fe40003f22270 */ /*02a0*/ IADD3 R19, R19, 0x4, RZ ; /* 0x0000000413137810 */ /* 0x000fe20007ffe0ff */ /*02b0*/ IMAD.WIDE.U32 R6, R6, R9, RZ ; /* 0x0000000906067225 */ /* 0x004fca00078e00ff */ /*02c0*/ IADD3 R6, P6, R6, R4, RZ ; /* 0x0000000406067210 */ /* 0x000fc80007fde0ff */ /*02d0*/ IADD3.X R4, R7, UR4, RZ, P6, !PT ; /* 0x0000000407047c10 */ /* 0x000fe2000b7fe4ff */ /*02e0*/ @P0 IMAD.MOV.U32 R8, RZ, RZ, R6.reuse ; /* 0x000000ffff080224 */ /* 0x100fe200078e0006 */ /*02f0*/ ISETP.NE.AND P6, PT, R13, 0x6, PT ; /* 0x000000060d00780c */ /* 0x000fe20003fc5270 */ /*0300*/ @P4 IMAD.MOV.U32 R16, RZ, RZ, R6.reuse ; /* 0x000000ffff104224 */ /* 0x100fe200078e0006 */ /*0310*/ @P5 MOV R15, R6 ; /* 0x00000006000f5202 */ /* 0x000fe20000000f00 */ /*0320*/ @P3 IMAD.MOV.U32 R17, RZ, RZ, R6.reuse ; /* 0x000000ffff113224 */ /* 0x100fe400078e0006 */ /*0330*/ @P2 IMAD.MOV.U32 R18, RZ, RZ, R6.reuse ; /* 0x000000ffff122224 */ /* 0x100fe400078e0006 */ /*0340*/ @P1 IMAD.MOV.U32 R14, RZ, RZ, R6 ; /* 0x000000ffff0e1224 */ /* 0x000fcc00078e0006 */ /*0350*/ @P6 BRA 0x1e0 ; /* 0xfffffe8000006947 */ /* 0x000fea000383ffff */ /*0360*/ IADD3 R6, -R12, 0x6, RZ ; /* 0x000000060c067810 */ /* 0x000fe20007ffe1ff */ /*0370*/ BSSY B1, 0x6a0 ; /* 0x0000032000017945 */ /* 0x000fe80003800000 */ /*0380*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fca00078e00ff */ /*0390*/ ISETP.EQ.AND P0, PT, R6.reuse, RZ, PT ; /* 0x000000ff0600720c */ /* 0x040fe40003f02270 */ /*03a0*/ ISETP.EQ.AND P3, PT, R6.reuse, 0x4, PT ; /* 0x000000040600780c */ /* 0x040fe40003f62270 */ /*03b0*/ ISETP.EQ.AND P4, PT, R6.reuse, 0x8, PT ; /* 0x000000080600780c */ /* 0x040fe40003f82270 */ /*03c0*/ ISETP.EQ.AND P2, PT, R6.reuse, 0xc, PT ; /* 0x0000000c0600780c */ /* 0x040fe40003f42270 */ /*03d0*/ ISETP.EQ.AND P1, PT, R6, 0x10, PT ; /* 0x000000100600780c */ /* 0x000fca0003f22270 */ /*03e0*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R8.reuse ; /* 0x000000ffff070224 */ /* 0x100fe200078e0008 */ /*03f0*/ ISETP.EQ.AND P0, PT, R6.reuse, 0x14, PT ; /* 0x000000140600780c */ /* 0x040fe20003f02270 */ /*0400*/ @P3 IMAD.MOV.U32 R9, RZ, RZ, R8 ; /* 0x000000ffff093224 */ /* 0x000fe400078e0008 */ /*0410*/ @P3 IMAD.MOV.U32 R7, RZ, RZ, R15 ; /* 0x000000ffff073224 */ /* 0x000fe200078e000f */ /*0420*/ @P4 MOV R9, R15 ; /* 0x0000000f00094202 */ /* 0x000fe20000000f00 */ /*0430*/ @P4 IMAD.MOV.U32 R7, RZ, RZ, R16.reuse ; /* 0x000000ffff074224 */ /* 0x100fe200078e0010 */ /*0440*/ ISETP.EQ.AND P3, PT, R6.reuse, 0x18, PT ; /* 0x000000180600780c */ /* 0x040fe20003f62270 */ /*0450*/ @P2 IMAD.MOV.U32 R9, RZ, RZ, R16 ; /* 0x000000ffff092224 */ /* 0x000fe200078e0010 */ /*0460*/ ISETP.EQ.AND P4, PT, R6, 0x1c, PT ; /* 0x0000001c0600780c */ /* 0x000fe20003f82270 */ /*0470*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R17 ; /* 0x000000ffff072224 */ /* 0x000fc400078e0011 */ /*0480*/ @P1 IMAD.MOV.U32 R9, RZ, RZ, R17 ; /* 0x000000ffff091224 */ /* 0x000fe400078e0011 */ /*0490*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R18 ; /* 0x000000ffff071224 */ /* 0x000fe200078e0012 */ /*04a0*/ LOP3.LUT P1, R6, R5, 0x1f, RZ, 0xc0, !PT ; /* 0x0000001f05067812 */ /* 0x000fe2000782c0ff */ /*04b0*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R14 ; /* 0x000000ffff070224 */ /* 0x000fe400078e000e */ /*04c0*/ @P0 IMAD.MOV.U32 R9, RZ, RZ, R18 ; /* 0x000000ffff090224 */ /* 0x000fe400078e0012 */ /*04d0*/ @P3 IMAD.MOV.U32 R7, RZ, RZ, R4.reuse ; /* 0x000000ffff073224 */ /* 0x100fe200078e0004 */ /*04e0*/ @P3 MOV R9, R14 ; /* 0x0000000e00093202 */ /* 0x000fe20000000f00 */ /*04f0*/ @P4 IMAD.MOV.U32 R9, RZ, RZ, R4 ; /* 0x000000ffff094224 */ /* 0x000fc400078e0004 */ /*0500*/ IMAD.MOV.U32 R5, RZ, RZ, R7 ; /* 0x000000ffff057224 */ /* 0x000fc800078e0007 */ /*0510*/ @!P1 BRA 0x690 ; /* 0x0000017000009947 */ /* 0x000fea0003800000 */ /*0520*/ IADD3 R12, -R12, 0x4, RZ ; /* 0x000000040c0c7810 */ /* 0x000fca0007ffe1ff */ /*0530*/ IMAD.SHL.U32 R12, R12, 0x4, RZ ; /* 0x000000040c0c7824 */ /* 0x000fca00078e00ff */ /*0540*/ ISETP.EQ.AND P0, PT, R12.reuse, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x040fe40003f02270 */ /*0550*/ ISETP.EQ.AND P1, PT, R12.reuse, 0x4, PT ; /* 0x000000040c00780c */ /* 0x040fe40003f22270 */ /*0560*/ ISETP.EQ.AND P2, PT, R12.reuse, 0x8, PT ; /* 0x000000080c00780c */ /* 0x040fe40003f42270 */ /*0570*/ ISETP.EQ.AND P3, PT, R12, 0xc, PT ; /* 0x0000000c0c00780c */ /* 0x000fce0003f62270 */ /*0580*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R8 ; /* 0x000000ffff070224 */ /* 0x000fe200078e0008 */ /*0590*/ ISETP.EQ.AND P0, PT, R12.reuse, 0x10, PT ; /* 0x000000100c00780c */ /* 0x040fe20003f02270 */ /*05a0*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R15 ; /* 0x000000ffff071224 */ /* 0x000fe200078e000f */ /*05b0*/ ISETP.EQ.AND P1, PT, R12.reuse, 0x14, PT ; /* 0x000000140c00780c */ /* 0x040fe20003f22270 */ /*05c0*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R16 ; /* 0x000000ffff072224 */ /* 0x000fe200078e0010 */ /*05d0*/ ISETP.EQ.AND P2, PT, R12, 0x18, PT ; /* 0x000000180c00780c */ /* 0x000fe40003f42270 */ /*05e0*/ @P3 MOV R7, R17 ; /* 0x0000001100073202 */ /* 0x000fe40000000f00 */ /*05f0*/ IADD3 R8, -R6, 0x20, RZ ; /* 0x0000002006087810 */ /* 0x000fca0007ffe1ff */ /*0600*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R18 ; /* 0x000000ffff070224 */ /* 0x000fe400078e0012 */ /*0610*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R14 ; /* 0x000000ffff071224 */ /* 0x000fe400078e000e */ /*0620*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R4 ; /* 0x000000ffff072224 */ /* 0x000fe200078e0004 */ /*0630*/ SHF.L.U32 R4, R5, R6.reuse, RZ ; /* 0x0000000605047219 */ /* 0x080fe400000006ff */ /*0640*/ SHF.L.U32 R6, R9, R6, RZ ; /* 0x0000000609067219 */ /* 0x000fe400000006ff */ /*0650*/ SHF.R.U32.HI R5, RZ, R8.reuse, R9 ; /* 0x00000008ff057219 */ /* 0x080fe40000011609 */ /*0660*/ SHF.R.U32.HI R7, RZ, R8, R7 ; /* 0x00000008ff077219 */ /* 0x000fc60000011607 */ /*0670*/ IMAD.IADD R5, R5, 0x1, R4 ; /* 0x0000000105057824 */ /* 0x000fe400078e0204 */ /*0680*/ IMAD.IADD R9, R7, 0x1, R6 ; /* 0x0000000107097824 */ /* 0x000fe400078e0206 */ /*0690*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*06a0*/ SHF.L.U32.HI R4, R9.reuse, 0x2, R5 ; /* 0x0000000209047819 */ /* 0x040fe20000010605 */ /*06b0*/ IMAD.SHL.U32 R8, R9, 0x4, RZ ; /* 0x0000000409087824 */ /* 0x000fe200078e00ff */ /*06c0*/ LOP3.LUT P1, R0, R0, 0x80000000, RZ, 0xc0, !PT ; /* 0x8000000000007812 */ /* 0x000fe4000782c0ff */ /*06d0*/ SHF.R.U32.HI R10, RZ, 0x1f, R4 ; /* 0x0000001fff0a7819 */ /* 0x000fc80000011604 */ /*06e0*/ ISETP.NE.AND P0, PT, R10, RZ, PT ; /* 0x000000ff0a00720c */ /* 0x000fda0003f05270 */ /*06f0*/ @P0 LOP3.LUT R4, RZ, R4, RZ, 0x33, !PT ; /* 0x00000004ff040212 */ /* 0x000fe400078e33ff */ /*0700*/ @P0 LOP3.LUT R8, RZ, R8, RZ, 0x33, !PT ; /* 0x00000008ff080212 */ /* 0x000fe400078e33ff */ /*0710*/ @P0 LOP3.LUT R0, R0, 0x80000000, RZ, 0x3c, !PT ; /* 0x8000000000000812 */ /* 0x000fe200078e3cff */ /*0720*/ IMAD.MOV.U32 R9, RZ, RZ, R4 ; /* 0x000000ffff097224 */ /* 0x000fe200078e0004 */ /*0730*/ LEA.HI R4, R5, R10, RZ, 0x2 ; /* 0x0000000a05047211 */ /* 0x000fe400078f10ff */ /*0740*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe20003f05270 */ /*0750*/ I2F.F64.S64 R6, R8 ; /* 0x0000000800067312 */ /* 0x000e220000301c00 */ /*0760*/ IADD3 R0, -R4, RZ, RZ ; /* 0x000000ff04007210 */ /* 0x000fca0007ffe1ff */ /*0770*/ @P1 IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff041224 */ /* 0x000fe200078e0000 */ /*0780*/ DMUL R6, R6, c[0x2][0x0] ; /* 0x0080000006067a28 */ /* 0x001e140000000000 */ /*0790*/ F2F.F32.F64 R6, R6 ; /* 0x0000000600067310 */ /* 0x001e240000301000 */ /*07a0*/ FSEL R5, R6, -R6, !P0 ; /* 0x8000000606057208 */ /* 0x001fe20004000000 */ /*07b0*/ BRA 0x7e0 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*07c0*/ FMUL R5, RZ, R0 ; /* 0x00000000ff057220 */ /* 0x000fe40000400000 */ /*07d0*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe400078e00ff */ /*07e0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*07f0*/ LOP3.LUT P1, RZ, R4.reuse, 0x1, RZ, 0xc0, !PT ; /* 0x0000000104ff7812 */ /* 0x040fe2000782c0ff */ /*0800*/ IMAD.MOV.U32 R7, RZ, RZ, 0x3c0885e4 ; /* 0x3c0885e4ff077424 */ /* 0x000fe200078e00ff */ /*0810*/ LOP3.LUT P0, RZ, R4, 0x2, RZ, 0xc0, !PT ; /* 0x0000000204ff7812 */ /* 0x000fe2000780c0ff */ /*0820*/ FMUL R9, R5.reuse, R5 ; /* 0x0000000505097220 */ /* 0x040fe20000400000 */ /*0830*/ FSEL R4, R5, 1, !P1 ; /* 0x3f80000005047808 */ /* 0x000fe20004800000 */ /*0840*/ IMAD.MOV.U32 R8, RZ, RZ, 0x3e2aaaa8 ; /* 0x3e2aaaa8ff087424 */ /* 0x000fe200078e00ff */ /*0850*/ FSEL R7, R7, 0.041666727513074874878, !P1 ; /* 0x3d2aaabb07077808 */ /* 0x000fe20004800000 */ /*0860*/ IMAD.MOV.U32 R0, RZ, RZ, -0x46b2bead ; /* 0xb94d4153ff007424 */ /* 0x000fc600078e00ff */ /*0870*/ FSEL R5, -R8, -0.4999999701976776123, !P1 ; /* 0xbeffffff08057808 */ /* 0x000fc60004800100 */ /*0880*/ @P1 IMAD.MOV.U32 R6, RZ, RZ, 0x37cbac00 ; /* 0x37cbac00ff061424 */ /* 0x000fc800078e00ff */ /*0890*/ @P1 FFMA R0, R9, R6, -0.0013887860113754868507 ; /* 0xbab607ed09001423 */ /* 0x000fc80000000006 */ /*08a0*/ FFMA R0, R9.reuse, R0, R7 ; /* 0x0000000009007223 */ /* 0x040fe40000000007 */ /*08b0*/ FFMA R7, R4, R9, RZ ; /* 0x0000000904077223 */ /* 0x000fe400000000ff */ /*08c0*/ FFMA R0, R9, R0, R5 ; /* 0x0000000009007223 */ /* 0x000fc80000000005 */ /*08d0*/ FFMA R7, R0, R7, R4 ; /* 0x0000000700077223 */ /* 0x000fc80000000004 */ /*08e0*/ @P0 FFMA R7, R7, -1, RZ ; /* 0xbf80000007070823 */ /* 0x000fca00000000ff */ /*08f0*/ STG.E [R2.64], R7 ; /* 0x0000000702007986 */ /* 0x000fe2000c101906 */ /*0900*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0910*/ BRA 0x910; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0920*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0930*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0940*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0950*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0960*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0970*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0980*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0990*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
.file "tmpxft_000dde16_00000000-6_sine_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z30__device_stub__Z10SineKernelPfPf .type _Z30__device_stub__Z10SineKernelPfPf, @function _Z30__device_stub__Z10SineKernelPfPf: .LFB2081: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10SineKernelPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z30__device_stub__Z10SineKernelPfPf, .-_Z30__device_stub__Z10SineKernelPfPf .globl _Z10SineKernelPf .type _Z10SineKernelPf, @function _Z10SineKernelPf: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z10SineKernelPfPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size _Z10SineKernelPf, .-_Z10SineKernelPf .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10SineKernelPf" .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 _Z10SineKernelPf(%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> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10SineKernelPf .globl _Z10SineKernelPf .p2align 8 .type _Z10SineKernelPf,@function _Z10SineKernelPf: s_clause 0x1 s_load_b32 s2, s[0:1], 0x14 s_load_b64 s[0:1], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_mov_b32_e32 v2, 0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_mov_b32 s1, exec_lo global_load_b32 v2, v[0:1], off s_waitcnt vmcnt(0) v_and_b32_e32 v3, 0x7fffffff, v2 v_cmpx_ngt_f32_e64 0x48000000, |v2| s_xor_b32 s2, exec_lo, s1 s_cbranch_execz .LBB0_2 s_mov_b32 s0, 0x7fffff v_mov_b32_e32 v6, 0 v_and_or_b32 v14, v3, s0, 0x800000 v_lshrrev_b32_e32 v11, 23, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mad_u64_u32 v[4:5], null, v14, 0xfe5163ab, 0 v_add_nc_u32_e32 v12, 0xffffff88, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cmp_lt_u32_e32 vcc_lo, 63, v12 v_mad_u64_u32 v[7:8], null, v14, 0x3c439041, v[5:6] v_cndmask_b32_e64 v13, 0, 0xffffffc0, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mov_b32_e32 v5, v8 v_add_nc_u32_e32 v13, v13, v12 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mad_u64_u32 v[8:9], null, v14, 0xdb629599, v[5:6] v_cmp_lt_u32_e64 s0, 31, v13 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v15, 0, 0xffffffe0, s0 v_dual_mov_b32 v5, v9 :: v_dual_cndmask_b32 v4, v8, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v15, v15, v13 v_mad_u64_u32 v[9:10], null, v14, 0xf534ddc0, v[5:6] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cmp_lt_u32_e64 s1, 31, v15 v_mov_b32_e32 v5, v10 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v7, v9, v7, vcc_lo v_mad_u64_u32 v[10:11], null, v14, 0xfc2757d1, v[5:6] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e64 v4, v7, v4, s0 v_mov_b32_e32 v5, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[11:12], null, v14, 0x4e441529, v[5:6] v_mov_b32_e32 v5, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[12:13], null, v14, 0xa2f9836e, v[5:6] v_cndmask_b32_e64 v5, 0, 0xffffffe0, s1 v_dual_cndmask_b32 v6, v11, v9 :: v_dual_add_nc_u32 v5, v5, v15 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_dual_cndmask_b32 v12, v12, v10 :: v_dual_cndmask_b32 v11, v13, v11 v_cndmask_b32_e32 v10, v10, v8, vcc_lo v_cmp_eq_u32_e32 vcc_lo, 0, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4) v_cndmask_b32_e64 v9, v12, v6, s0 v_cndmask_b32_e64 v11, v11, v12, s0 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_2) | instid1(VALU_DEP_4) v_cndmask_b32_e64 v6, v6, v10, s0 v_sub_nc_u32_e32 v12, 32, v5 v_cndmask_b32_e64 v10, v10, v7, s0 v_cndmask_b32_e64 v11, v11, v9, s1 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v9, v9, v6, s1 v_cndmask_b32_e64 v6, v6, v10, s1 v_cndmask_b32_e64 v4, v10, v4, s1 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_alignbit_b32 v13, v11, v9, v12 v_alignbit_b32 v8, v9, v6, v12 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_4) v_cndmask_b32_e32 v5, v13, v11, vcc_lo v_alignbit_b32 v11, v6, v4, v12 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v7, v8, v9, vcc_lo v_bfe_u32 v8, v5, 29, 1 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v6, v11, v6, vcc_lo v_alignbit_b32 v9, v5, v7, 30 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_sub_nc_u32_e32 v10, 0, v8 v_alignbit_b32 v7, v7, v6, 30 v_alignbit_b32 v4, v6, v4, 30 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_xor_b32_e32 v9, v9, v10 v_xor_b32_e32 v6, v7, v10 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_xor_b32_e32 v4, v4, v10 v_clz_i32_u32_e32 v11, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_min_u32_e32 v11, 32, v11 v_sub_nc_u32_e32 v7, 31, v11 v_lshlrev_b32_e32 v13, 23, v11 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_alignbit_b32 v9, v9, v6, v7 v_alignbit_b32 v4, v6, v4, v7 v_lshrrev_b32_e32 v7, 29, v5 v_alignbit_b32 v6, v9, v4, 9 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_lshlrev_b32_e32 v7, 31, v7 v_lshrrev_b32_e32 v9, 9, v9 v_clz_i32_u32_e32 v10, v6 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_or_b32_e32 v12, 0.5, v7 v_min_u32_e32 v10, 32, v10 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v12, v12, v13 v_sub_nc_u32_e32 v14, 31, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_alignbit_b32 v4, v6, v4, v14 v_or_b32_e32 v6, v9, v12 v_add_lshl_u32 v9, v10, v11, 23 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_lshrrev_b32_e32 v4, 9, v4 v_mul_f32_e32 v10, 0x3fc90fda, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v4, v4, v9 v_fma_f32 v9, v6, 0x3fc90fda, -v10 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, 0x33000000, v4 v_fmamk_f32 v6, v6, 0x33a22168, v9 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_or_b32_e32 v4, v4, v7 v_fmac_f32_e32 v6, 0x3fc90fda, v4 v_lshrrev_b32_e32 v5, 30, v5 s_delay_alu instid0(VALU_DEP_1) v_dual_add_f32 v4, v10, v6 :: v_dual_add_nc_u32 v5, v8, v5 .LBB0_2: s_and_not1_saveexec_b32 s0, s2 v_mul_f32_e64 v4, 0x3f22f983, |v2| s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_rndne_f32_e32 v5, v4 v_fma_f32 v4, v5, 0xbfc90fda, |v2| s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmamk_f32 v4, v5, 0xb3a22168, v4 v_fmamk_f32 v4, v5, 0xa7c234c4, v4 v_cvt_i32_f32_e32 v5, v5 s_or_b32 exec_lo, exec_lo, s0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_dual_mul_f32 v6, v4, v4 :: v_dual_and_b32 v9, 1, v5 s_mov_b32 s0, 0xb94c1982 s_mov_b32 s1, 0x37d75334 v_xor_b32_e32 v3, v3, v2 v_fmaak_f32 v7, s0, v6, 0x3c0881c4 v_cmp_eq_u32_e32 vcc_lo, 0, v9 v_lshlrev_b32_e32 v5, 30, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_fmaak_f32 v7, v6, v7, 0xbe2aaa9d v_fmaak_f32 v8, s1, v6, 0xbab64f3b v_and_b32_e32 v5, 0x80000000, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_mul_f32_e32 v7, v6, v7 v_fmaak_f32 v8, v6, v8, 0x3d2aabf7 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_fmac_f32_e32 v4, v4, v7 v_fmaak_f32 v8, v6, v8, 0xbf000004 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, v6, v8, 1.0 v_cndmask_b32_e32 v4, v6, v4, vcc_lo v_cmp_class_f32_e64 vcc_lo, v2, 0x1f8 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_xor3_b32 v3, v3, v5, v4 v_cndmask_b32_e32 v2, 0x7fc00000, v3, 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 _Z10SineKernelPf .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 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10SineKernelPf, .Lfunc_end0-_Z10SineKernelPf .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: _Z10SineKernelPf .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10SineKernelPf.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void SineKernel(float *A) {{ // Block index const uint bx = blockIdx.x; const uint bw = blockDim.x; // Thread index const uint tx = threadIdx.x; // Stride access locations const uint aBegin = bw*bx; const uint idx = aBegin+tx; // Grab a value from global memory float this_val = A[idx]; float sinval = sin(this_val); A[idx] = sinval; // printf("Sin(%f) = %f\n", this_val, sinval); }}
.text .file "sine_kernel.hip" .globl _Z25__device_stub__SineKernelPf # -- Begin function _Z25__device_stub__SineKernelPf .p2align 4, 0x90 .type _Z25__device_stub__SineKernelPf,@function _Z25__device_stub__SineKernelPf: # @_Z25__device_stub__SineKernelPf .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 $_Z10SineKernelPf, %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 _Z25__device_stub__SineKernelPf, .Lfunc_end0-_Z25__device_stub__SineKernelPf .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 $_Z10SineKernelPf, %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 _Z10SineKernelPf,@object # @_Z10SineKernelPf .section .rodata,"a",@progbits .globl _Z10SineKernelPf .p2align 3, 0x0 _Z10SineKernelPf: .quad _Z25__device_stub__SineKernelPf .size _Z10SineKernelPf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z10SineKernelPf" .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 _Z25__device_stub__SineKernelPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10SineKernelPf .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 : _Z10SineKernelPf .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ IMAD.MOV.U32 R1, RZ, RZ, c[0x0][0x28] ; /* 0x00000a00ff017624 */ /* 0x000fe400078e00ff */ /*0010*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000e220000002500 */ /*0020*/ IMAD.MOV.U32 R5, RZ, RZ, 0x4 ; /* 0x00000004ff057424 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR6, c[0x0][0x118] ; /* 0x0000460000067ab9 */ /* 0x000fe40000000a00 */ /*0040*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0050*/ IMAD R2, R2, c[0x0][0x0], R3 ; /* 0x0000000002027a24 */ /* 0x001fc800078e0203 */ /*0060*/ IMAD.WIDE.U32 R2, R2, R5, c[0x0][0x160] ; /* 0x0000580002027625 */ /* 0x000fca00078e0005 */ /*0070*/ LDG.E R0, [R2.64] ; /* 0x0000000602007981 */ /* 0x000ea2000c1e1900 */ /*0080*/ BSSY B0, 0x7f0 ; /* 0x0000076000007945 */ /* 0x000fe20003800000 */ /*0090*/ FMUL R4, R0.reuse, 0.63661974668502807617 ; /* 0x3f22f98300047820 */ /* 0x044fe20000400000 */ /*00a0*/ FSETP.GE.AND P0, PT, |R0|, 105615, PT ; /* 0x47ce47800000780b */ /* 0x000fca0003f06200 */ /*00b0*/ F2I.NTZ R4, R4 ; /* 0x0000000400047305 */ /* 0x000e300000203100 */ /*00c0*/ I2F R5, R4 ; /* 0x0000000400057306 */ /* 0x001e240000201400 */ /*00d0*/ FFMA R6, R5, -1.5707962512969970703, R0 ; /* 0xbfc90fda05067823 */ /* 0x001fc80000000000 */ /*00e0*/ FFMA R6, R5, -7.5497894158615963534e-08, R6 ; /* 0xb3a2216805067823 */ /* 0x000fc80000000006 */ /*00f0*/ FFMA R5, R5, -5.3903029534742383927e-15, R6 ; /* 0xa7c234c505057823 */ /* 0x000fe20000000006 */ /*0100*/ @!P0 BRA 0x7e0 ; /* 0x000006d000008947 */ /* 0x000fea0003800000 */ /*0110*/ FSETP.NEU.AND P0, PT, |R0|, +INF , PT ; /* 0x7f8000000000780b */ /* 0x000fda0003f0d200 */ /*0120*/ @!P0 BRA 0x7c0 ; /* 0x0000069000008947 */ /* 0x000fea0003800000 */ /*0130*/ SHF.R.U32.HI R4, RZ, 0x17, R0 ; /* 0x00000017ff047819 */ /* 0x000fe20000011600 */ /*0140*/ IMAD.SHL.U32 R6, R0, 0x100, RZ ; /* 0x0000010000067824 */ /* 0x000fe200078e00ff */ /*0150*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*0160*/ IMAD.MOV.U32 R13, RZ, RZ, RZ ; /* 0x000000ffff0d7224 */ /* 0x000fe200078e00ff */ /*0170*/ LOP3.LUT R5, R4, 0xff, RZ, 0xc0, !PT ; /* 0x000000ff04057812 */ /* 0x000fe200078ec0ff */ /*0180*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe200078e00ff */ /*0190*/ LOP3.LUT R9, R6, 0x80000000, RZ, 0xfc, !PT ; /* 0x8000000006097812 */ /* 0x000fe200078efcff */ /*01a0*/ IMAD.MOV.U32 R19, RZ, RZ, RZ ; /* 0x000000ffff137224 */ /* 0x000fe200078e00ff */ /*01b0*/ IADD3 R5, R5, -0x80, RZ ; /* 0xffffff8005057810 */ /* 0x000fe20007ffe0ff */ /*01c0*/ ULDC.64 UR8, c[0x4][0x0] ; /* 0x0100000000087ab9 */ /* 0x000fc60000000a00 */ /*01d0*/ SHF.R.U32.HI R12, RZ, 0x5, R5 ; /* 0x00000005ff0c7819 */ /* 0x000fe40000011605 */ /*01e0*/ IMAD.U32 R10, RZ, RZ, UR8 ; /* 0x00000008ff0a7e24 */ /* 0x000fe4000f8e00ff */ /*01f0*/ IMAD.U32 R11, RZ, RZ, UR9 ; /* 0x00000009ff0b7e24 */ /* 0x000fca000f8e00ff */ /*0200*/ LDG.E.CONSTANT R6, [R10.64] ; /* 0x000000060a067981 */ /* 0x000ea2000c1e9900 */ /*0210*/ IADD3 R13, R13, 0x1, RZ ; /* 0x000000010d0d7810 */ /* 0x000fe20007ffe0ff */ /*0220*/ UIADD3 UR8, UP0, UR8, 0x4, URZ ; /* 0x0000000408087890 */ /* 0x000fe2000ff1e03f */ /*0230*/ ISETP.EQ.AND P0, PT, R19.reuse, RZ, PT ; /* 0x000000ff1300720c */ /* 0x040fe40003f02270 */ /*0240*/ ISETP.EQ.AND P5, PT, R19.reuse, 0x4, PT ; /* 0x000000041300780c */ /* 0x040fe20003fa2270 */ /*0250*/ UIADD3.X UR9, URZ, UR9, URZ, UP0, !UPT ; /* 0x000000093f097290 */ /* 0x000fe200087fe43f */ /*0260*/ ISETP.EQ.AND P4, PT, R19.reuse, 0x8, PT ; /* 0x000000081300780c */ /* 0x040fe40003f82270 */ /*0270*/ ISETP.EQ.AND P3, PT, R19.reuse, 0xc, PT ; /* 0x0000000c1300780c */ /* 0x040fe40003f62270 */ /*0280*/ ISETP.EQ.AND P2, PT, R19, 0x10, PT ; /* 0x000000101300780c */ /* 0x000fc40003f42270 */ /*0290*/ ISETP.EQ.AND P1, PT, R19.reuse, 0x14, PT ; /* 0x000000141300780c */ /* 0x040fe40003f22270 */ /*02a0*/ IADD3 R19, R19, 0x4, RZ ; /* 0x0000000413137810 */ /* 0x000fe20007ffe0ff */ /*02b0*/ IMAD.WIDE.U32 R6, R6, R9, RZ ; /* 0x0000000906067225 */ /* 0x004fca00078e00ff */ /*02c0*/ IADD3 R6, P6, R6, R4, RZ ; /* 0x0000000406067210 */ /* 0x000fc80007fde0ff */ /*02d0*/ IADD3.X R4, R7, UR4, RZ, P6, !PT ; /* 0x0000000407047c10 */ /* 0x000fe2000b7fe4ff */ /*02e0*/ @P0 IMAD.MOV.U32 R8, RZ, RZ, R6.reuse ; /* 0x000000ffff080224 */ /* 0x100fe200078e0006 */ /*02f0*/ ISETP.NE.AND P6, PT, R13, 0x6, PT ; /* 0x000000060d00780c */ /* 0x000fe20003fc5270 */ /*0300*/ @P4 IMAD.MOV.U32 R16, RZ, RZ, R6.reuse ; /* 0x000000ffff104224 */ /* 0x100fe200078e0006 */ /*0310*/ @P5 MOV R15, R6 ; /* 0x00000006000f5202 */ /* 0x000fe20000000f00 */ /*0320*/ @P3 IMAD.MOV.U32 R17, RZ, RZ, R6.reuse ; /* 0x000000ffff113224 */ /* 0x100fe400078e0006 */ /*0330*/ @P2 IMAD.MOV.U32 R18, RZ, RZ, R6.reuse ; /* 0x000000ffff122224 */ /* 0x100fe400078e0006 */ /*0340*/ @P1 IMAD.MOV.U32 R14, RZ, RZ, R6 ; /* 0x000000ffff0e1224 */ /* 0x000fcc00078e0006 */ /*0350*/ @P6 BRA 0x1e0 ; /* 0xfffffe8000006947 */ /* 0x000fea000383ffff */ /*0360*/ IADD3 R6, -R12, 0x6, RZ ; /* 0x000000060c067810 */ /* 0x000fe20007ffe1ff */ /*0370*/ BSSY B1, 0x6a0 ; /* 0x0000032000017945 */ /* 0x000fe80003800000 */ /*0380*/ IMAD.SHL.U32 R6, R6, 0x4, RZ ; /* 0x0000000406067824 */ /* 0x000fca00078e00ff */ /*0390*/ ISETP.EQ.AND P0, PT, R6.reuse, RZ, PT ; /* 0x000000ff0600720c */ /* 0x040fe40003f02270 */ /*03a0*/ ISETP.EQ.AND P3, PT, R6.reuse, 0x4, PT ; /* 0x000000040600780c */ /* 0x040fe40003f62270 */ /*03b0*/ ISETP.EQ.AND P4, PT, R6.reuse, 0x8, PT ; /* 0x000000080600780c */ /* 0x040fe40003f82270 */ /*03c0*/ ISETP.EQ.AND P2, PT, R6.reuse, 0xc, PT ; /* 0x0000000c0600780c */ /* 0x040fe40003f42270 */ /*03d0*/ ISETP.EQ.AND P1, PT, R6, 0x10, PT ; /* 0x000000100600780c */ /* 0x000fca0003f22270 */ /*03e0*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R8.reuse ; /* 0x000000ffff070224 */ /* 0x100fe200078e0008 */ /*03f0*/ ISETP.EQ.AND P0, PT, R6.reuse, 0x14, PT ; /* 0x000000140600780c */ /* 0x040fe20003f02270 */ /*0400*/ @P3 IMAD.MOV.U32 R9, RZ, RZ, R8 ; /* 0x000000ffff093224 */ /* 0x000fe400078e0008 */ /*0410*/ @P3 IMAD.MOV.U32 R7, RZ, RZ, R15 ; /* 0x000000ffff073224 */ /* 0x000fe200078e000f */ /*0420*/ @P4 MOV R9, R15 ; /* 0x0000000f00094202 */ /* 0x000fe20000000f00 */ /*0430*/ @P4 IMAD.MOV.U32 R7, RZ, RZ, R16.reuse ; /* 0x000000ffff074224 */ /* 0x100fe200078e0010 */ /*0440*/ ISETP.EQ.AND P3, PT, R6.reuse, 0x18, PT ; /* 0x000000180600780c */ /* 0x040fe20003f62270 */ /*0450*/ @P2 IMAD.MOV.U32 R9, RZ, RZ, R16 ; /* 0x000000ffff092224 */ /* 0x000fe200078e0010 */ /*0460*/ ISETP.EQ.AND P4, PT, R6, 0x1c, PT ; /* 0x0000001c0600780c */ /* 0x000fe20003f82270 */ /*0470*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R17 ; /* 0x000000ffff072224 */ /* 0x000fc400078e0011 */ /*0480*/ @P1 IMAD.MOV.U32 R9, RZ, RZ, R17 ; /* 0x000000ffff091224 */ /* 0x000fe400078e0011 */ /*0490*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R18 ; /* 0x000000ffff071224 */ /* 0x000fe200078e0012 */ /*04a0*/ LOP3.LUT P1, R6, R5, 0x1f, RZ, 0xc0, !PT ; /* 0x0000001f05067812 */ /* 0x000fe2000782c0ff */ /*04b0*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R14 ; /* 0x000000ffff070224 */ /* 0x000fe400078e000e */ /*04c0*/ @P0 IMAD.MOV.U32 R9, RZ, RZ, R18 ; /* 0x000000ffff090224 */ /* 0x000fe400078e0012 */ /*04d0*/ @P3 IMAD.MOV.U32 R7, RZ, RZ, R4.reuse ; /* 0x000000ffff073224 */ /* 0x100fe200078e0004 */ /*04e0*/ @P3 MOV R9, R14 ; /* 0x0000000e00093202 */ /* 0x000fe20000000f00 */ /*04f0*/ @P4 IMAD.MOV.U32 R9, RZ, RZ, R4 ; /* 0x000000ffff094224 */ /* 0x000fc400078e0004 */ /*0500*/ IMAD.MOV.U32 R5, RZ, RZ, R7 ; /* 0x000000ffff057224 */ /* 0x000fc800078e0007 */ /*0510*/ @!P1 BRA 0x690 ; /* 0x0000017000009947 */ /* 0x000fea0003800000 */ /*0520*/ IADD3 R12, -R12, 0x4, RZ ; /* 0x000000040c0c7810 */ /* 0x000fca0007ffe1ff */ /*0530*/ IMAD.SHL.U32 R12, R12, 0x4, RZ ; /* 0x000000040c0c7824 */ /* 0x000fca00078e00ff */ /*0540*/ ISETP.EQ.AND P0, PT, R12.reuse, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x040fe40003f02270 */ /*0550*/ ISETP.EQ.AND P1, PT, R12.reuse, 0x4, PT ; /* 0x000000040c00780c */ /* 0x040fe40003f22270 */ /*0560*/ ISETP.EQ.AND P2, PT, R12.reuse, 0x8, PT ; /* 0x000000080c00780c */ /* 0x040fe40003f42270 */ /*0570*/ ISETP.EQ.AND P3, PT, R12, 0xc, PT ; /* 0x0000000c0c00780c */ /* 0x000fce0003f62270 */ /*0580*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R8 ; /* 0x000000ffff070224 */ /* 0x000fe200078e0008 */ /*0590*/ ISETP.EQ.AND P0, PT, R12.reuse, 0x10, PT ; /* 0x000000100c00780c */ /* 0x040fe20003f02270 */ /*05a0*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R15 ; /* 0x000000ffff071224 */ /* 0x000fe200078e000f */ /*05b0*/ ISETP.EQ.AND P1, PT, R12.reuse, 0x14, PT ; /* 0x000000140c00780c */ /* 0x040fe20003f22270 */ /*05c0*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R16 ; /* 0x000000ffff072224 */ /* 0x000fe200078e0010 */ /*05d0*/ ISETP.EQ.AND P2, PT, R12, 0x18, PT ; /* 0x000000180c00780c */ /* 0x000fe40003f42270 */ /*05e0*/ @P3 MOV R7, R17 ; /* 0x0000001100073202 */ /* 0x000fe40000000f00 */ /*05f0*/ IADD3 R8, -R6, 0x20, RZ ; /* 0x0000002006087810 */ /* 0x000fca0007ffe1ff */ /*0600*/ @P0 IMAD.MOV.U32 R7, RZ, RZ, R18 ; /* 0x000000ffff070224 */ /* 0x000fe400078e0012 */ /*0610*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, R14 ; /* 0x000000ffff071224 */ /* 0x000fe400078e000e */ /*0620*/ @P2 IMAD.MOV.U32 R7, RZ, RZ, R4 ; /* 0x000000ffff072224 */ /* 0x000fe200078e0004 */ /*0630*/ SHF.L.U32 R4, R5, R6.reuse, RZ ; /* 0x0000000605047219 */ /* 0x080fe400000006ff */ /*0640*/ SHF.L.U32 R6, R9, R6, RZ ; /* 0x0000000609067219 */ /* 0x000fe400000006ff */ /*0650*/ SHF.R.U32.HI R5, RZ, R8.reuse, R9 ; /* 0x00000008ff057219 */ /* 0x080fe40000011609 */ /*0660*/ SHF.R.U32.HI R7, RZ, R8, R7 ; /* 0x00000008ff077219 */ /* 0x000fc60000011607 */ /*0670*/ IMAD.IADD R5, R5, 0x1, R4 ; /* 0x0000000105057824 */ /* 0x000fe400078e0204 */ /*0680*/ IMAD.IADD R9, R7, 0x1, R6 ; /* 0x0000000107097824 */ /* 0x000fe400078e0206 */ /*0690*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*06a0*/ SHF.L.U32.HI R4, R9.reuse, 0x2, R5 ; /* 0x0000000209047819 */ /* 0x040fe20000010605 */ /*06b0*/ IMAD.SHL.U32 R8, R9, 0x4, RZ ; /* 0x0000000409087824 */ /* 0x000fe200078e00ff */ /*06c0*/ LOP3.LUT P1, R0, R0, 0x80000000, RZ, 0xc0, !PT ; /* 0x8000000000007812 */ /* 0x000fe4000782c0ff */ /*06d0*/ SHF.R.U32.HI R10, RZ, 0x1f, R4 ; /* 0x0000001fff0a7819 */ /* 0x000fc80000011604 */ /*06e0*/ ISETP.NE.AND P0, PT, R10, RZ, PT ; /* 0x000000ff0a00720c */ /* 0x000fda0003f05270 */ /*06f0*/ @P0 LOP3.LUT R4, RZ, R4, RZ, 0x33, !PT ; /* 0x00000004ff040212 */ /* 0x000fe400078e33ff */ /*0700*/ @P0 LOP3.LUT R8, RZ, R8, RZ, 0x33, !PT ; /* 0x00000008ff080212 */ /* 0x000fe400078e33ff */ /*0710*/ @P0 LOP3.LUT R0, R0, 0x80000000, RZ, 0x3c, !PT ; /* 0x8000000000000812 */ /* 0x000fe200078e3cff */ /*0720*/ IMAD.MOV.U32 R9, RZ, RZ, R4 ; /* 0x000000ffff097224 */ /* 0x000fe200078e0004 */ /*0730*/ LEA.HI R4, R5, R10, RZ, 0x2 ; /* 0x0000000a05047211 */ /* 0x000fe400078f10ff */ /*0740*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe20003f05270 */ /*0750*/ I2F.F64.S64 R6, R8 ; /* 0x0000000800067312 */ /* 0x000e220000301c00 */ /*0760*/ IADD3 R0, -R4, RZ, RZ ; /* 0x000000ff04007210 */ /* 0x000fca0007ffe1ff */ /*0770*/ @P1 IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff041224 */ /* 0x000fe200078e0000 */ /*0780*/ DMUL R6, R6, c[0x2][0x0] ; /* 0x0080000006067a28 */ /* 0x001e140000000000 */ /*0790*/ F2F.F32.F64 R6, R6 ; /* 0x0000000600067310 */ /* 0x001e240000301000 */ /*07a0*/ FSEL R5, R6, -R6, !P0 ; /* 0x8000000606057208 */ /* 0x001fe20004000000 */ /*07b0*/ BRA 0x7e0 ; /* 0x0000002000007947 */ /* 0x000fea0003800000 */ /*07c0*/ FMUL R5, RZ, R0 ; /* 0x00000000ff057220 */ /* 0x000fe40000400000 */ /*07d0*/ IMAD.MOV.U32 R4, RZ, RZ, RZ ; /* 0x000000ffff047224 */ /* 0x000fe400078e00ff */ /*07e0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*07f0*/ LOP3.LUT P1, RZ, R4.reuse, 0x1, RZ, 0xc0, !PT ; /* 0x0000000104ff7812 */ /* 0x040fe2000782c0ff */ /*0800*/ IMAD.MOV.U32 R7, RZ, RZ, 0x3c0885e4 ; /* 0x3c0885e4ff077424 */ /* 0x000fe200078e00ff */ /*0810*/ LOP3.LUT P0, RZ, R4, 0x2, RZ, 0xc0, !PT ; /* 0x0000000204ff7812 */ /* 0x000fe2000780c0ff */ /*0820*/ FMUL R9, R5.reuse, R5 ; /* 0x0000000505097220 */ /* 0x040fe20000400000 */ /*0830*/ FSEL R4, R5, 1, !P1 ; /* 0x3f80000005047808 */ /* 0x000fe20004800000 */ /*0840*/ IMAD.MOV.U32 R8, RZ, RZ, 0x3e2aaaa8 ; /* 0x3e2aaaa8ff087424 */ /* 0x000fe200078e00ff */ /*0850*/ FSEL R7, R7, 0.041666727513074874878, !P1 ; /* 0x3d2aaabb07077808 */ /* 0x000fe20004800000 */ /*0860*/ IMAD.MOV.U32 R0, RZ, RZ, -0x46b2bead ; /* 0xb94d4153ff007424 */ /* 0x000fc600078e00ff */ /*0870*/ FSEL R5, -R8, -0.4999999701976776123, !P1 ; /* 0xbeffffff08057808 */ /* 0x000fc60004800100 */ /*0880*/ @P1 IMAD.MOV.U32 R6, RZ, RZ, 0x37cbac00 ; /* 0x37cbac00ff061424 */ /* 0x000fc800078e00ff */ /*0890*/ @P1 FFMA R0, R9, R6, -0.0013887860113754868507 ; /* 0xbab607ed09001423 */ /* 0x000fc80000000006 */ /*08a0*/ FFMA R0, R9.reuse, R0, R7 ; /* 0x0000000009007223 */ /* 0x040fe40000000007 */ /*08b0*/ FFMA R7, R4, R9, RZ ; /* 0x0000000904077223 */ /* 0x000fe400000000ff */ /*08c0*/ FFMA R0, R9, R0, R5 ; /* 0x0000000009007223 */ /* 0x000fc80000000005 */ /*08d0*/ FFMA R7, R0, R7, R4 ; /* 0x0000000700077223 */ /* 0x000fc80000000004 */ /*08e0*/ @P0 FFMA R7, R7, -1, RZ ; /* 0xbf80000007070823 */ /* 0x000fca00000000ff */ /*08f0*/ STG.E [R2.64], R7 ; /* 0x0000000702007986 */ /* 0x000fe2000c101906 */ /*0900*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0910*/ BRA 0x910; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0920*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0930*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0940*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0950*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0960*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0970*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0980*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0990*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*09f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z10SineKernelPf .globl _Z10SineKernelPf .p2align 8 .type _Z10SineKernelPf,@function _Z10SineKernelPf: s_clause 0x1 s_load_b32 s2, s[0:1], 0x14 s_load_b64 s[0:1], s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] v_mov_b32_e32 v2, 0 v_lshlrev_b64 v[0:1], 2, v[1:2] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_mov_b32 s1, exec_lo global_load_b32 v2, v[0:1], off s_waitcnt vmcnt(0) v_and_b32_e32 v3, 0x7fffffff, v2 v_cmpx_ngt_f32_e64 0x48000000, |v2| s_xor_b32 s2, exec_lo, s1 s_cbranch_execz .LBB0_2 s_mov_b32 s0, 0x7fffff v_mov_b32_e32 v6, 0 v_and_or_b32 v14, v3, s0, 0x800000 v_lshrrev_b32_e32 v11, 23, v3 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mad_u64_u32 v[4:5], null, v14, 0xfe5163ab, 0 v_add_nc_u32_e32 v12, 0xffffff88, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cmp_lt_u32_e32 vcc_lo, 63, v12 v_mad_u64_u32 v[7:8], null, v14, 0x3c439041, v[5:6] v_cndmask_b32_e64 v13, 0, 0xffffffc0, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mov_b32_e32 v5, v8 v_add_nc_u32_e32 v13, v13, v12 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_mad_u64_u32 v[8:9], null, v14, 0xdb629599, v[5:6] v_cmp_lt_u32_e64 s0, 31, v13 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v15, 0, 0xffffffe0, s0 v_dual_mov_b32 v5, v9 :: v_dual_cndmask_b32 v4, v8, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v15, v15, v13 v_mad_u64_u32 v[9:10], null, v14, 0xf534ddc0, v[5:6] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cmp_lt_u32_e64 s1, 31, v15 v_mov_b32_e32 v5, v10 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e32 v7, v9, v7, vcc_lo v_mad_u64_u32 v[10:11], null, v14, 0xfc2757d1, v[5:6] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cndmask_b32_e64 v4, v7, v4, s0 v_mov_b32_e32 v5, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[11:12], null, v14, 0x4e441529, v[5:6] v_mov_b32_e32 v5, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[12:13], null, v14, 0xa2f9836e, v[5:6] v_cndmask_b32_e64 v5, 0, 0xffffffe0, s1 v_dual_cndmask_b32 v6, v11, v9 :: v_dual_add_nc_u32 v5, v5, v15 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_dual_cndmask_b32 v12, v12, v10 :: v_dual_cndmask_b32 v11, v13, v11 v_cndmask_b32_e32 v10, v10, v8, vcc_lo v_cmp_eq_u32_e32 vcc_lo, 0, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4) v_cndmask_b32_e64 v9, v12, v6, s0 v_cndmask_b32_e64 v11, v11, v12, s0 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_2) | instid1(VALU_DEP_4) v_cndmask_b32_e64 v6, v6, v10, s0 v_sub_nc_u32_e32 v12, 32, v5 v_cndmask_b32_e64 v10, v10, v7, s0 v_cndmask_b32_e64 v11, v11, v9, s1 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v9, v9, v6, s1 v_cndmask_b32_e64 v6, v6, v10, s1 v_cndmask_b32_e64 v4, v10, v4, s1 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_alignbit_b32 v13, v11, v9, v12 v_alignbit_b32 v8, v9, v6, v12 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_4) v_cndmask_b32_e32 v5, v13, v11, vcc_lo v_alignbit_b32 v11, v6, v4, v12 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v7, v8, v9, vcc_lo v_bfe_u32 v8, v5, 29, 1 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v6, v11, v6, vcc_lo v_alignbit_b32 v9, v5, v7, 30 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_sub_nc_u32_e32 v10, 0, v8 v_alignbit_b32 v7, v7, v6, 30 v_alignbit_b32 v4, v6, v4, 30 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_xor_b32_e32 v9, v9, v10 v_xor_b32_e32 v6, v7, v10 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_xor_b32_e32 v4, v4, v10 v_clz_i32_u32_e32 v11, v9 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_min_u32_e32 v11, 32, v11 v_sub_nc_u32_e32 v7, 31, v11 v_lshlrev_b32_e32 v13, 23, v11 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_alignbit_b32 v9, v9, v6, v7 v_alignbit_b32 v4, v6, v4, v7 v_lshrrev_b32_e32 v7, 29, v5 v_alignbit_b32 v6, v9, v4, 9 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_lshlrev_b32_e32 v7, 31, v7 v_lshrrev_b32_e32 v9, 9, v9 v_clz_i32_u32_e32 v10, v6 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_or_b32_e32 v12, 0.5, v7 v_min_u32_e32 v10, 32, v10 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v12, v12, v13 v_sub_nc_u32_e32 v14, 31, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_alignbit_b32 v4, v6, v4, v14 v_or_b32_e32 v6, v9, v12 v_add_lshl_u32 v9, v10, v11, 23 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_lshrrev_b32_e32 v4, 9, v4 v_mul_f32_e32 v10, 0x3fc90fda, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v4, v4, v9 v_fma_f32 v9, v6, 0x3fc90fda, -v10 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, 0x33000000, v4 v_fmamk_f32 v6, v6, 0x33a22168, v9 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_or_b32_e32 v4, v4, v7 v_fmac_f32_e32 v6, 0x3fc90fda, v4 v_lshrrev_b32_e32 v5, 30, v5 s_delay_alu instid0(VALU_DEP_1) v_dual_add_f32 v4, v10, v6 :: v_dual_add_nc_u32 v5, v8, v5 .LBB0_2: s_and_not1_saveexec_b32 s0, s2 v_mul_f32_e64 v4, 0x3f22f983, |v2| s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_rndne_f32_e32 v5, v4 v_fma_f32 v4, v5, 0xbfc90fda, |v2| s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmamk_f32 v4, v5, 0xb3a22168, v4 v_fmamk_f32 v4, v5, 0xa7c234c4, v4 v_cvt_i32_f32_e32 v5, v5 s_or_b32 exec_lo, exec_lo, s0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_dual_mul_f32 v6, v4, v4 :: v_dual_and_b32 v9, 1, v5 s_mov_b32 s0, 0xb94c1982 s_mov_b32 s1, 0x37d75334 v_xor_b32_e32 v3, v3, v2 v_fmaak_f32 v7, s0, v6, 0x3c0881c4 v_cmp_eq_u32_e32 vcc_lo, 0, v9 v_lshlrev_b32_e32 v5, 30, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_fmaak_f32 v7, v6, v7, 0xbe2aaa9d v_fmaak_f32 v8, s1, v6, 0xbab64f3b v_and_b32_e32 v5, 0x80000000, v5 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3) v_mul_f32_e32 v7, v6, v7 v_fmaak_f32 v8, v6, v8, 0x3d2aabf7 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_fmac_f32_e32 v4, v4, v7 v_fmaak_f32 v8, v6, v8, 0xbf000004 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v6, v6, v8, 1.0 v_cndmask_b32_e32 v4, v6, v4, vcc_lo v_cmp_class_f32_e64 vcc_lo, v2, 0x1f8 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_xor3_b32 v3, v3, v5, v4 v_cndmask_b32_e32 v2, 0x7fc00000, v3, 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 _Z10SineKernelPf .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 16 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z10SineKernelPf, .Lfunc_end0-_Z10SineKernelPf .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: _Z10SineKernelPf .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z10SineKernelPf.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 16 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000dde16_00000000-6_sine_kernel.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2059: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2059: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z30__device_stub__Z10SineKernelPfPf .type _Z30__device_stub__Z10SineKernelPfPf, @function _Z30__device_stub__Z10SineKernelPfPf: .LFB2081: .cfi_startproc endbr64 subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 8(%rsp) movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 88(%rsp), %rax subq %fs:40, %rax jne .L8 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 120 pushq 24(%rsp) .cfi_def_cfa_offset 128 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z10SineKernelPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2081: .size _Z30__device_stub__Z10SineKernelPfPf, .-_Z30__device_stub__Z10SineKernelPfPf .globl _Z10SineKernelPf .type _Z10SineKernelPf, @function _Z10SineKernelPf: .LFB2082: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z30__device_stub__Z10SineKernelPfPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2082: .size _Z10SineKernelPf, .-_Z10SineKernelPf .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z10SineKernelPf" .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 _Z10SineKernelPf(%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 "sine_kernel.hip" .globl _Z25__device_stub__SineKernelPf # -- Begin function _Z25__device_stub__SineKernelPf .p2align 4, 0x90 .type _Z25__device_stub__SineKernelPf,@function _Z25__device_stub__SineKernelPf: # @_Z25__device_stub__SineKernelPf .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 $_Z10SineKernelPf, %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 _Z25__device_stub__SineKernelPf, .Lfunc_end0-_Z25__device_stub__SineKernelPf .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 $_Z10SineKernelPf, %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 _Z10SineKernelPf,@object # @_Z10SineKernelPf .section .rodata,"a",@progbits .globl _Z10SineKernelPf .p2align 3, 0x0 _Z10SineKernelPf: .quad _Z25__device_stub__SineKernelPf .size _Z10SineKernelPf, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z10SineKernelPf" .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 _Z25__device_stub__SineKernelPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z10SineKernelPf .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/device_vector.h> #include <thrust/host_vector.h> #include <iostream> #include <chrono> int main() { int N = 2518; thrust::host_vector<double> host(N); double tmp; for(int i = 0; i < N; i++){ std::cin >> tmp; host[i] = tmp; } /* na linha abaixo os dados são copiados para GPU */ auto start_time = std::chrono::high_resolution_clock::now(); thrust::device_vector<double> dev(host); auto end_time = std::chrono::high_resolution_clock::now(); auto runtime = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time); std::cerr << "Alocação e cópia para GPU: " << runtime.count() << "ms\n"; double mediaAnual = thrust::reduce(dev.end() -365, dev.end(), 0.0, thrust::plus<double>()) /N; std::cout << "Media Anual: " << mediaAnual << "\n"; }
#include <hip/hip_runtime.h> #include <thrust/device_vector.h> #include <thrust/host_vector.h> #include <iostream> #include <chrono> int main() { int N = 2518; thrust::host_vector<double> host(N); double tmp; for(int i = 0; i < N; i++){ std::cin >> tmp; host[i] = tmp; } /* na linha abaixo os dados são copiados para GPU */ auto start_time = std::chrono::high_resolution_clock::now(); thrust::device_vector<double> dev(host); auto end_time = std::chrono::high_resolution_clock::now(); auto runtime = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time); std::cerr << "Alocação e cópia para GPU: " << runtime.count() << "ms\n"; double mediaAnual = thrust::reduce(dev.end() -365, dev.end(), 0.0, thrust::plus<double>()) /N; std::cout << "Media Anual: " << mediaAnual << "\n"; }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
code for sm_80 Function : _Z28mergePartialHistogramsKernelPjS_jj .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 R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0020*/ ISETP.GE.U32.AND P0, PT, R3, c[0x0][0x174], PT ; /* 0x00005d0003007a0c */ /* 0x001fda0003f06070 */ /*0030*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0040*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e220000002100 */ /*0050*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0060*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x001fe20003f06070 */ /*0070*/ BSSY B0, 0x4a0 ; /* 0x0000042000007945 */ /* 0x000fe20003800000 */ /*0080*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fd600078e00ff */ /*0090*/ @P0 BRA 0x490 ; /* 0x000003f000000947 */ /* 0x000fea0003800000 */ /*00a0*/ ISETP.GE.U32.AND P1, PT, R0.reuse, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x040fe20003f26070 */ /*00b0*/ IMAD R4, R0, c[0x0][0x174], R3 ; /* 0x00005d0000047a24 */ /* 0x000fd800078e0203 */ /*00c0*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff071424 */ /* 0x000fc800078e00ff */ /*00d0*/ @P1 IMAD.WIDE.U32 R6, R4, R7, c[0x0][0x168] ; /* 0x00005a0004061625 */ /* 0x000fcc00078e0007 */ /*00e0*/ @P1 LDG.E R6, [R6.64] ; /* 0x0000000406061981 */ /* 0x000ea2000c1e1900 */ /*00f0*/ MOV R2, R0 ; /* 0x0000000000027202 */ /* 0x000fe20000000f00 */ /*0100*/ @P1 IMAD.MOV.U32 R9, RZ, RZ, 0x80 ; /* 0x00000080ff091424 */ /* 0x000fe200078e00ff */ /*0110*/ BSSY B1, 0x320 ; /* 0x0000020000017945 */ /* 0x000fe40003800000 */ /*0120*/ @P1 IADD3 R2, R2, 0x80, RZ ; /* 0x0000008002021810 */ /* 0x000fe20007ffe0ff */ /*0130*/ @P1 IMAD R4, R9, c[0x0][0x174], R4 ; /* 0x00005d0009041a24 */ /* 0x000fc600078e0204 */ /*0140*/ ISETP.GE.U32.AND P0, PT, R2.reuse, c[0x0][0x170], PT ; /* 0x00005c0002007a0c */ /* 0x040fe40003f06070 */ /*0150*/ IADD3 R5, -R2, c[0x0][0x170], RZ ; /* 0x00005c0002057a10 */ /* 0x000fc80007ffe1ff */ /*0160*/ ISETP.LE.U32.OR P0, PT, R5, 0x180, P0 ; /* 0x000001800500780c */ /* 0x000fe20000703470 */ /*0170*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc800078e00ff */ /*0180*/ @P1 IMAD.IADD R5, R5, 0x1, R6 ; /* 0x0000000105051824 */ /* 0x004fe200078e0206 */ /*0190*/ PLOP3.LUT P1, PT, P1, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fce0000f2e170 */ /*01a0*/ @P0 BRA 0x310 ; /* 0x0000016000000947 */ /* 0x000fea0003800000 */ /*01b0*/ MOV R15, c[0x0][0x170] ; /* 0x00005c00000f7a02 */ /* 0x000fe40000000f00 */ /*01c0*/ PLOP3.LUT P1, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f2e170 */ /*01d0*/ IADD3 R15, R15, -0x180, RZ ; /* 0xfffffe800f0f7810 */ /* 0x000fe40007ffe0ff */ /*01e0*/ IMAD.MOV.U32 R17, RZ, RZ, 0x80 ; /* 0x00000080ff117424 */ /* 0x000fe400078e00ff */ /*01f0*/ IMAD.MOV.U32 R13, RZ, RZ, 0x4 ; /* 0x00000004ff0d7424 */ /* 0x000fe400078e00ff */ /*0200*/ IMAD R8, R17, c[0x0][0x174], R4 ; /* 0x00005d0011087a24 */ /* 0x000fe400078e0204 */ /*0210*/ IMAD.WIDE.U32 R6, R4, R13, c[0x0][0x168] ; /* 0x00005a0004067625 */ /* 0x000fc800078e000d */ /*0220*/ IMAD R10, R17.reuse, c[0x0][0x174], R8 ; /* 0x00005d00110a7a24 */ /* 0x040fe400078e0208 */ /*0230*/ IMAD.WIDE.U32 R8, R8, R13.reuse, c[0x0][0x168] ; /* 0x00005a0008087625 */ /* 0x080fe200078e000d */ /*0240*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea6000c1e1900 */ /*0250*/ IMAD R4, R17, c[0x0][0x174], R10 ; /* 0x00005d0011047a24 */ /* 0x000fe400078e020a */ /*0260*/ IMAD.WIDE.U32 R10, R10, R13.reuse, c[0x0][0x168] ; /* 0x00005a000a0a7625 */ /* 0x080fe200078e000d */ /*0270*/ LDG.E R8, [R8.64] ; /* 0x0000000408087981 */ /* 0x000ea6000c1e1900 */ /*0280*/ IMAD.WIDE.U32 R12, R4, R13, c[0x0][0x168] ; /* 0x00005a00040c7625 */ /* 0x000fc400078e000d */ /*0290*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ee8000c1e1900 */ /*02a0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000ee2000c1e1900 */ /*02b0*/ IADD3 R2, R2, 0x200, RZ ; /* 0x0000020002027810 */ /* 0x000fc80007ffe0ff */ /*02c0*/ ISETP.GE.U32.AND P0, PT, R2, R15, PT ; /* 0x0000000f0200720c */ /* 0x000fe20003f06070 */ /*02d0*/ IMAD R4, R17, c[0x0][0x174], R4 ; /* 0x00005d0011047a24 */ /* 0x000fe200078e0204 */ /*02e0*/ IADD3 R5, R8, R6, R5 ; /* 0x0000000608057210 */ /* 0x004fc80007ffe005 */ /*02f0*/ IADD3 R5, R12, R10, R5 ; /* 0x0000000a0c057210 */ /* 0x008fce0007ffe005 */ /*0300*/ @!P0 BRA 0x1e0 ; /* 0xfffffed000008947 */ /* 0x000fea000383ffff */ /*0310*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0320*/ ISETP.GE.U32.AND P0, PT, R2.reuse, c[0x0][0x170], PT ; /* 0x00005c0002007a0c */ /* 0x040fe20003f06070 */ /*0330*/ BSSY B1, 0x430 ; /* 0x000000f000017945 */ /* 0x000fe20003800000 */ /*0340*/ IADD3 R6, -R2, c[0x0][0x170], RZ ; /* 0x00005c0002067a10 */ /* 0x000fc80007ffe1ff */ /*0350*/ ISETP.LE.U32.OR P0, PT, R6, 0x80, P0 ; /* 0x000000800600780c */ /* 0x000fda0000703470 */ /*0360*/ @P0 BRA 0x420 ; /* 0x000000b000000947 */ /* 0x000fea0003800000 */ /*0370*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0380*/ IMAD.MOV.U32 R11, RZ, RZ, 0x80 ; /* 0x00000080ff0b7424 */ /* 0x000fc800078e00ff */ /*0390*/ IMAD R10, R11, c[0x0][0x174], R4 ; /* 0x00005d000b0a7a24 */ /* 0x000fca00078e0204 */ /*03a0*/ IMAD.WIDE.U32 R8, R4, R7, c[0x0][0x168] ; /* 0x00005a0004087625 */ /* 0x000fc800078e0007 */ /*03b0*/ IMAD.WIDE.U32 R6, R10, R7, c[0x0][0x168] ; /* 0x00005a000a067625 */ /* 0x000fe400078e0007 */ /*03c0*/ LDG.E R8, [R8.64] ; /* 0x0000000408087981 */ /* 0x000ea8000c1e1900 */ /*03d0*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea2000c1e1900 */ /*03e0*/ PLOP3.LUT P1, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe20003f2e170 */ /*03f0*/ IMAD R4, R11, c[0x0][0x174], R10 ; /* 0x00005d000b047a24 */ /* 0x000fe200078e020a */ /*0400*/ IADD3 R2, R2, 0x100, RZ ; /* 0x0000010002027810 */ /* 0x000fe40007ffe0ff */ /*0410*/ IADD3 R5, R6, R8, R5 ; /* 0x0000000806057210 */ /* 0x004fc80007ffe005 */ /*0420*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0430*/ ISETP.LT.U32.OR P1, PT, R2, c[0x0][0x170], P1 ; /* 0x00005c0002007a0c */ /* 0x000fda0000f21470 */ /*0440*/ @!P1 BRA 0x490 ; /* 0x0000004000009947 */ /* 0x000fea0003800000 */ /*0450*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fc800078e00ff */ /*0460*/ IMAD.WIDE.U32 R6, R4, R7, c[0x0][0x168] ; /* 0x00005a0004067625 */ /* 0x000fcc00078e0007 */ /*0470*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea4000c1e1900 */ /*0480*/ IMAD.IADD R5, R5, 0x1, R6 ; /* 0x0000000105057824 */ /* 0x004fe400078e0206 */ /*0490*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*04a0*/ STS [R0.X4], R5 ; /* 0x0000000500007388 */ /* 0x000fe80000004800 */ /*04b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*04c0*/ ISETP.GT.U32.AND P1, PT, R0.reuse, 0x3f, PT ; /* 0x0000003f0000780c */ /* 0x040fe40003f24070 */ /*04d0*/ ISETP.GT.U32.AND P0, PT, R0, 0x1f, PT ; /* 0x0000001f0000780c */ /* 0x000fd60003f04070 */ /*04e0*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*04f0*/ @!P1 LDS R7, [R0.X4+0x100] ; /* 0x0001000000079984 */ /* 0x000e240000004800 */ /*0500*/ @!P1 IADD3 R7, R2, R7, RZ ; /* 0x0000000702079210 */ /* 0x001fca0007ffe0ff */ /*0510*/ @!P1 STS [R0.X4], R7 ; /* 0x0000000700009388 */ /* 0x000fe80000004800 */ /*0520*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0530*/ ISETP.GT.U32.AND P1, PT, R0, 0xf, PT ; /* 0x0000000f0000780c */ /* 0x000fca0003f24070 */ /*0540*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*0550*/ @!P0 LDS R9, [R0.X4+0x80] ; /* 0x0000800000098984 */ /* 0x000e240000004800 */ /*0560*/ @!P0 IMAD.IADD R9, R2, 0x1, R9 ; /* 0x0000000102098824 */ /* 0x001fca00078e0209 */ /*0570*/ @!P0 STS [R0.X4], R9 ; /* 0x0000000900008388 */ /* 0x000fe80000004800 */ /*0580*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0590*/ ISETP.GT.U32.AND P0, PT, R0, 0x7, PT ; /* 0x000000070000780c */ /* 0x000fca0003f04070 */ /*05a0*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*05b0*/ @!P1 LDS R5, [R0.X4+0x40] ; /* 0x0000400000059984 */ /* 0x000e240000004800 */ /*05c0*/ @!P1 IMAD.IADD R5, R2, 0x1, R5 ; /* 0x0000000102059824 */ /* 0x001fca00078e0205 */ /*05d0*/ @!P1 STS [R0.X4], R5 ; /* 0x0000000500009388 */ /* 0x000fe80000004800 */ /*05e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*05f0*/ ISETP.GT.U32.AND P1, PT, R0, 0x3, PT ; /* 0x000000030000780c */ /* 0x000fca0003f24070 */ /*0600*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*0610*/ @!P0 LDS R7, [R0.X4+0x20] ; /* 0x0000200000078984 */ /* 0x000e240000004800 */ /*0620*/ @!P0 IADD3 R7, R2, R7, RZ ; /* 0x0000000702078210 */ /* 0x001fca0007ffe0ff */ /*0630*/ @!P0 STS [R0.X4], R7 ; /* 0x0000000700008388 */ /* 0x000fe80000004800 */ /*0640*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0650*/ ISETP.GT.U32.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fca0003f04070 */ /*0660*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*0670*/ @!P1 LDS R9, [R0.X4+0x10] ; /* 0x0000100000099984 */ /* 0x000e240000004800 */ /*0680*/ @!P1 IMAD.IADD R9, R2, 0x1, R9 ; /* 0x0000000102099824 */ /* 0x001fca00078e0209 */ /*0690*/ @!P1 STS [R0.X4], R9 ; /* 0x0000000900009388 */ /* 0x000fe80000004800 */ /*06a0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*06b0*/ ISETP.NE.AND P1, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f25270 */ /*06c0*/ @!P1 IMAD.MOV.U32 R4, RZ, RZ, 0x4 ; /* 0x00000004ff049424 */ /* 0x000fe200078e00ff */ /*06d0*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*06e0*/ @!P0 LDS R5, [R0.X4+0x8] ; /* 0x0000080000058984 */ /* 0x000e240000004800 */ /*06f0*/ @!P0 IMAD.IADD R7, R2, 0x1, R5 ; /* 0x0000000102078824 */ /* 0x001fca00078e0205 */ /*0700*/ @!P0 STS [R0.X4], R7 ; /* 0x0000000700008388 */ /* 0x000fe80000004800 */ /*0710*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0720*/ @!P1 LDS R2, [0x4] ; /* 0x00000400ff029984 */ /* 0x000fe80000000800 */ /*0730*/ @!P1 LDS R5, [R0.X4] ; /* 0x0000000000059984 */ /* 0x000e240000004800 */ /*0740*/ @!P1 IADD3 R11, R2, R5, RZ ; /* 0x00000005020b9210 */ /* 0x001fe20007ffe0ff */ /*0750*/ @!P1 IMAD.WIDE.U32 R4, R3.reuse, R4, c[0x0][0x160] ; /* 0x0000580003049625 */ /* 0x040fe200078e0004 */ /*0760*/ IADD3 R3, R3, c[0x0][0xc], RZ ; /* 0x0000030003037a10 */ /* 0x000fc60007ffe0ff */ /*0770*/ @!P1 STS [R0.X4], R11 ; /* 0x0000000b00009388 */ /* 0x000fe20000004800 */ /*0780*/ ISETP.GE.U32.AND P0, PT, R3, c[0x0][0x174], PT ; /* 0x00005d0003007a0c */ /* 0x000fc60003f06070 */ /*0790*/ @!P1 LDS R9, [RZ] ; /* 0x00000000ff099984 */ /* 0x000e280000000800 */ /*07a0*/ @!P1 STG.E [R4.64], R9 ; /* 0x0000000904009986 */ /* 0x0011ec000c101904 */ /*07b0*/ @!P0 BRA 0x60 ; /* 0xfffff8a000008947 */ /* 0x000fea000383ffff */ /*07c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*07d0*/ BRA 0x7d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*07e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0800*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0810*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0820*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0830*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0840*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0850*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0860*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0870*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
.file "tmpxft_00096cea_00000000-6_mergePartialHistogramsKernel.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 _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj .type _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj, @function _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) 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 _Z28mergePartialHistogramsKernelPjS_jj(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj, .-_Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj .globl _Z28mergePartialHistogramsKernelPjS_jj .type _Z28mergePartialHistogramsKernelPjS_jj, @function _Z28mergePartialHistogramsKernelPjS_jj: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z28mergePartialHistogramsKernelPjS_jj, .-_Z28mergePartialHistogramsKernelPjS_jj .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z28mergePartialHistogramsKernelPjS_jj" .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 _Z28mergePartialHistogramsKernelPjS_jj(%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" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
#include <hip/hip_runtime.h> #include "includes.h" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z28mergePartialHistogramsKernelPjS_jj .globl _Z28mergePartialHistogramsKernelPjS_jj .p2align 8 .type _Z28mergePartialHistogramsKernelPjS_jj,@function _Z28mergePartialHistogramsKernelPjS_jj: s_load_b32 s8, s[0:1], 0x14 s_waitcnt lgkmcnt(0) s_cmp_ge_u32 s15, s8 s_cbranch_scc1 .LBB0_13 s_clause 0x2 s_load_b32 s9, s[0:1], 0x10 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b32 s10, s[0:1], 0x18 s_mov_b32 s2, s15 s_mov_b32 s3, 0 v_dual_mov_b32 v3, 0 :: v_dual_lshlrev_b32 v4, 2, v0 v_mad_u64_u32 v[1:2], null, v0, s8, s[2:3] v_cmp_eq_u32_e32 vcc_lo, 0, v0 s_lshl_b32 s11, s8, 7 s_waitcnt lgkmcnt(0) v_cmp_gt_u32_e64 s0, s9, v0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s1 v_add_nc_u32_e32 v1, s10, v1 s_add_i32 s2, s10, s2 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_ge_u32 s2, s8 s_cbranch_scc1 .LBB0_13 .LBB0_3: v_mov_b32_e32 v5, v3 s_delay_alu instid0(VALU_DEP_2) s_and_saveexec_b32 s12, s0 s_cbranch_execz .LBB0_7 v_dual_mov_b32 v5, 0 :: v_dual_mov_b32 v2, v1 v_mov_b32_e32 v6, v0 s_mov_b32 s13, 0 .p2align 6 .LBB0_5: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_lshlrev_b64 v[7:8], 2, v[2:3] v_add_nc_u32_e32 v6, 0x80, v6 v_add_nc_u32_e32 v2, s11, v2 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v7, s1, s6, v7 v_add_co_ci_u32_e64 v8, s1, s7, v8, s1 s_delay_alu instid0(VALU_DEP_4) v_cmp_le_u32_e64 s1, s9, v6 global_load_b32 v7, v[7:8], off s_or_b32 s13, s1, s13 s_waitcnt vmcnt(0) v_add_nc_u32_e32 v5, v7, v5 s_and_not1_b32 exec_lo, exec_lo, s13 s_cbranch_execnz .LBB0_5 s_or_b32 exec_lo, exec_lo, s13 .LBB0_7: s_delay_alu instid0(SALU_CYCLE_1) s_or_b32 exec_lo, exec_lo, s12 s_mov_b32 s12, 64 ds_store_b32 v4, v5 s_branch .LBB0_9 .p2align 6 .LBB0_8: s_or_b32 exec_lo, exec_lo, s13 s_lshr_b32 s1, s12, 1 s_cmp_lt_u32 s12, 2 s_mov_b32 s12, s1 s_cbranch_scc1 .LBB0_11 .LBB0_9: s_mov_b32 s13, exec_lo s_waitcnt lgkmcnt(0) s_waitcnt_vscnt null, 0x0 s_barrier buffer_gl0_inv v_cmpx_gt_u32_e64 s12, v0 s_cbranch_execz .LBB0_8 v_add_lshl_u32 v2, s12, v0, 2 ds_load_b32 v2, v2 ds_load_b32 v5, v4 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v2, v5, v2 ds_store_b32 v4, v2 s_branch .LBB0_8 .LBB0_11: s_and_saveexec_b32 s1, vcc_lo s_cbranch_execz .LBB0_2 ds_load_b32 v2, v3 s_lshl_b64 s[12:13], s[2:3], 2 s_delay_alu instid0(SALU_CYCLE_1) s_add_u32 s12, s4, s12 s_addc_u32 s13, s5, s13 s_waitcnt lgkmcnt(0) global_store_b32 v3, v2, s[12:13] s_branch .LBB0_2 .LBB0_13: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z28mergePartialHistogramsKernelPjS_jj .amdhsa_group_segment_fixed_size 512 .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 9 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z28mergePartialHistogramsKernelPjS_jj, .Lfunc_end0-_Z28mergePartialHistogramsKernelPjS_jj .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: 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: 512 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z28mergePartialHistogramsKernelPjS_jj .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z28mergePartialHistogramsKernelPjS_jj.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 9 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" #define UMUL(a, b) ( (a) * (b) ) #define UMAD(a, b, c) ( UMUL((a), (b)) + (c) ) typedef unsigned int uint; typedef unsigned short ushort; typedef unsigned char uchar; #define SHARED_MEMORY_SIZE 49152 #define MERGE_THREADBLOCK_SIZE 128 static uint *d_PartialHistograms; /* * Function that maps value to bin in range 0 inclusive to binCOunt exclusive */ __global__ void mergePartialHistogramsKernel(uint *d_Histogram, uint *d_PartialHistograms, uint histogramCount, uint binCount) { for (uint bin = blockIdx.x; bin < binCount; bin += gridDim.x) { uint sum = 0; for (uint histogramIndex = threadIdx.x; histogramIndex < histogramCount; histogramIndex += MERGE_THREADBLOCK_SIZE) { sum += d_PartialHistograms[bin + histogramIndex * binCount]; } __shared__ uint data[MERGE_THREADBLOCK_SIZE]; data[threadIdx.x] = sum; for (uint stride = MERGE_THREADBLOCK_SIZE / 2; stride > 0; stride >>= 1) { __syncthreads(); if (threadIdx.x < stride) { data[threadIdx.x] += data[threadIdx.x + stride]; } } if (threadIdx.x == 0) { d_Histogram[bin] = data[0]; } } }
.text .file "mergePartialHistogramsKernel.hip" .globl _Z43__device_stub__mergePartialHistogramsKernelPjS_jj # -- Begin function _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .p2align 4, 0x90 .type _Z43__device_stub__mergePartialHistogramsKernelPjS_jj,@function _Z43__device_stub__mergePartialHistogramsKernelPjS_jj: # @_Z43__device_stub__mergePartialHistogramsKernelPjS_jj .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) 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 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 $_Z28mergePartialHistogramsKernelPjS_jj, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z43__device_stub__mergePartialHistogramsKernelPjS_jj, .Lfunc_end0-_Z43__device_stub__mergePartialHistogramsKernelPjS_jj .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 $_Z28mergePartialHistogramsKernelPjS_jj, %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 _Z28mergePartialHistogramsKernelPjS_jj,@object # @_Z28mergePartialHistogramsKernelPjS_jj .section .rodata,"a",@progbits .globl _Z28mergePartialHistogramsKernelPjS_jj .p2align 3, 0x0 _Z28mergePartialHistogramsKernelPjS_jj: .quad _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .size _Z28mergePartialHistogramsKernelPjS_jj, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z28mergePartialHistogramsKernelPjS_jj" .size .L__unnamed_1, 39 .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 _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z28mergePartialHistogramsKernelPjS_jj .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 : _Z28mergePartialHistogramsKernelPjS_jj .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 R3, SR_CTAID.X ; /* 0x0000000000037919 */ /* 0x000e240000002500 */ /*0020*/ ISETP.GE.U32.AND P0, PT, R3, c[0x0][0x174], PT ; /* 0x00005d0003007a0c */ /* 0x001fda0003f06070 */ /*0030*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0040*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000e220000002100 */ /*0050*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc60000000a00 */ /*0060*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x001fe20003f06070 */ /*0070*/ BSSY B0, 0x4a0 ; /* 0x0000042000007945 */ /* 0x000fe20003800000 */ /*0080*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fd600078e00ff */ /*0090*/ @P0 BRA 0x490 ; /* 0x000003f000000947 */ /* 0x000fea0003800000 */ /*00a0*/ ISETP.GE.U32.AND P1, PT, R0.reuse, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x040fe20003f26070 */ /*00b0*/ IMAD R4, R0, c[0x0][0x174], R3 ; /* 0x00005d0000047a24 */ /* 0x000fd800078e0203 */ /*00c0*/ @P1 IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff071424 */ /* 0x000fc800078e00ff */ /*00d0*/ @P1 IMAD.WIDE.U32 R6, R4, R7, c[0x0][0x168] ; /* 0x00005a0004061625 */ /* 0x000fcc00078e0007 */ /*00e0*/ @P1 LDG.E R6, [R6.64] ; /* 0x0000000406061981 */ /* 0x000ea2000c1e1900 */ /*00f0*/ MOV R2, R0 ; /* 0x0000000000027202 */ /* 0x000fe20000000f00 */ /*0100*/ @P1 IMAD.MOV.U32 R9, RZ, RZ, 0x80 ; /* 0x00000080ff091424 */ /* 0x000fe200078e00ff */ /*0110*/ BSSY B1, 0x320 ; /* 0x0000020000017945 */ /* 0x000fe40003800000 */ /*0120*/ @P1 IADD3 R2, R2, 0x80, RZ ; /* 0x0000008002021810 */ /* 0x000fe20007ffe0ff */ /*0130*/ @P1 IMAD R4, R9, c[0x0][0x174], R4 ; /* 0x00005d0009041a24 */ /* 0x000fc600078e0204 */ /*0140*/ ISETP.GE.U32.AND P0, PT, R2.reuse, c[0x0][0x170], PT ; /* 0x00005c0002007a0c */ /* 0x040fe40003f06070 */ /*0150*/ IADD3 R5, -R2, c[0x0][0x170], RZ ; /* 0x00005c0002057a10 */ /* 0x000fc80007ffe1ff */ /*0160*/ ISETP.LE.U32.OR P0, PT, R5, 0x180, P0 ; /* 0x000001800500780c */ /* 0x000fe20000703470 */ /*0170*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc800078e00ff */ /*0180*/ @P1 IMAD.IADD R5, R5, 0x1, R6 ; /* 0x0000000105051824 */ /* 0x004fe200078e0206 */ /*0190*/ PLOP3.LUT P1, PT, P1, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fce0000f2e170 */ /*01a0*/ @P0 BRA 0x310 ; /* 0x0000016000000947 */ /* 0x000fea0003800000 */ /*01b0*/ MOV R15, c[0x0][0x170] ; /* 0x00005c00000f7a02 */ /* 0x000fe40000000f00 */ /*01c0*/ PLOP3.LUT P1, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe40003f2e170 */ /*01d0*/ IADD3 R15, R15, -0x180, RZ ; /* 0xfffffe800f0f7810 */ /* 0x000fe40007ffe0ff */ /*01e0*/ IMAD.MOV.U32 R17, RZ, RZ, 0x80 ; /* 0x00000080ff117424 */ /* 0x000fe400078e00ff */ /*01f0*/ IMAD.MOV.U32 R13, RZ, RZ, 0x4 ; /* 0x00000004ff0d7424 */ /* 0x000fe400078e00ff */ /*0200*/ IMAD R8, R17, c[0x0][0x174], R4 ; /* 0x00005d0011087a24 */ /* 0x000fe400078e0204 */ /*0210*/ IMAD.WIDE.U32 R6, R4, R13, c[0x0][0x168] ; /* 0x00005a0004067625 */ /* 0x000fc800078e000d */ /*0220*/ IMAD R10, R17.reuse, c[0x0][0x174], R8 ; /* 0x00005d00110a7a24 */ /* 0x040fe400078e0208 */ /*0230*/ IMAD.WIDE.U32 R8, R8, R13.reuse, c[0x0][0x168] ; /* 0x00005a0008087625 */ /* 0x080fe200078e000d */ /*0240*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea6000c1e1900 */ /*0250*/ IMAD R4, R17, c[0x0][0x174], R10 ; /* 0x00005d0011047a24 */ /* 0x000fe400078e020a */ /*0260*/ IMAD.WIDE.U32 R10, R10, R13.reuse, c[0x0][0x168] ; /* 0x00005a000a0a7625 */ /* 0x080fe200078e000d */ /*0270*/ LDG.E R8, [R8.64] ; /* 0x0000000408087981 */ /* 0x000ea6000c1e1900 */ /*0280*/ IMAD.WIDE.U32 R12, R4, R13, c[0x0][0x168] ; /* 0x00005a00040c7625 */ /* 0x000fc400078e000d */ /*0290*/ LDG.E R10, [R10.64] ; /* 0x000000040a0a7981 */ /* 0x000ee8000c1e1900 */ /*02a0*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x000ee2000c1e1900 */ /*02b0*/ IADD3 R2, R2, 0x200, RZ ; /* 0x0000020002027810 */ /* 0x000fc80007ffe0ff */ /*02c0*/ ISETP.GE.U32.AND P0, PT, R2, R15, PT ; /* 0x0000000f0200720c */ /* 0x000fe20003f06070 */ /*02d0*/ IMAD R4, R17, c[0x0][0x174], R4 ; /* 0x00005d0011047a24 */ /* 0x000fe200078e0204 */ /*02e0*/ IADD3 R5, R8, R6, R5 ; /* 0x0000000608057210 */ /* 0x004fc80007ffe005 */ /*02f0*/ IADD3 R5, R12, R10, R5 ; /* 0x0000000a0c057210 */ /* 0x008fce0007ffe005 */ /*0300*/ @!P0 BRA 0x1e0 ; /* 0xfffffed000008947 */ /* 0x000fea000383ffff */ /*0310*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0320*/ ISETP.GE.U32.AND P0, PT, R2.reuse, c[0x0][0x170], PT ; /* 0x00005c0002007a0c */ /* 0x040fe20003f06070 */ /*0330*/ BSSY B1, 0x430 ; /* 0x000000f000017945 */ /* 0x000fe20003800000 */ /*0340*/ IADD3 R6, -R2, c[0x0][0x170], RZ ; /* 0x00005c0002067a10 */ /* 0x000fc80007ffe1ff */ /*0350*/ ISETP.LE.U32.OR P0, PT, R6, 0x80, P0 ; /* 0x000000800600780c */ /* 0x000fda0000703470 */ /*0360*/ @P0 BRA 0x420 ; /* 0x000000b000000947 */ /* 0x000fea0003800000 */ /*0370*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0380*/ IMAD.MOV.U32 R11, RZ, RZ, 0x80 ; /* 0x00000080ff0b7424 */ /* 0x000fc800078e00ff */ /*0390*/ IMAD R10, R11, c[0x0][0x174], R4 ; /* 0x00005d000b0a7a24 */ /* 0x000fca00078e0204 */ /*03a0*/ IMAD.WIDE.U32 R8, R4, R7, c[0x0][0x168] ; /* 0x00005a0004087625 */ /* 0x000fc800078e0007 */ /*03b0*/ IMAD.WIDE.U32 R6, R10, R7, c[0x0][0x168] ; /* 0x00005a000a067625 */ /* 0x000fe400078e0007 */ /*03c0*/ LDG.E R8, [R8.64] ; /* 0x0000000408087981 */ /* 0x000ea8000c1e1900 */ /*03d0*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea2000c1e1900 */ /*03e0*/ PLOP3.LUT P1, PT, PT, PT, PT, 0x8, 0x0 ; /* 0x000000000000781c */ /* 0x000fe20003f2e170 */ /*03f0*/ IMAD R4, R11, c[0x0][0x174], R10 ; /* 0x00005d000b047a24 */ /* 0x000fe200078e020a */ /*0400*/ IADD3 R2, R2, 0x100, RZ ; /* 0x0000010002027810 */ /* 0x000fe40007ffe0ff */ /*0410*/ IADD3 R5, R6, R8, R5 ; /* 0x0000000806057210 */ /* 0x004fc80007ffe005 */ /*0420*/ BSYNC B1 ; /* 0x0000000000017941 */ /* 0x000fea0003800000 */ /*0430*/ ISETP.LT.U32.OR P1, PT, R2, c[0x0][0x170], P1 ; /* 0x00005c0002007a0c */ /* 0x000fda0000f21470 */ /*0440*/ @!P1 BRA 0x490 ; /* 0x0000004000009947 */ /* 0x000fea0003800000 */ /*0450*/ IMAD.MOV.U32 R7, RZ, RZ, 0x4 ; /* 0x00000004ff077424 */ /* 0x000fc800078e00ff */ /*0460*/ IMAD.WIDE.U32 R6, R4, R7, c[0x0][0x168] ; /* 0x00005a0004067625 */ /* 0x000fcc00078e0007 */ /*0470*/ LDG.E R6, [R6.64] ; /* 0x0000000406067981 */ /* 0x000ea4000c1e1900 */ /*0480*/ IMAD.IADD R5, R5, 0x1, R6 ; /* 0x0000000105057824 */ /* 0x004fe400078e0206 */ /*0490*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*04a0*/ STS [R0.X4], R5 ; /* 0x0000000500007388 */ /* 0x000fe80000004800 */ /*04b0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*04c0*/ ISETP.GT.U32.AND P1, PT, R0.reuse, 0x3f, PT ; /* 0x0000003f0000780c */ /* 0x040fe40003f24070 */ /*04d0*/ ISETP.GT.U32.AND P0, PT, R0, 0x1f, PT ; /* 0x0000001f0000780c */ /* 0x000fd60003f04070 */ /*04e0*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*04f0*/ @!P1 LDS R7, [R0.X4+0x100] ; /* 0x0001000000079984 */ /* 0x000e240000004800 */ /*0500*/ @!P1 IADD3 R7, R2, R7, RZ ; /* 0x0000000702079210 */ /* 0x001fca0007ffe0ff */ /*0510*/ @!P1 STS [R0.X4], R7 ; /* 0x0000000700009388 */ /* 0x000fe80000004800 */ /*0520*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0530*/ ISETP.GT.U32.AND P1, PT, R0, 0xf, PT ; /* 0x0000000f0000780c */ /* 0x000fca0003f24070 */ /*0540*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*0550*/ @!P0 LDS R9, [R0.X4+0x80] ; /* 0x0000800000098984 */ /* 0x000e240000004800 */ /*0560*/ @!P0 IMAD.IADD R9, R2, 0x1, R9 ; /* 0x0000000102098824 */ /* 0x001fca00078e0209 */ /*0570*/ @!P0 STS [R0.X4], R9 ; /* 0x0000000900008388 */ /* 0x000fe80000004800 */ /*0580*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0590*/ ISETP.GT.U32.AND P0, PT, R0, 0x7, PT ; /* 0x000000070000780c */ /* 0x000fca0003f04070 */ /*05a0*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*05b0*/ @!P1 LDS R5, [R0.X4+0x40] ; /* 0x0000400000059984 */ /* 0x000e240000004800 */ /*05c0*/ @!P1 IMAD.IADD R5, R2, 0x1, R5 ; /* 0x0000000102059824 */ /* 0x001fca00078e0205 */ /*05d0*/ @!P1 STS [R0.X4], R5 ; /* 0x0000000500009388 */ /* 0x000fe80000004800 */ /*05e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*05f0*/ ISETP.GT.U32.AND P1, PT, R0, 0x3, PT ; /* 0x000000030000780c */ /* 0x000fca0003f24070 */ /*0600*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*0610*/ @!P0 LDS R7, [R0.X4+0x20] ; /* 0x0000200000078984 */ /* 0x000e240000004800 */ /*0620*/ @!P0 IADD3 R7, R2, R7, RZ ; /* 0x0000000702078210 */ /* 0x001fca0007ffe0ff */ /*0630*/ @!P0 STS [R0.X4], R7 ; /* 0x0000000700008388 */ /* 0x000fe80000004800 */ /*0640*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*0650*/ ISETP.GT.U32.AND P0, PT, R0, 0x1, PT ; /* 0x000000010000780c */ /* 0x000fca0003f04070 */ /*0660*/ @!P1 LDS R2, [R0.X4] ; /* 0x0000000000029984 */ /* 0x000fe80000004800 */ /*0670*/ @!P1 LDS R9, [R0.X4+0x10] ; /* 0x0000100000099984 */ /* 0x000e240000004800 */ /*0680*/ @!P1 IMAD.IADD R9, R2, 0x1, R9 ; /* 0x0000000102099824 */ /* 0x001fca00078e0209 */ /*0690*/ @!P1 STS [R0.X4], R9 ; /* 0x0000000900009388 */ /* 0x000fe80000004800 */ /*06a0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe20000010000 */ /*06b0*/ ISETP.NE.AND P1, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fda0003f25270 */ /*06c0*/ @!P1 IMAD.MOV.U32 R4, RZ, RZ, 0x4 ; /* 0x00000004ff049424 */ /* 0x000fe200078e00ff */ /*06d0*/ @!P0 LDS R2, [R0.X4] ; /* 0x0000000000028984 */ /* 0x000fe80000004800 */ /*06e0*/ @!P0 LDS R5, [R0.X4+0x8] ; /* 0x0000080000058984 */ /* 0x000e240000004800 */ /*06f0*/ @!P0 IMAD.IADD R7, R2, 0x1, R5 ; /* 0x0000000102078824 */ /* 0x001fca00078e0205 */ /*0700*/ @!P0 STS [R0.X4], R7 ; /* 0x0000000700008388 */ /* 0x000fe80000004800 */ /*0710*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0720*/ @!P1 LDS R2, [0x4] ; /* 0x00000400ff029984 */ /* 0x000fe80000000800 */ /*0730*/ @!P1 LDS R5, [R0.X4] ; /* 0x0000000000059984 */ /* 0x000e240000004800 */ /*0740*/ @!P1 IADD3 R11, R2, R5, RZ ; /* 0x00000005020b9210 */ /* 0x001fe20007ffe0ff */ /*0750*/ @!P1 IMAD.WIDE.U32 R4, R3.reuse, R4, c[0x0][0x160] ; /* 0x0000580003049625 */ /* 0x040fe200078e0004 */ /*0760*/ IADD3 R3, R3, c[0x0][0xc], RZ ; /* 0x0000030003037a10 */ /* 0x000fc60007ffe0ff */ /*0770*/ @!P1 STS [R0.X4], R11 ; /* 0x0000000b00009388 */ /* 0x000fe20000004800 */ /*0780*/ ISETP.GE.U32.AND P0, PT, R3, c[0x0][0x174], PT ; /* 0x00005d0003007a0c */ /* 0x000fc60003f06070 */ /*0790*/ @!P1 LDS R9, [RZ] ; /* 0x00000000ff099984 */ /* 0x000e280000000800 */ /*07a0*/ @!P1 STG.E [R4.64], R9 ; /* 0x0000000904009986 */ /* 0x0011ec000c101904 */ /*07b0*/ @!P0 BRA 0x60 ; /* 0xfffff8a000008947 */ /* 0x000fea000383ffff */ /*07c0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*07d0*/ BRA 0x7d0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*07e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*07f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0800*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0810*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0820*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0830*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0840*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0850*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0860*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0870*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z28mergePartialHistogramsKernelPjS_jj .globl _Z28mergePartialHistogramsKernelPjS_jj .p2align 8 .type _Z28mergePartialHistogramsKernelPjS_jj,@function _Z28mergePartialHistogramsKernelPjS_jj: s_load_b32 s8, s[0:1], 0x14 s_waitcnt lgkmcnt(0) s_cmp_ge_u32 s15, s8 s_cbranch_scc1 .LBB0_13 s_clause 0x2 s_load_b32 s9, s[0:1], 0x10 s_load_b128 s[4:7], s[0:1], 0x0 s_load_b32 s10, s[0:1], 0x18 s_mov_b32 s2, s15 s_mov_b32 s3, 0 v_dual_mov_b32 v3, 0 :: v_dual_lshlrev_b32 v4, 2, v0 v_mad_u64_u32 v[1:2], null, v0, s8, s[2:3] v_cmp_eq_u32_e32 vcc_lo, 0, v0 s_lshl_b32 s11, s8, 7 s_waitcnt lgkmcnt(0) v_cmp_gt_u32_e64 s0, s9, v0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s1 v_add_nc_u32_e32 v1, s10, v1 s_add_i32 s2, s10, s2 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_ge_u32 s2, s8 s_cbranch_scc1 .LBB0_13 .LBB0_3: v_mov_b32_e32 v5, v3 s_delay_alu instid0(VALU_DEP_2) s_and_saveexec_b32 s12, s0 s_cbranch_execz .LBB0_7 v_dual_mov_b32 v5, 0 :: v_dual_mov_b32 v2, v1 v_mov_b32_e32 v6, v0 s_mov_b32 s13, 0 .p2align 6 .LBB0_5: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_lshlrev_b64 v[7:8], 2, v[2:3] v_add_nc_u32_e32 v6, 0x80, v6 v_add_nc_u32_e32 v2, s11, v2 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v7, s1, s6, v7 v_add_co_ci_u32_e64 v8, s1, s7, v8, s1 s_delay_alu instid0(VALU_DEP_4) v_cmp_le_u32_e64 s1, s9, v6 global_load_b32 v7, v[7:8], off s_or_b32 s13, s1, s13 s_waitcnt vmcnt(0) v_add_nc_u32_e32 v5, v7, v5 s_and_not1_b32 exec_lo, exec_lo, s13 s_cbranch_execnz .LBB0_5 s_or_b32 exec_lo, exec_lo, s13 .LBB0_7: s_delay_alu instid0(SALU_CYCLE_1) s_or_b32 exec_lo, exec_lo, s12 s_mov_b32 s12, 64 ds_store_b32 v4, v5 s_branch .LBB0_9 .p2align 6 .LBB0_8: s_or_b32 exec_lo, exec_lo, s13 s_lshr_b32 s1, s12, 1 s_cmp_lt_u32 s12, 2 s_mov_b32 s12, s1 s_cbranch_scc1 .LBB0_11 .LBB0_9: s_mov_b32 s13, exec_lo s_waitcnt lgkmcnt(0) s_waitcnt_vscnt null, 0x0 s_barrier buffer_gl0_inv v_cmpx_gt_u32_e64 s12, v0 s_cbranch_execz .LBB0_8 v_add_lshl_u32 v2, s12, v0, 2 ds_load_b32 v2, v2 ds_load_b32 v5, v4 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v2, v5, v2 ds_store_b32 v4, v2 s_branch .LBB0_8 .LBB0_11: s_and_saveexec_b32 s1, vcc_lo s_cbranch_execz .LBB0_2 ds_load_b32 v2, v3 s_lshl_b64 s[12:13], s[2:3], 2 s_delay_alu instid0(SALU_CYCLE_1) s_add_u32 s12, s4, s12 s_addc_u32 s13, s5, s13 s_waitcnt lgkmcnt(0) global_store_b32 v3, v2, s[12:13] s_branch .LBB0_2 .LBB0_13: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z28mergePartialHistogramsKernelPjS_jj .amdhsa_group_segment_fixed_size 512 .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 9 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z28mergePartialHistogramsKernelPjS_jj, .Lfunc_end0-_Z28mergePartialHistogramsKernelPjS_jj .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: 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: 512 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z28mergePartialHistogramsKernelPjS_jj .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z28mergePartialHistogramsKernelPjS_jj.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 9 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00096cea_00000000-6_mergePartialHistogramsKernel.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 _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj .type _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj, @function _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) 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 _Z28mergePartialHistogramsKernelPjS_jj(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj, .-_Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj .globl _Z28mergePartialHistogramsKernelPjS_jj .type _Z28mergePartialHistogramsKernelPjS_jj, @function _Z28mergePartialHistogramsKernelPjS_jj: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z52__device_stub__Z28mergePartialHistogramsKernelPjS_jjPjS_jj addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z28mergePartialHistogramsKernelPjS_jj, .-_Z28mergePartialHistogramsKernelPjS_jj .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z28mergePartialHistogramsKernelPjS_jj" .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 _Z28mergePartialHistogramsKernelPjS_jj(%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 "mergePartialHistogramsKernel.hip" .globl _Z43__device_stub__mergePartialHistogramsKernelPjS_jj # -- Begin function _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .p2align 4, 0x90 .type _Z43__device_stub__mergePartialHistogramsKernelPjS_jj,@function _Z43__device_stub__mergePartialHistogramsKernelPjS_jj: # @_Z43__device_stub__mergePartialHistogramsKernelPjS_jj .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) 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 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 $_Z28mergePartialHistogramsKernelPjS_jj, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z43__device_stub__mergePartialHistogramsKernelPjS_jj, .Lfunc_end0-_Z43__device_stub__mergePartialHistogramsKernelPjS_jj .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 $_Z28mergePartialHistogramsKernelPjS_jj, %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 _Z28mergePartialHistogramsKernelPjS_jj,@object # @_Z28mergePartialHistogramsKernelPjS_jj .section .rodata,"a",@progbits .globl _Z28mergePartialHistogramsKernelPjS_jj .p2align 3, 0x0 _Z28mergePartialHistogramsKernelPjS_jj: .quad _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .size _Z28mergePartialHistogramsKernelPjS_jj, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z28mergePartialHistogramsKernelPjS_jj" .size .L__unnamed_1, 39 .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 _Z43__device_stub__mergePartialHistogramsKernelPjS_jj .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z28mergePartialHistogramsKernelPjS_jj .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> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
code for sm_80
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> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
.file "tmpxft_0004c39b_00000000-6_generator.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2074: .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 .LFE2074: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Usage: %s row_len col_len folder\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Cannot allocate filepath" .LC2: .string "%s/data.mem" .LC3: .string "w+" .LC4: .string "%s was not opened\n" .LC5: .string "error: can not truncate %s\n" .LC6: .string "ftruncate" .LC7: .string "Cannot mmap %s.\n" .text .globl _Z4initiPPc .type _Z4initiPPc, @function _Z4initiPPc: .LFB2070: .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 $8, %rsp .cfi_def_cfa_offset 48 movq %rsi, %rbx cmpl $4, %edi je .L21 movq (%rsi), %rdx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L21: movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, cols(%rip) movq 16(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, rows(%rip) movq 24(%rbx), %rbp movq %rbp, %rdi call strlen@PLT leaq 128(%rax), %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r13 testq %rax, %rax je .L22 movq %rbp, %r8 leaq .LC2(%rip), %rcx movq %rbx, %rdx movl $2, %esi movq %rax, %rdi movl $0, %eax call __sprintf_chk@PLT leaq .LC3(%rip), %rsi movq %r13, %rdi call fopen@PLT movq %rax, %r12 testq %rax, %rax je .L23 movq rows(%rip), %rbx imulq cols(%rip), %rbx salq $2, %rbx movq %rax, %rdi call fileno@PLT movl %eax, %edi movq %rbx, %rsi call ftruncate@PLT testl %eax, %eax jne .L24 movq %r12, %rdi call fileno@PLT movl %eax, %r8d movq rows(%rip), %rsi imulq cols(%rip), %rsi salq $2, %rsi movl $0, %r9d movl $16385, %ecx movl $3, %edx movl $0, %edi call mmap@PLT movq %rax, data(%rip) testq %rax, %rax je .L25 movl $9, %edi call srand@PLT movl $0, %ebp cmpq $0, rows(%rip) jg .L10 .L11: movq %r12, %rdi call fflush@PLT movq %r12, %rdi call fclose@PLT movq %r13, %rdi call free@PLT addq $8, %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 .L22: .cfi_restore_state leaq .LC1(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L23: movq %r13, %rcx leaq .LC4(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L24: movq %r13, %rcx leaq .LC5(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT leaq .LC6(%rip), %rdi call perror@PLT movl $1, %edi call exit@PLT .L25: movq %r13, %rcx leaq .LC7(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L12: call rand@PLT movq %rbp, %rcx imulq cols(%rip), %rcx addq %rbx, %rcx movslq %eax, %rdx imulq $1717986919, %rdx, %rdx sarq $34, %rdx movl %eax, %esi sarl $31, %esi subl %esi, %edx leal (%rdx,%rdx,4), %edx addl %edx, %edx subl %edx, %eax movq data(%rip), %rdx movl %eax, (%rdx,%rcx,4) addq $1, %rbx cmpq %rbx, cols(%rip) jg .L12 .L13: addq $1, %rbp cmpq %rbp, rows(%rip) jle .L11 .L10: movl $0, %ebx cmpq $0, cols(%rip) jg .L12 jmp .L13 .cfi_endproc .LFE2070: .size _Z4initiPPc, .-_Z4initiPPc .globl main .type main, @function main: .LFB2071: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z4initiPPc movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2071: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2097: .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 .LFE2097: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .globl data .bss .align 8 .type data, @object .size data, 8 data: .zero 8 .globl cols .align 8 .type cols, @object .size cols, 8 cols: .zero 8 .globl rows .align 8 .type rows, @object .size rows, 8 rows: .zero 8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <stdlib.h> #include <time.h> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
.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 <stdlib.h> #include <time.h> #include <assert.h> #include <string.h> #include <sys/mman.h> #include <unistd.h> #include <sys/types.h> #define BLOCK_SIZE 256 #define STR_SIZE 256 #define HALO 1 // halo width along one direction when advancing to the next iteration #define BENCH_PRINT void run(int argc, char** argv); long rows, cols; int* data; #define M_SEED 9 void init(int argc, char** argv) { char *folder; char *filepath; if (argc == 4) { cols = atol(argv[1]); rows = atol(argv[2]); folder = argv[3]; } else { printf("Usage: %s row_len col_len folder\n", argv[0]); exit(0); } filepath = (char *)malloc(sizeof(char) * (strlen(folder) + 128)); if (!filepath) { fprintf(stderr, "Cannot allocate filepath"); exit(EXIT_FAILURE); } sprintf(filepath, "%s/data.mem", folder); FILE *fp; if ((fp = fopen(filepath, "w+")) == 0) { fprintf(stderr, "%s was not opened\n", filepath); exit(EXIT_FAILURE); } if (ftruncate(fileno(fp), sizeof(int) * rows * cols) != 0) { fprintf(stderr, "error: can not truncate %s\n", filepath); perror("ftruncate"); exit(EXIT_FAILURE); } data = (int *)mmap(NULL, sizeof(int) * rows * cols, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_NORESERVE, fileno(fp), 0); if (!data) { fprintf(stderr, "Cannot mmap %s.\n", filepath); exit(EXIT_FAILURE); } int seed = M_SEED; srand(seed); for (long i = 0; i < rows; i++) { for (long j = 0; j < cols; j++) { data[i * cols + j] = rand() % 10; } } fflush(fp); fclose(fp); free(filepath); } int main(int argc, char** argv) { init(argc, argv); return EXIT_SUCCESS; }
.text .file "generator.hip" .globl _Z4initiPPc # -- Begin function _Z4initiPPc .p2align 4, 0x90 .type _Z4initiPPc,@function _Z4initiPPc: # @_Z4initiPPc .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rsi, %rbx cmpl $4, %edi jne .LBB0_15 # %bb.1: movq 8(%rbx), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, cols(%rip) movq 16(%rbx), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, rows(%rip) movq 24(%rbx), %r14 movq %r14, %rdi callq strlen leaq 128(%rax), %rdi callq malloc testq %rax, %rax je .LBB0_16 # %bb.2: movq %rax, %rbx movl $.L.str.2, %esi movq %rax, %rdi movq %r14, %rdx xorl %eax, %eax callq sprintf movl $.L.str.3, %esi movq %rbx, %rdi callq fopen testq %rax, %rax je .LBB0_3 # %bb.5: movq %rax, %r14 movq %rax, %rdi callq fileno movq rows(%rip), %rsi imulq cols(%rip), %rsi shlq $2, %rsi movl %eax, %edi callq ftruncate testl %eax, %eax jne .LBB0_17 # %bb.6: movq rows(%rip), %r15 imulq cols(%rip), %r15 shlq $2, %r15 movq %r14, %rdi callq fileno xorl %edi, %edi movq %r15, %rsi movl $3, %edx movl $16385, %ecx # imm = 0x4001 movl %eax, %r8d xorl %r9d, %r9d callq mmap movq %rax, data(%rip) testq %rax, %rax je .LBB0_7 # %bb.8: movl $9, %edi callq srand cmpq $0, rows(%rip) jle .LBB0_14 # %bb.9: # %.preheader.preheader xorl %r15d, %r15d xorl %r12d, %r12d jmp .LBB0_10 .p2align 4, 0x90 .LBB0_13: # %._crit_edge # in Loop: Header=BB0_10 Depth=1 incq %r12 addq $4, %r15 cmpq rows(%rip), %r12 jge .LBB0_14 .LBB0_10: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_12 Depth 2 cmpq $0, cols(%rip) jle .LBB0_13 # %bb.11: # %.lr.ph.preheader # in Loop: Header=BB0_10 Depth=1 xorl %r13d, %r13d .p2align 4, 0x90 .LBB0_12: # %.lr.ph # Parent Loop BB0_10 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx movq cols(%rip), %rdx movq %r15, %rsi imulq %rdx, %rsi addq data(%rip), %rsi subl %ecx, %eax movl %eax, (%rsi,%r13,4) incq %r13 cmpq %rdx, %r13 jl .LBB0_12 jmp .LBB0_13 .LBB0_14: # %._crit_edge31 movq %r14, %rdi callq fflush movq %r14, %rdi callq fclose movq %rbx, %rdi popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 jmp free # TAILCALL .LBB0_15: .cfi_def_cfa_offset 48 movq (%rbx), %rsi movl $.L.str, %edi xorl %eax, %eax callq printf xorl %edi, %edi callq exit .LBB0_16: movq stderr(%rip), %rcx movl $.L.str.1, %edi movl $24, %esi movl $1, %edx callq fwrite@PLT movl $1, %edi callq exit .LBB0_3: movq stderr(%rip), %rdi movl $.L.str.4, %esi jmp .LBB0_4 .LBB0_17: movq stderr(%rip), %rdi movl $.L.str.5, %esi movq %rbx, %rdx xorl %eax, %eax callq fprintf movl $.L.str.6, %edi callq perror movl $1, %edi callq exit .LBB0_7: movq stderr(%rip), %rdi movl $.L.str.7, %esi .LBB0_4: movq %rbx, %rdx xorl %eax, %eax callq fprintf movl $1, %edi callq exit .Lfunc_end0: .size _Z4initiPPc, .Lfunc_end0-_Z4initiPPc .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 callq _Z4initiPPc xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .type rows,@object # @rows .bss .globl rows .p2align 3, 0x0 rows: .quad 0 # 0x0 .size rows, 8 .type cols,@object # @cols .globl cols .p2align 3, 0x0 cols: .quad 0 # 0x0 .size cols, 8 .type data,@object # @data .globl data .p2align 3, 0x0 data: .quad 0 .size data, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Usage: %s row_len col_len folder\n" .size .L.str, 34 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Cannot allocate filepath" .size .L.str.1, 25 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "%s/data.mem" .size .L.str.2, 12 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "w+" .size .L.str.3, 3 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%s was not opened\n" .size .L.str.4, 19 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "error: can not truncate %s\n" .size .L.str.5, 28 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "ftruncate" .size .L.str.6, 10 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "Cannot mmap %s.\n" .size .L.str.7, 17 .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_0004c39b_00000000-6_generator.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2074: .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 .LFE2074: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "Usage: %s row_len col_len folder\n" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "Cannot allocate filepath" .LC2: .string "%s/data.mem" .LC3: .string "w+" .LC4: .string "%s was not opened\n" .LC5: .string "error: can not truncate %s\n" .LC6: .string "ftruncate" .LC7: .string "Cannot mmap %s.\n" .text .globl _Z4initiPPc .type _Z4initiPPc, @function _Z4initiPPc: .LFB2070: .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 $8, %rsp .cfi_def_cfa_offset 48 movq %rsi, %rbx cmpl $4, %edi je .L21 movq (%rsi), %rdx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L21: movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, cols(%rip) movq 16(%rbx), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, rows(%rip) movq 24(%rbx), %rbp movq %rbp, %rdi call strlen@PLT leaq 128(%rax), %rbx movq %rbx, %rdi call malloc@PLT movq %rax, %r13 testq %rax, %rax je .L22 movq %rbp, %r8 leaq .LC2(%rip), %rcx movq %rbx, %rdx movl $2, %esi movq %rax, %rdi movl $0, %eax call __sprintf_chk@PLT leaq .LC3(%rip), %rsi movq %r13, %rdi call fopen@PLT movq %rax, %r12 testq %rax, %rax je .L23 movq rows(%rip), %rbx imulq cols(%rip), %rbx salq $2, %rbx movq %rax, %rdi call fileno@PLT movl %eax, %edi movq %rbx, %rsi call ftruncate@PLT testl %eax, %eax jne .L24 movq %r12, %rdi call fileno@PLT movl %eax, %r8d movq rows(%rip), %rsi imulq cols(%rip), %rsi salq $2, %rsi movl $0, %r9d movl $16385, %ecx movl $3, %edx movl $0, %edi call mmap@PLT movq %rax, data(%rip) testq %rax, %rax je .L25 movl $9, %edi call srand@PLT movl $0, %ebp cmpq $0, rows(%rip) jg .L10 .L11: movq %r12, %rdi call fflush@PLT movq %r12, %rdi call fclose@PLT movq %r13, %rdi call free@PLT addq $8, %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 .L22: .cfi_restore_state leaq .LC1(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L23: movq %r13, %rcx leaq .LC4(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L24: movq %r13, %rcx leaq .LC5(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT leaq .LC6(%rip), %rdi call perror@PLT movl $1, %edi call exit@PLT .L25: movq %r13, %rcx leaq .LC7(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi call __fprintf_chk@PLT movl $1, %edi call exit@PLT .L12: call rand@PLT movq %rbp, %rcx imulq cols(%rip), %rcx addq %rbx, %rcx movslq %eax, %rdx imulq $1717986919, %rdx, %rdx sarq $34, %rdx movl %eax, %esi sarl $31, %esi subl %esi, %edx leal (%rdx,%rdx,4), %edx addl %edx, %edx subl %edx, %eax movq data(%rip), %rdx movl %eax, (%rdx,%rcx,4) addq $1, %rbx cmpq %rbx, cols(%rip) jg .L12 .L13: addq $1, %rbp cmpq %rbp, rows(%rip) jle .L11 .L10: movl $0, %ebx cmpq $0, cols(%rip) jg .L12 jmp .L13 .cfi_endproc .LFE2070: .size _Z4initiPPc, .-_Z4initiPPc .globl main .type main, @function main: .LFB2071: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z4initiPPc movl $0, %eax addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2071: .size main, .-main .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2097: .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 .LFE2097: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .globl data .bss .align 8 .type data, @object .size data, 8 data: .zero 8 .globl cols .align 8 .type cols, @object .size cols, 8 cols: .zero 8 .globl rows .align 8 .type rows, @object .size rows, 8 rows: .zero 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 "generator.hip" .globl _Z4initiPPc # -- Begin function _Z4initiPPc .p2align 4, 0x90 .type _Z4initiPPc,@function _Z4initiPPc: # @_Z4initiPPc .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rsi, %rbx cmpl $4, %edi jne .LBB0_15 # %bb.1: movq 8(%rbx), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, cols(%rip) movq 16(%rbx), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, rows(%rip) movq 24(%rbx), %r14 movq %r14, %rdi callq strlen leaq 128(%rax), %rdi callq malloc testq %rax, %rax je .LBB0_16 # %bb.2: movq %rax, %rbx movl $.L.str.2, %esi movq %rax, %rdi movq %r14, %rdx xorl %eax, %eax callq sprintf movl $.L.str.3, %esi movq %rbx, %rdi callq fopen testq %rax, %rax je .LBB0_3 # %bb.5: movq %rax, %r14 movq %rax, %rdi callq fileno movq rows(%rip), %rsi imulq cols(%rip), %rsi shlq $2, %rsi movl %eax, %edi callq ftruncate testl %eax, %eax jne .LBB0_17 # %bb.6: movq rows(%rip), %r15 imulq cols(%rip), %r15 shlq $2, %r15 movq %r14, %rdi callq fileno xorl %edi, %edi movq %r15, %rsi movl $3, %edx movl $16385, %ecx # imm = 0x4001 movl %eax, %r8d xorl %r9d, %r9d callq mmap movq %rax, data(%rip) testq %rax, %rax je .LBB0_7 # %bb.8: movl $9, %edi callq srand cmpq $0, rows(%rip) jle .LBB0_14 # %bb.9: # %.preheader.preheader xorl %r15d, %r15d xorl %r12d, %r12d jmp .LBB0_10 .p2align 4, 0x90 .LBB0_13: # %._crit_edge # in Loop: Header=BB0_10 Depth=1 incq %r12 addq $4, %r15 cmpq rows(%rip), %r12 jge .LBB0_14 .LBB0_10: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_12 Depth 2 cmpq $0, cols(%rip) jle .LBB0_13 # %bb.11: # %.lr.ph.preheader # in Loop: Header=BB0_10 Depth=1 xorl %r13d, %r13d .p2align 4, 0x90 .LBB0_12: # %.lr.ph # Parent Loop BB0_10 Depth=1 # => This Inner Loop Header: Depth=2 callq rand cltq imulq $1717986919, %rax, %rcx # imm = 0x66666667 movq %rcx, %rdx shrq $63, %rdx sarq $34, %rcx addl %edx, %ecx addl %ecx, %ecx leal (%rcx,%rcx,4), %ecx movq cols(%rip), %rdx movq %r15, %rsi imulq %rdx, %rsi addq data(%rip), %rsi subl %ecx, %eax movl %eax, (%rsi,%r13,4) incq %r13 cmpq %rdx, %r13 jl .LBB0_12 jmp .LBB0_13 .LBB0_14: # %._crit_edge31 movq %r14, %rdi callq fflush movq %r14, %rdi callq fclose movq %rbx, %rdi popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 jmp free # TAILCALL .LBB0_15: .cfi_def_cfa_offset 48 movq (%rbx), %rsi movl $.L.str, %edi xorl %eax, %eax callq printf xorl %edi, %edi callq exit .LBB0_16: movq stderr(%rip), %rcx movl $.L.str.1, %edi movl $24, %esi movl $1, %edx callq fwrite@PLT movl $1, %edi callq exit .LBB0_3: movq stderr(%rip), %rdi movl $.L.str.4, %esi jmp .LBB0_4 .LBB0_17: movq stderr(%rip), %rdi movl $.L.str.5, %esi movq %rbx, %rdx xorl %eax, %eax callq fprintf movl $.L.str.6, %edi callq perror movl $1, %edi callq exit .LBB0_7: movq stderr(%rip), %rdi movl $.L.str.7, %esi .LBB0_4: movq %rbx, %rdx xorl %eax, %eax callq fprintf movl $1, %edi callq exit .Lfunc_end0: .size _Z4initiPPc, .Lfunc_end0-_Z4initiPPc .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %rax .cfi_def_cfa_offset 16 callq _Z4initiPPc xorl %eax, %eax popq %rcx .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .type rows,@object # @rows .bss .globl rows .p2align 3, 0x0 rows: .quad 0 # 0x0 .size rows, 8 .type cols,@object # @cols .globl cols .p2align 3, 0x0 cols: .quad 0 # 0x0 .size cols, 8 .type data,@object # @data .globl data .p2align 3, 0x0 data: .quad 0 .size data, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "Usage: %s row_len col_len folder\n" .size .L.str, 34 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "Cannot allocate filepath" .size .L.str.1, 25 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "%s/data.mem" .size .L.str.2, 12 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "w+" .size .L.str.3, 3 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%s was not opened\n" .size .L.str.4, 19 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "error: can not truncate %s\n" .size .L.str.5, 28 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "ftruncate" .size .L.str.6, 10 .type .L.str.7,@object # @.str.7 .L.str.7: .asciz "Cannot mmap %s.\n" .size .L.str.7, 17 .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> #define cudaErrorCheck(call) \ do { \ cudaError_t cuErr = call; \ if (cudaSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ cudaGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(cudaGetLastError()); cudaDeviceSynchronize(); int *count_addr; cudaErrorCheck(cudaGetSymbolAddress((void**)&count_addr, count)); cudaErrorCheck(cudaMemcpy(&result, count_addr, sizeof(int), cudaMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); return 0; }
code for sm_80 Function : _Z12getThreadNumv .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*/ MOV R2, c[0x4][0x0] ; /* 0x0100000000027a02 */ /* 0x000fe20000000f00 */ /*0020*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x4][0x4] ; /* 0x01000100ff037624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc80000000a00 */ /*0040*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ea2000c1e1900 */ /*0050*/ YIELD ; /* 0x0000000000007946 */ /* 0x000fe20003800000 */ /*0060*/ IADD3 R5, R4, 0x1, RZ ; /* 0x0000000104057810 */ /* 0x004fcc0007ffe0ff */ /*0070*/ ATOMG.E.CAS.STRONG.GPU PT, R5, [R2], R4, R5 ; /* 0x00000004020573a9 */ /* 0x000ea400001ee105 */ /*0080*/ ISETP.NE.AND P0, PT, R4, R5, PT ; /* 0x000000050400720c */ /* 0x004fda0003f05270 */ /*0090*/ @P0 BRA 0x40 ; /* 0xffffffa000000947 */ /* 0x000fea000383ffff */ /*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 <stdio.h> #define cudaErrorCheck(call) \ do { \ cudaError_t cuErr = call; \ if (cudaSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ cudaGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(cudaGetLastError()); cudaDeviceSynchronize(); int *count_addr; cudaErrorCheck(cudaGetSymbolAddress((void**)&count_addr, count)); cudaErrorCheck(cudaMemcpy(&result, count_addr, sizeof(int), cudaMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); return 0; }
.file "tmpxft_00052755_00000000-6_testAtomic.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z12getThreadNumvv .type _Z31__device_stub__Z12getThreadNumvv, @function _Z31__device_stub__Z12getThreadNumvv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z12getThreadNumv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z31__device_stub__Z12getThreadNumvv, .-_Z31__device_stub__Z12getThreadNumvv .globl _Z12getThreadNumv .type _Z12getThreadNumv, @function _Z12getThreadNumv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z12getThreadNumvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z12getThreadNumv, .-_Z12getThreadNumv .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "/home/ubuntu/Datasets/stackv2/train-structured/FuriousBerserker/ConcurrencyError/master/CUDA/testAtomic/testAtomic.cu" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "CUDA Error - %s:%d: '%s'\n" .section .rodata.str1.8 .align 8 .LC2: .string "thread num, expected = %d, actual = %d\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $56, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $-1, 8(%rsp) movl $512, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $100, 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 .L18 .L12: call cudaGetLastError@PLT testl %eax, %eax je .L13 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $30, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L18: call _Z31__device_stub__Z12getThreadNumvv jmp .L12 .L13: call cudaDeviceSynchronize@PLT leaq 24(%rsp), %rdi leaq _ZL5count(%rip), %rsi call cudaGetSymbolAddress@PLT testl %eax, %eax je .L14 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $33, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L14: leaq 8(%rsp), %rdi movl $2, %ecx movl $4, %edx movq 24(%rsp), %rsi call cudaMemcpy@PLT testl %eax, %eax je .L15 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $34, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L15: movl 8(%rsp), %ecx movl $51200, %edx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L19 movl $0, %eax addq $56, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z12getThreadNumv" .LC4: .string "count" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z12getThreadNumv(%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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _ZL5count(%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 .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL5count .comm _ZL5count,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 <stdio.h> #define cudaErrorCheck(call) \ do { \ cudaError_t cuErr = call; \ if (cudaSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ cudaGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(cudaGetLastError()); cudaDeviceSynchronize(); int *count_addr; cudaErrorCheck(cudaGetSymbolAddress((void**)&count_addr, count)); cudaErrorCheck(cudaMemcpy(&result, count_addr, sizeof(int), cudaMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #define cudaErrorCheck(call) \ do { \ hipError_t cuErr = call; \ if (hipSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ hipGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(hipGetLastError()); hipDeviceSynchronize(); int *count_addr; cudaErrorCheck(hipGetSymbolAddress((void**)&count_addr, HIP_SYMBOL(count))); cudaErrorCheck(hipMemcpy(&result, count_addr, sizeof(int), hipMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); 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> #define cudaErrorCheck(call) \ do { \ hipError_t cuErr = call; \ if (hipSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ hipGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(hipGetLastError()); hipDeviceSynchronize(); int *count_addr; cudaErrorCheck(hipGetSymbolAddress((void**)&count_addr, HIP_SYMBOL(count))); cudaErrorCheck(hipMemcpy(&result, count_addr, sizeof(int), hipMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z12getThreadNumv .globl _Z12getThreadNumv .p2align 8 .type _Z12getThreadNumv,@function _Z12getThreadNumv: v_mov_b32_e32 v0, 0 s_mov_b32 s0, 0 s_getpc_b64 s[2:3] s_add_u32 s2, s2, count@rel32@lo+4 s_addc_u32 s3, s3, count@rel32@hi+12 .LBB0_1: global_load_b32 v2, v0, s[2:3] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, 1, v2 global_atomic_cmpswap_b32 v1, v0, v[1:2], s[2:3] glc s_waitcnt vmcnt(0) v_cmp_eq_u32_e32 vcc_lo, v2, v1 s_or_b32 s0, vcc_lo, s0 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s0 s_cbranch_execnz .LBB0_1 s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z12getThreadNumv .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 4 .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 _Z12getThreadNumv, .Lfunc_end0-_Z12getThreadNumv .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .protected count .type count,@object .section .bss,"aw",@nobits .globl count .p2align 2, 0x0 count: .long 0 .size count, 4 .type __hip_cuid_,@object .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 count .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z12getThreadNumv .private_segment_fixed_size: 0 .sgpr_count: 6 .sgpr_spill_count: 0 .symbol: _Z12getThreadNumv.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 <stdio.h> #define cudaErrorCheck(call) \ do { \ hipError_t cuErr = call; \ if (hipSuccess != cuErr) { \ printf("CUDA Error - %s:%d: '%s'\n", __FILE__, __LINE__, \ hipGetErrorString(cuErr)); \ exit(0); \ } \ } while (0) #define BLOCK_NUM 100 #define BLOCK_DIM 512 __device__ int count = 0; __global__ void getThreadNum() { int oldVal, returnVal; do { oldVal = count; returnVal = atomicCAS(&count, oldVal, oldVal + 1); } while (oldVal != returnVal); } int main() { int result = -1; getThreadNum<<<BLOCK_NUM, BLOCK_DIM>>>(); cudaErrorCheck(hipGetLastError()); hipDeviceSynchronize(); int *count_addr; cudaErrorCheck(hipGetSymbolAddress((void**)&count_addr, HIP_SYMBOL(count))); cudaErrorCheck(hipMemcpy(&result, count_addr, sizeof(int), hipMemcpyDeviceToHost)); printf("thread num, expected = %d, actual = %d\n", BLOCK_NUM * BLOCK_DIM, result); return 0; }
.text .file "testAtomic.hip" .globl _Z27__device_stub__getThreadNumv # -- Begin function _Z27__device_stub__getThreadNumv .p2align 4, 0x90 .type _Z27__device_stub__getThreadNumv,@function _Z27__device_stub__getThreadNumv: # @_Z27__device_stub__getThreadNumv .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 $_Z12getThreadNumv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $72, %rsp .cfi_adjust_cfa_offset -72 retq .Lfunc_end0: .size _Z27__device_stub__getThreadNumv, .Lfunc_end0-_Z27__device_stub__getThreadNumv .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 $-1, 12(%rsp) movabsq $4294967396, %rdi # imm = 0x100000064 leaq 412(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 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 $_Z12getThreadNumv, %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: callq hipGetLastError testl %eax, %eax jne .LBB1_3 # %bb.5: callq hipDeviceSynchronize leaq 16(%rsp), %rdi movl $count, %esi callq hipGetSymbolAddress testl %eax, %eax jne .LBB1_6 # %bb.7: movq 16(%rsp), %rsi leaq 12(%rsp), %rdi movl $4, %edx movl $2, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_8 # %bb.9: movl 12(%rsp), %edx movl $.L.str.2, %edi movl $51200, %esi # imm = 0xC800 xorl %eax, %eax callq printf xorl %eax, %eax addq $72, %rsp .cfi_def_cfa_offset 8 retq .LBB1_3: .cfi_def_cfa_offset 80 movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $32, %edx jmp .LBB1_4 .LBB1_6: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $35, %edx jmp .LBB1_4 .LBB1_8: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $36, %edx .LBB1_4: movq %rax, %rcx xorl %eax, %eax callq printf xorl %edi, %edi callq exit .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 $_Z12getThreadNumv, %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 $count, %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 count,@object # @count .local count .comm count,4,4 .type _Z12getThreadNumv,@object # @_Z12getThreadNumv .section .rodata,"a",@progbits .globl _Z12getThreadNumv .p2align 3, 0x0 _Z12getThreadNumv: .quad _Z27__device_stub__getThreadNumv .size _Z12getThreadNumv, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "CUDA Error - %s:%d: '%s'\n" .size .L.str, 26 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/FuriousBerserker/ConcurrencyError/master/CUDA/testAtomic/testAtomic.hip" .size .L.str.1, 129 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "thread num, expected = %d, actual = %d\n" .size .L.str.2, 40 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z12getThreadNumv" .size .L__unnamed_1, 18 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "count" .size .L__unnamed_2, 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 _Z27__device_stub__getThreadNumv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym count .addrsig_sym _Z12getThreadNumv .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 : _Z12getThreadNumv .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*/ MOV R2, c[0x4][0x0] ; /* 0x0100000000027a02 */ /* 0x000fe20000000f00 */ /*0020*/ IMAD.MOV.U32 R3, RZ, RZ, c[0x4][0x4] ; /* 0x01000100ff037624 */ /* 0x000fe200078e00ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fc80000000a00 */ /*0040*/ LDG.E R4, [R2.64] ; /* 0x0000000402047981 */ /* 0x000ea2000c1e1900 */ /*0050*/ YIELD ; /* 0x0000000000007946 */ /* 0x000fe20003800000 */ /*0060*/ IADD3 R5, R4, 0x1, RZ ; /* 0x0000000104057810 */ /* 0x004fcc0007ffe0ff */ /*0070*/ ATOMG.E.CAS.STRONG.GPU PT, R5, [R2], R4, R5 ; /* 0x00000004020573a9 */ /* 0x000ea400001ee105 */ /*0080*/ ISETP.NE.AND P0, PT, R4, R5, PT ; /* 0x000000050400720c */ /* 0x004fda0003f05270 */ /*0090*/ @P0 BRA 0x40 ; /* 0xffffffa000000947 */ /* 0x000fea000383ffff */ /*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 _Z12getThreadNumv .globl _Z12getThreadNumv .p2align 8 .type _Z12getThreadNumv,@function _Z12getThreadNumv: v_mov_b32_e32 v0, 0 s_mov_b32 s0, 0 s_getpc_b64 s[2:3] s_add_u32 s2, s2, count@rel32@lo+4 s_addc_u32 s3, s3, count@rel32@hi+12 .LBB0_1: global_load_b32 v2, v0, s[2:3] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, 1, v2 global_atomic_cmpswap_b32 v1, v0, v[1:2], s[2:3] glc s_waitcnt vmcnt(0) v_cmp_eq_u32_e32 vcc_lo, v2, v1 s_or_b32 s0, vcc_lo, s0 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s0 s_cbranch_execnz .LBB0_1 s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z12getThreadNumv .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 4 .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 _Z12getThreadNumv, .Lfunc_end0-_Z12getThreadNumv .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .protected count .type count,@object .section .bss,"aw",@nobits .globl count .p2align 2, 0x0 count: .long 0 .size count, 4 .type __hip_cuid_,@object .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 count .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z12getThreadNumv .private_segment_fixed_size: 0 .sgpr_count: 6 .sgpr_spill_count: 0 .symbol: _Z12getThreadNumv.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_00052755_00000000-6_testAtomic.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z12getThreadNumvv .type _Z31__device_stub__Z12getThreadNumvv, @function _Z31__device_stub__Z12getThreadNumvv: .LFB2082: .cfi_startproc endbr64 subq $88, %rsp .cfi_def_cfa_offset 96 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $1, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) leaq 8(%rsp), %rcx movq %rsp, %rdx leaq 28(%rsp), %rsi leaq 16(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 72(%rsp), %rax subq %fs:40, %rax jne .L8 addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 8(%rsp) .cfi_def_cfa_offset 104 pushq 8(%rsp) .cfi_def_cfa_offset 112 leaq 80(%rsp), %r9 movq 44(%rsp), %rcx movl 52(%rsp), %r8d movq 32(%rsp), %rsi movl 40(%rsp), %edx leaq _Z12getThreadNumv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z31__device_stub__Z12getThreadNumvv, .-_Z31__device_stub__Z12getThreadNumvv .globl _Z12getThreadNumv .type _Z12getThreadNumv, @function _Z12getThreadNumv: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z12getThreadNumvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z12getThreadNumv, .-_Z12getThreadNumv .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "/home/ubuntu/Datasets/stackv2/train-structured/FuriousBerserker/ConcurrencyError/master/CUDA/testAtomic/testAtomic.cu" .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "CUDA Error - %s:%d: '%s'\n" .section .rodata.str1.8 .align 8 .LC2: .string "thread num, expected = %d, actual = %d\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $56, %rsp .cfi_def_cfa_offset 64 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $-1, 8(%rsp) movl $512, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $100, 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 .L18 .L12: call cudaGetLastError@PLT testl %eax, %eax je .L13 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $30, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L18: call _Z31__device_stub__Z12getThreadNumvv jmp .L12 .L13: call cudaDeviceSynchronize@PLT leaq 24(%rsp), %rdi leaq _ZL5count(%rip), %rsi call cudaGetSymbolAddress@PLT testl %eax, %eax je .L14 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $33, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L14: leaq 8(%rsp), %rdi movl $2, %ecx movl $4, %edx movq 24(%rsp), %rsi call cudaMemcpy@PLT testl %eax, %eax je .L15 movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %r8 movl $34, %ecx leaq .LC0(%rip), %rdx leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %edi call exit@PLT .L15: movl 8(%rsp), %ecx movl $51200, %edx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L19 movl $0, %eax addq $56, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L19: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z12getThreadNumv" .LC4: .string "count" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC3(%rip), %rdx movq %rdx, %rcx leaq _Z12getThreadNumv(%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 .LC4(%rip), %rdx movq %rdx, %rcx leaq _ZL5count(%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 .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL5count .comm _ZL5count,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 "testAtomic.hip" .globl _Z27__device_stub__getThreadNumv # -- Begin function _Z27__device_stub__getThreadNumv .p2align 4, 0x90 .type _Z27__device_stub__getThreadNumv,@function _Z27__device_stub__getThreadNumv: # @_Z27__device_stub__getThreadNumv .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 $_Z12getThreadNumv, %edi pushq (%rsp) .cfi_adjust_cfa_offset 8 pushq 16(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $72, %rsp .cfi_adjust_cfa_offset -72 retq .Lfunc_end0: .size _Z27__device_stub__getThreadNumv, .Lfunc_end0-_Z27__device_stub__getThreadNumv .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 $-1, 12(%rsp) movabsq $4294967396, %rdi # imm = 0x100000064 leaq 412(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_2 # %bb.1: leaq 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 $_Z12getThreadNumv, %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: callq hipGetLastError testl %eax, %eax jne .LBB1_3 # %bb.5: callq hipDeviceSynchronize leaq 16(%rsp), %rdi movl $count, %esi callq hipGetSymbolAddress testl %eax, %eax jne .LBB1_6 # %bb.7: movq 16(%rsp), %rsi leaq 12(%rsp), %rdi movl $4, %edx movl $2, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_8 # %bb.9: movl 12(%rsp), %edx movl $.L.str.2, %edi movl $51200, %esi # imm = 0xC800 xorl %eax, %eax callq printf xorl %eax, %eax addq $72, %rsp .cfi_def_cfa_offset 8 retq .LBB1_3: .cfi_def_cfa_offset 80 movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $32, %edx jmp .LBB1_4 .LBB1_6: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $35, %edx jmp .LBB1_4 .LBB1_8: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movl $.L.str.1, %esi movl $36, %edx .LBB1_4: movq %rax, %rcx xorl %eax, %eax callq printf xorl %edi, %edi callq exit .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 $_Z12getThreadNumv, %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 $count, %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 count,@object # @count .local count .comm count,4,4 .type _Z12getThreadNumv,@object # @_Z12getThreadNumv .section .rodata,"a",@progbits .globl _Z12getThreadNumv .p2align 3, 0x0 _Z12getThreadNumv: .quad _Z27__device_stub__getThreadNumv .size _Z12getThreadNumv, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "CUDA Error - %s:%d: '%s'\n" .size .L.str, 26 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/FuriousBerserker/ConcurrencyError/master/CUDA/testAtomic/testAtomic.hip" .size .L.str.1, 129 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "thread num, expected = %d, actual = %d\n" .size .L.str.2, 40 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z12getThreadNumv" .size .L__unnamed_1, 18 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "count" .size .L__unnamed_2, 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 _Z27__device_stub__getThreadNumv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym count .addrsig_sym _Z12getThreadNumv .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.
__device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
code for sm_80 Function : EVAPORATION .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R3, SR_CTAID.Y ; /* 0x0000000000037919 */ /* 0x000e280000002600 */ /*0020*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e280000002200 */ /*0030*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e680000002500 */ /*0040*/ S2R R5, SR_TID.X ; /* 0x0000000000057919 */ /* 0x000e620000002100 */ /*0050*/ IMAD R3, R3, c[0x0][0x4], R2 ; /* 0x0000010003037a24 */ /* 0x001fca00078e0202 */ /*0060*/ ISETP.GE.AND P0, PT, R3, c[0x0][0x164], PT ; /* 0x0000590003007a0c */ /* 0x000fe20003f06270 */ /*0070*/ IMAD R0, R0, c[0x0][0x0], R5 ; /* 0x0000000000007a24 */ /* 0x002fca00078e0205 */ /*0080*/ ISETP.GE.OR P0, PT, R0, c[0x0][0x160], P0 ; /* 0x0000580000007a0c */ /* 0x000fda0000706670 */ /*0090*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*00a0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fe200000001ff */ /*00b0*/ IMAD R3, R3, c[0x0][0x160], R0 ; /* 0x0000580003037a24 */ /* 0x000fe200078e0200 */ /*00c0*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd00000000a00 */ /*00d0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x168] ; /* 0x00005a0003027625 */ /* 0x000fca00078e0202 */ /*00e0*/ LDG.E R0, [R2.64] ; /* 0x0000000402007981 */ /* 0x000ea4000c1e1900 */ /*00f0*/ FFMA R5, -R0, c[0x0][0x170], R0 ; /* 0x00005c0000057a23 */ /* 0x004fca0000000100 */ /*0100*/ STG.E [R2.64], R5 ; /* 0x0000000502007986 */ /* 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 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
__device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
.file "tmpxft_000a65b1_00000000-6_Evaporation_2D.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 _Z3getiii .type _Z3getiii, @function _Z3getiii: .LFB2027: .cfi_startproc endbr64 pushq %rax .cfi_def_cfa_offset 16 popq %rax .cfi_def_cfa_offset 8 subq $24, %rsp .cfi_def_cfa_offset 32 movl $1, 12(%rsp) movl 12(%rsp), %edi call exit@PLT .cfi_endproc .LFE2027: .size _Z3getiii, .-_Z3getiii .globl _Z34__device_stub__Z11EVAPORATIONiiPffiiPff .type _Z34__device_stub__Z11EVAPORATIONiiPffiiPff, @function _Z34__device_stub__Z11EVAPORATIONiiPffiiPff: .LFB2052: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 28(%rsp) movl %esi, 24(%rsp) movq %rdx, 16(%rsp) movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 24(%rsp), %rax movq %rax, 104(%rsp) leaq 16(%rsp), %rax movq %rax, 112(%rsp) leaq 12(%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 .L9 .L5: movq 136(%rsp), %rax subq %fs:40, %rax jne .L10 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L9: .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 EVAPORATION(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L5 .L10: call __stack_chk_fail@PLT .cfi_endproc .LFE2052: .size _Z34__device_stub__Z11EVAPORATIONiiPffiiPff, .-_Z34__device_stub__Z11EVAPORATIONiiPffiiPff .globl EVAPORATION .type EVAPORATION, @function EVAPORATION: .LFB2053: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z34__device_stub__Z11EVAPORATIONiiPffiiPff addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2053: .size EVAPORATION, .-EVAPORATION .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "EVAPORATION" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2055: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq EVAPORATION(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2055: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
__device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
#include <hip/hip_runtime.h> __device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> __device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected EVAPORATION .globl EVAPORATION .p2align 8 .type EVAPORATION,@function EVAPORATION: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b64 s[4:5], s[0:1], 0x0 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s3, s2, 0xffff s_lshr_b32 s2, s2, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_mad_u64_u32 v[0:1], null, s14, s3, v[2:3] v_mad_u64_u32 v[1:2], null, s15, s2, v[3:4] v_cmp_gt_i32_e32 vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cmp_gt_i32_e64 s2, s5, v1 s_and_b32 s2, vcc_lo, s2 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_2 s_load_b64 s[2:3], s[0:1], 0x8 v_mad_u64_u32 v[2:3], null, v1, s4, v[0:1] s_load_b32 s0, s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v3, 31, v2 v_lshlrev_b64 v[0:1], 2, v[2:3] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s2, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s3, v1, vcc_lo global_load_b32 v2, v[0:1], off s_waitcnt vmcnt(0) v_fma_f32 v2, -v2, s0, v2 global_store_b32 v[0:1], v2, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel EVAPORATION .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 280 .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 5 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size EVAPORATION, .Lfunc_end0-EVAPORATION .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .offset: 0 .size: 4 .value_kind: by_value - .offset: 4 .size: 4 .value_kind: by_value - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .offset: 16 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .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: EVAPORATION .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: EVAPORATION.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 5 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> __device__ int get(int x, int y,int width){ return y * width +x; } extern "C" __global__ void EVAPORATION( int width, int height, float *values, float evapCoef) { int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; if (i < width && j < height ){//TODO + alone int k = get(i,j,width); values[k] -= values[k] * evapCoef; } }
.text .file "Evaporation_2D.hip" .globl __device_stub__EVAPORATION # -- Begin function __device_stub__EVAPORATION .p2align 4, 0x90 .type __device_stub__EVAPORATION,@function __device_stub__EVAPORATION: # @__device_stub__EVAPORATION .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 20(%rsp) movl %esi, 16(%rsp) movq %rdx, 72(%rsp) movss %xmm0, 12(%rsp) leaq 20(%rsp), %rax movq %rax, 80(%rsp) leaq 16(%rsp), %rax movq %rax, 88(%rsp) leaq 72(%rsp), %rax movq %rax, 96(%rsp) leaq 12(%rsp), %rax movq %rax, 104(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $EVAPORATION, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size __device_stub__EVAPORATION, .Lfunc_end0-__device_stub__EVAPORATION .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 $EVAPORATION, %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 EVAPORATION,@object # @EVAPORATION .section .rodata,"a",@progbits .globl EVAPORATION .p2align 3, 0x0 EVAPORATION: .quad __device_stub__EVAPORATION .size EVAPORATION, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "EVAPORATION" .size .L__unnamed_1, 12 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __device_stub__EVAPORATION .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym EVAPORATION .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_