system
stringclasses
7 values
input
stringlengths
10
94.5k
output
stringlengths
14
55.5k
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <iostream> #include <cstdlib> #include <cstdio> #include <curand_kernel.h> #include <thrust/reduce.h> #include <thrust/functional.h> #include <thrust/execution_policy.h> #include <thrust/extrema.h> #include <thrust/device_ptr.h> #include <sys/time.h> using namespace std; __global__ void partition_step (curandState * state, unsigned long seed ) { int i= blockDim.x * blockIdx.x + threadIdx.x; curand_init (seed, i, 0, &state[i]); } __global__ void randomColouring (curandState* globalState, int *degreeCount, int n, int limit){ int i= blockDim.x * blockIdx.x + threadIdx.x; curandState localState = globalState[i]; float RANDOM = curand_uniform( &localState ); globalState[i] = localState; RANDOM *= (limit - 1 + 0.999999); RANDOM += 1; degreeCount[i] = (int) RANDOM; } __global__ void conflictDetection (int *vertexArray, int *neighbourArray, int *degreeCount, int n, int m, int *detectConflict){ int i= blockDim.x * blockIdx.x + threadIdx.x; if (i>=n){ return; } int myColour = degreeCount[i]; int incoming = -1, stop = -1; incoming = vertexArray[i]; if (i==n-1){ stop = m; } else{ stop = vertexArray[i+1]; } for (int j=incoming; j<stop; j++){ if (degreeCount[neighbourArray[j]-1] == myColour){ detectConflict[i]=1; break; } } } __global__ void degreeCalc (int *vertexArray, int *neighbourArray, int *degreeCount, int n, int m){ int i= blockDim.x * blockIdx.x + threadIdx.x; if (i>=n){ return; } int incoming = -1, stop = -1; int diff=0; incoming = vertexArray[i]; if (i==n-1){ stop = m; } else{ stop = vertexArray[i+1]; } diff = stop-incoming; atomicAdd(&degreeCount[i], diff); for (int j=incoming; j<stop; j++){ atomicAdd(&degreeCount[neighbourArray[j]-1], 1); } } int main(int argc, char const *argv[]) { int n, m; // Enter number of vertices and edges cin>>n>>m; int h_vertexArray[n]; int h_neighbourArray[m]; int h_degreeCount[n]; int h_detectConflict[n]; // Cuda memory allocation size_t bytes = n*sizeof(int); int *d_vertexArray = NULL; cudaMalloc((void **)&d_vertexArray, bytes); int *d_neighbourArray = NULL; cudaMalloc((void **)&d_neighbourArray, m*sizeof(int)); int *d_detectConflict = NULL; cudaMalloc((void **)&d_detectConflict, bytes); cudaMemset((void *)d_detectConflict, 0,bytes); int *d_degreeCount = NULL; cudaMalloc((void **)&d_degreeCount, bytes); cudaMemset((void *)d_degreeCount, 0, bytes); curandState* partition_states; cudaMalloc ( &partition_states, n*sizeof( curandState ) ); for (int i = 0; i < n; ++i) { /* code */ h_vertexArray[i]=m; } int temp = 0; int current = 0; int mark = 1; // Add the graph based on input file for (int i = 0; i < m; ++i) { /* code */ int incoming; int end; cin>>incoming>>end; incoming++; end++; if (incoming!=mark){ if (incoming == mark+1 && h_vertexArray[mark-1]!=m){ } else{ for (int j = mark; j<incoming; j++){ h_vertexArray[j-1]=temp; } } mark = incoming; } if (incoming==current){ h_neighbourArray[temp]=end; temp++; } else { current = incoming; h_vertexArray[current-1]=temp; h_neighbourArray[temp]=end; temp++; } } cudaMemcpy(d_vertexArray, h_vertexArray, n*sizeof(int), cudaMemcpyHostToDevice); cudaMemcpy(d_neighbourArray, h_neighbourArray, m*sizeof(int), cudaMemcpyHostToDevice); int threadsPerBlock = 512; int blocksPerGrid = (n + threadsPerBlock -1)/threadsPerBlock; struct timeval startTime; struct timeval endTime; struct timezone startZone; struct timezone endZone; long startt,endt; double overhead; cout<<threadsPerBlock<<" "<<blocksPerGrid<<endl; gettimeofday(&startTime,&startZone); // Step 0 : Calculate degree of each vertex degreeCalc<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m); thrust::device_ptr<int> d_ptr = thrust::device_pointer_cast(d_degreeCount); int max = *(thrust::max_element(d_ptr, d_ptr + n)); cout<<"Max number of colours = "<<max<<endl; partition_step <<<blocksPerGrid, threadsPerBlock>>> ( partition_states, time(NULL) ); // Step 1 - Randomly assign colours randomColouring<<<blocksPerGrid, threadsPerBlock>>>(partition_states, d_degreeCount, n, max); cudaMemcpy(h_degreeCount, d_degreeCount, n*sizeof(int), cudaMemcpyDeviceToHost); cout<<"randomColouring"<<endl; for (int i=0; i<n; i++){ cout<<"Color of"<<i<<": "<<h_degreeCount[i]<<endl; } cout<<endl; conflictDetection<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m, d_detectConflict); thrust::device_ptr<int> d_detectConflict_ptr = thrust::device_pointer_cast(d_detectConflict); int count1 = thrust::reduce(d_detectConflict_ptr, d_detectConflict_ptr + n); cudaMemcpy(h_detectConflict, d_detectConflict, n*sizeof(int), cudaMemcpyDeviceToHost); int countnew=0; int old_colors[n]; for (int i = 0; i < n; ++i) { /* code */ old_colors[i] = -1; } for (int i=0; i<n-1; i++){ if (h_detectConflict[i]==0){ continue; } countnew++; bool usedColours[n]; fill(usedColours, usedColours+n, false); int incoming = -1, stop = -1; incoming = h_vertexArray[i]; stop = h_vertexArray[i+1]; old_colors[i] = h_degreeCount[i]; for (int j=incoming; j<stop; j++){ usedColours[h_degreeCount[h_neighbourArray[j]-1]-1] = true; } for (int j=0; j<n; j++){ if (usedColours[j]==false){ h_degreeCount[i]=j+1; break; } } } if (h_detectConflict[n-1]!=0){ bool usedColours[n]; countnew++; fill(usedColours, usedColours+n, false); int incoming = -1, stop = -1; incoming = h_vertexArray[n-1]; stop = m; for (int j=incoming; j<stop; j++){ usedColours[h_degreeCount[h_neighbourArray[j]-1]-1] = true; } for (int j=0; j<n; j++){ if (usedColours[j]==false){ h_degreeCount[n-1]=j+1; break; } } } for (int i = 0; i < n; ++i) { cout<<"Colour of i from" <<i <<" "<<old_colors[i]<<":"<<h_degreeCount[i]<<endl; } cudaMemset((void *)d_detectConflict, 0, (n)*sizeof(int)); cudaMemcpy(d_degreeCount, h_degreeCount, n*sizeof(int), cudaMemcpyHostToDevice); conflictDetection<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m, d_detectConflict); gettimeofday(&endTime,&endZone); startt = startTime.tv_sec*1000000+startTime.tv_usec; endt = endTime.tv_sec*1000000+endTime.tv_usec; overhead = (endt-startt)/1000000.0; count1 = thrust::reduce(d_detectConflict_ptr, d_detectConflict_ptr + n); cout<<"Count: "<<count1<<" "<<countnew<<endl; cout<<"time taken is"<<overhead<<endl; cudaFree(d_neighbourArray); cudaFree(d_vertexArray); cudaFree(d_degreeCount); cudaFree(d_detectConflict); cudaDeviceReset(); return 0; }
#include <hip/hip_runtime.h> #include <iostream> #include <cstdlib> #include <cstdio> #include <hiprand/hiprand_kernel.h> #include <thrust/reduce.h> #include <thrust/functional.h> #include <thrust/execution_policy.h> #include <thrust/extrema.h> #include <thrust/device_ptr.h> #include <sys/time.h> using namespace std; __global__ void partition_step (hiprandState * state, unsigned long seed ) { int i= blockDim.x * blockIdx.x + threadIdx.x; hiprand_init (seed, i, 0, &state[i]); } __global__ void randomColouring (hiprandState* globalState, int *degreeCount, int n, int limit){ int i= blockDim.x * blockIdx.x + threadIdx.x; hiprandState localState = globalState[i]; float RANDOM = hiprand_uniform( &localState ); globalState[i] = localState; RANDOM *= (limit - 1 + 0.999999); RANDOM += 1; degreeCount[i] = (int) RANDOM; } __global__ void conflictDetection (int *vertexArray, int *neighbourArray, int *degreeCount, int n, int m, int *detectConflict){ int i= blockDim.x * blockIdx.x + threadIdx.x; if (i>=n){ return; } int myColour = degreeCount[i]; int incoming = -1, stop = -1; incoming = vertexArray[i]; if (i==n-1){ stop = m; } else{ stop = vertexArray[i+1]; } for (int j=incoming; j<stop; j++){ if (degreeCount[neighbourArray[j]-1] == myColour){ detectConflict[i]=1; break; } } } __global__ void degreeCalc (int *vertexArray, int *neighbourArray, int *degreeCount, int n, int m){ int i= blockDim.x * blockIdx.x + threadIdx.x; if (i>=n){ return; } int incoming = -1, stop = -1; int diff=0; incoming = vertexArray[i]; if (i==n-1){ stop = m; } else{ stop = vertexArray[i+1]; } diff = stop-incoming; atomicAdd(&degreeCount[i], diff); for (int j=incoming; j<stop; j++){ atomicAdd(&degreeCount[neighbourArray[j]-1], 1); } } int main(int argc, char const *argv[]) { int n, m; // Enter number of vertices and edges cin>>n>>m; int h_vertexArray[n]; int h_neighbourArray[m]; int h_degreeCount[n]; int h_detectConflict[n]; // Cuda memory allocation size_t bytes = n*sizeof(int); int *d_vertexArray = NULL; hipMalloc((void **)&d_vertexArray, bytes); int *d_neighbourArray = NULL; hipMalloc((void **)&d_neighbourArray, m*sizeof(int)); int *d_detectConflict = NULL; hipMalloc((void **)&d_detectConflict, bytes); hipMemset((void *)d_detectConflict, 0,bytes); int *d_degreeCount = NULL; hipMalloc((void **)&d_degreeCount, bytes); hipMemset((void *)d_degreeCount, 0, bytes); hiprandState* partition_states; hipMalloc ( &partition_states, n*sizeof( hiprandState ) ); for (int i = 0; i < n; ++i) { /* code */ h_vertexArray[i]=m; } int temp = 0; int current = 0; int mark = 1; // Add the graph based on input file for (int i = 0; i < m; ++i) { /* code */ int incoming; int end; cin>>incoming>>end; incoming++; end++; if (incoming!=mark){ if (incoming == mark+1 && h_vertexArray[mark-1]!=m){ } else{ for (int j = mark; j<incoming; j++){ h_vertexArray[j-1]=temp; } } mark = incoming; } if (incoming==current){ h_neighbourArray[temp]=end; temp++; } else { current = incoming; h_vertexArray[current-1]=temp; h_neighbourArray[temp]=end; temp++; } } hipMemcpy(d_vertexArray, h_vertexArray, n*sizeof(int), hipMemcpyHostToDevice); hipMemcpy(d_neighbourArray, h_neighbourArray, m*sizeof(int), hipMemcpyHostToDevice); int threadsPerBlock = 512; int blocksPerGrid = (n + threadsPerBlock -1)/threadsPerBlock; struct timeval startTime; struct timeval endTime; struct timezone startZone; struct timezone endZone; long startt,endt; double overhead; cout<<threadsPerBlock<<" "<<blocksPerGrid<<endl; gettimeofday(&startTime,&startZone); // Step 0 : Calculate degree of each vertex degreeCalc<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m); thrust::device_ptr<int> d_ptr = thrust::device_pointer_cast(d_degreeCount); int max = *(thrust::max_element(d_ptr, d_ptr + n)); cout<<"Max number of colours = "<<max<<endl; partition_step <<<blocksPerGrid, threadsPerBlock>>> ( partition_states, time(NULL) ); // Step 1 - Randomly assign colours randomColouring<<<blocksPerGrid, threadsPerBlock>>>(partition_states, d_degreeCount, n, max); hipMemcpy(h_degreeCount, d_degreeCount, n*sizeof(int), hipMemcpyDeviceToHost); cout<<"randomColouring"<<endl; for (int i=0; i<n; i++){ cout<<"Color of"<<i<<": "<<h_degreeCount[i]<<endl; } cout<<endl; conflictDetection<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m, d_detectConflict); thrust::device_ptr<int> d_detectConflict_ptr = thrust::device_pointer_cast(d_detectConflict); int count1 = thrust::reduce(d_detectConflict_ptr, d_detectConflict_ptr + n); hipMemcpy(h_detectConflict, d_detectConflict, n*sizeof(int), hipMemcpyDeviceToHost); int countnew=0; int old_colors[n]; for (int i = 0; i < n; ++i) { /* code */ old_colors[i] = -1; } for (int i=0; i<n-1; i++){ if (h_detectConflict[i]==0){ continue; } countnew++; bool usedColours[n]; fill(usedColours, usedColours+n, false); int incoming = -1, stop = -1; incoming = h_vertexArray[i]; stop = h_vertexArray[i+1]; old_colors[i] = h_degreeCount[i]; for (int j=incoming; j<stop; j++){ usedColours[h_degreeCount[h_neighbourArray[j]-1]-1] = true; } for (int j=0; j<n; j++){ if (usedColours[j]==false){ h_degreeCount[i]=j+1; break; } } } if (h_detectConflict[n-1]!=0){ bool usedColours[n]; countnew++; fill(usedColours, usedColours+n, false); int incoming = -1, stop = -1; incoming = h_vertexArray[n-1]; stop = m; for (int j=incoming; j<stop; j++){ usedColours[h_degreeCount[h_neighbourArray[j]-1]-1] = true; } for (int j=0; j<n; j++){ if (usedColours[j]==false){ h_degreeCount[n-1]=j+1; break; } } } for (int i = 0; i < n; ++i) { cout<<"Colour of i from" <<i <<" "<<old_colors[i]<<":"<<h_degreeCount[i]<<endl; } hipMemset((void *)d_detectConflict, 0, (n)*sizeof(int)); hipMemcpy(d_degreeCount, h_degreeCount, n*sizeof(int), hipMemcpyHostToDevice); conflictDetection<<<blocksPerGrid, threadsPerBlock>>>(d_vertexArray, d_neighbourArray, d_degreeCount, n, m, d_detectConflict); gettimeofday(&endTime,&endZone); startt = startTime.tv_sec*1000000+startTime.tv_usec; endt = endTime.tv_sec*1000000+endTime.tv_usec; overhead = (endt-startt)/1000000.0; count1 = thrust::reduce(d_detectConflict_ptr, d_detectConflict_ptr + n); cout<<"Count: "<<count1<<" "<<countnew<<endl; cout<<"time taken is"<<overhead<<endl; hipFree(d_neighbourArray); hipFree(d_vertexArray); hipFree(d_degreeCount); hipFree(d_detectConflict); hipDeviceReset(); return 0; }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
.file "tmpxft_0007addc_00000000-6_kernels.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2037: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2037: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
#include <hip/hip_runtime.h> template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
.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> template<typename T> __global__ void init_array_kernel(T *gpu_array, size_t n) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { new (&gpu_array[i]) T(); } } template<class T, class TransformedT, class TransformationT> __global__ void transform_kernel( T *from_array, size_t n, TransformedT *to_array, TransformationT transform) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { to_array[i] = transform(from_array[i], i); } } template<typename T, typename Transformation> __global__ void for_each_kernel(T *gpu_array, size_t n, Transformation fn) { for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; i += blockDim.x * gridDim.x) { fn(gpu_array[i], i); } } template<typename T, typename Reduction> __global__ void reduce_2step_kernel( T *gpu_array, size_t n, T *out, Reduction fn, T initial_value=T{}) { /* Log-reduction based from the one in the book "The CUDA Handbook" by Nicholas Wilt. */ extern __shared__ T partials[]; const int tid = threadIdx.x; auto reduced = initial_value; for (int i = blockIdx.x * blockDim.x + tid; i < n; i += blockDim.x * gridDim.x) { reduced = fn(reduced, gpu_array[i]); } partials[tid] = reduced; __syncthreads(); for (int active_threads = blockDim.x / 2; active_threads > 0; active_threads /= 2) { auto is_active_thread = tid < active_threads; if (is_active_thread) { partials[tid] = fn(partials[tid], partials[tid + active_threads]); } __syncthreads(); } if (tid == 0) { out[blockIdx.x] = partials[0]; } } /* ----------------------------------------------------------------------- These kernels are used and tested on the gpu_array and gpu_object classes. */
.text .file "kernels.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_0007addc_00000000-6_kernels.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2037: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2037: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "kernels.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 "includes.h" __global__ void sync_conv_groups() { }
code for sm_80 Function : _Z16sync_conv_groupsv .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void sync_conv_groups() { }
.file "tmpxft_000724d1_00000000-6_sync_conv_groups.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__Z16sync_conv_groupsvv .type _Z35__device_stub__Z16sync_conv_groupsvv, @function _Z35__device_stub__Z16sync_conv_groupsvv: .LFB2051: .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 _Z16sync_conv_groupsv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z35__device_stub__Z16sync_conv_groupsvv, .-_Z35__device_stub__Z16sync_conv_groupsvv .globl _Z16sync_conv_groupsv .type _Z16sync_conv_groupsv, @function _Z16sync_conv_groupsv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z16sync_conv_groupsvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z16sync_conv_groupsv, .-_Z16sync_conv_groupsv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z16sync_conv_groupsv" .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 _Z16sync_conv_groupsv(%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 sync_conv_groups() { }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void sync_conv_groups() { }
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 sync_conv_groups() { }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z16sync_conv_groupsv .globl _Z16sync_conv_groupsv .p2align 8 .type _Z16sync_conv_groupsv,@function _Z16sync_conv_groupsv: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16sync_conv_groupsv .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z16sync_conv_groupsv, .Lfunc_end0-_Z16sync_conv_groupsv .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z16sync_conv_groupsv .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z16sync_conv_groupsv.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void sync_conv_groups() { }
.text .file "sync_conv_groups.hip" .globl _Z31__device_stub__sync_conv_groupsv # -- Begin function _Z31__device_stub__sync_conv_groupsv .p2align 4, 0x90 .type _Z31__device_stub__sync_conv_groupsv,@function _Z31__device_stub__sync_conv_groupsv: # @_Z31__device_stub__sync_conv_groupsv .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 $_Z16sync_conv_groupsv, %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 _Z31__device_stub__sync_conv_groupsv, .Lfunc_end0-_Z31__device_stub__sync_conv_groupsv .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 $_Z16sync_conv_groupsv, %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 _Z16sync_conv_groupsv,@object # @_Z16sync_conv_groupsv .section .rodata,"a",@progbits .globl _Z16sync_conv_groupsv .p2align 3, 0x0 _Z16sync_conv_groupsv: .quad _Z31__device_stub__sync_conv_groupsv .size _Z16sync_conv_groupsv, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z16sync_conv_groupsv" .size .L__unnamed_1, 22 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__sync_conv_groupsv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16sync_conv_groupsv .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 : _Z16sync_conv_groupsv .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0020*/ BRA 0x20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0030*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0040*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0050*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0060*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0070*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0080*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0090*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z16sync_conv_groupsv .globl _Z16sync_conv_groupsv .p2align 8 .type _Z16sync_conv_groupsv,@function _Z16sync_conv_groupsv: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16sync_conv_groupsv .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 0 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 0 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 1 .amdhsa_next_free_sgpr 1 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z16sync_conv_groupsv, .Lfunc_end0-_Z16sync_conv_groupsv .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: [] .group_segment_fixed_size: 0 .kernarg_segment_align: 4 .kernarg_segment_size: 0 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z16sync_conv_groupsv .private_segment_fixed_size: 0 .sgpr_count: 0 .sgpr_spill_count: 0 .symbol: _Z16sync_conv_groupsv.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 0 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000724d1_00000000-6_sync_conv_groups.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__Z16sync_conv_groupsvv .type _Z35__device_stub__Z16sync_conv_groupsvv, @function _Z35__device_stub__Z16sync_conv_groupsvv: .LFB2051: .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 _Z16sync_conv_groupsv(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 96 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z35__device_stub__Z16sync_conv_groupsvv, .-_Z35__device_stub__Z16sync_conv_groupsvv .globl _Z16sync_conv_groupsv .type _Z16sync_conv_groupsv, @function _Z16sync_conv_groupsv: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z35__device_stub__Z16sync_conv_groupsvv addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z16sync_conv_groupsv, .-_Z16sync_conv_groupsv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z16sync_conv_groupsv" .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 _Z16sync_conv_groupsv(%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 "sync_conv_groups.hip" .globl _Z31__device_stub__sync_conv_groupsv # -- Begin function _Z31__device_stub__sync_conv_groupsv .p2align 4, 0x90 .type _Z31__device_stub__sync_conv_groupsv,@function _Z31__device_stub__sync_conv_groupsv: # @_Z31__device_stub__sync_conv_groupsv .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 $_Z16sync_conv_groupsv, %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 _Z31__device_stub__sync_conv_groupsv, .Lfunc_end0-_Z31__device_stub__sync_conv_groupsv .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 $_Z16sync_conv_groupsv, %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 _Z16sync_conv_groupsv,@object # @_Z16sync_conv_groupsv .section .rodata,"a",@progbits .globl _Z16sync_conv_groupsv .p2align 3, 0x0 _Z16sync_conv_groupsv: .quad _Z31__device_stub__sync_conv_groupsv .size _Z16sync_conv_groupsv, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z16sync_conv_groupsv" .size .L__unnamed_1, 22 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__sync_conv_groupsv .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16sync_conv_groupsv .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 <cstdio> #include <math.h> #include <stdlib.h> #include <time.h> #include <cuda_runtime_api.h> #define restrict __restrict__ using namespace std; int error(const char *msg) { fprintf(stderr, "%s\n", msg); exit(1); } void cuda_check(cudaError_t err, const char *msg) { if (err != cudaSuccess) { fprintf(stderr, "%s: errore %d - %s\n", msg, err, cudaGetErrorString(err)); exit(1); } } //inizializzazione su CPU con numeri random void init_random(int vett[],int nels,int max) { srand(time(NULL)); for(int i=0;i<nels;++i) vett[i]=rand()% max+ 1; } //verifica con numeri random bool verify_random(const int* scan_out, int nels) { for(int i=0;i<nels-1;++i) if(scan_out[i]>scan_out[i+1]) { fprintf(stderr, "errore tra le posizioni %d e %d \n", i,i+1); return false; } return true; } //verifica con numeri ordinati al contrario partendo da nels fino ad arrivare ad 1 bool verify(const int* scan_out, int nels) { int err=0; for (int i = 0; i < nels; ++i) { if(i+1!=scan_out[i]) { fprintf(stderr, "verify,idx=%d: val_scan:%d \n", i,scan_out[i]); err=1; } } if(err) return false; return true; } //inizializzazione su GPU __global__ void init(int *vec, int nels) { int idx = threadIdx.x + blockDim.x*blockIdx.x; if (idx < nels) vec[idx] = nels-idx; } extern __shared__ int4 shared[]; __device__ void scan_delle_code(int4 coda) { __syncthreads(); shared[threadIdx.x] = coda; for (int offset = 1; offset < blockDim.x; offset *= 2) { __syncthreads(); if (threadIdx.x >= offset) { coda.x += shared[threadIdx.x - offset].x; coda.y += shared[threadIdx.x - offset].y; coda.z += shared[threadIdx.x - offset].z; coda.w += shared[threadIdx.x - offset].w; } __syncthreads(); shared[threadIdx.x] = coda; } __syncthreads(); } //primo scan __global__ void scan_step1(int4 * restrict out0,int4 * restrict out1,int4 * restrict out2,int4 * restrict out3, const int4 * restrict in, int nels /* numero di quartine */, int * restrict code0,int * restrict code1,int * restrict code2,int * restrict code3, int nbit) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 val,val0,val1,val2,val3; int4 correzione_dal_blocco_precedente = make_int4(0,0,0,0); int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { val = (idx < elemento_limite ? in[idx] : make_int4(0, 0, 0, 0)); val0= make_int4(0,0,0,0); val1= make_int4(0,0,0,0); val2= make_int4(0,0,0,0); val3= make_int4(0,0,0,0); //basta fare solo 3 controlli sui bit,il quarto cofronto (00 per comodità) è complementare! //controllo sul primo valore della quartina (x) if(((val.x>>nbit)&3)==1) val1.x=1; else if(((val.x>>nbit)&3)==2) val2.x=1; else if(((val.x>>nbit)&3)==3) val3.x=1; else val0.x=1; //controllo sulla seconda componente della quartina (y) if(((val.y>>nbit)&3)==1) val1.y=1; else if(((val.y>>nbit)&3)==2) val2.y=1; else if(((val.y>>nbit)&3)==3) val3.y=1; else val0.y=1; //controllo sulla terza componente della quartina (z) if(((val.z>>nbit)&3)==1) val1.z=1; else if(((val.z>>nbit)&3)==2) val2.z=1; else if(((val.z>>nbit)&3)==3) val3.z=1; else val0.z=1; //controllo sulla quarta componente della quartina (w) if(((val.w>>nbit)&3)==1) val1.w=1; else if(((val.w>>nbit)&3)==2) val2.w=1; else if(((val.w>>nbit)&3)==3) val3.w=1; else val0.w=1; /* scan delle componenti dei val */ val0.y += val0.x; val0.z += val0.y; val0.w += val0.z; val1.y += val1.x; val1.z += val1.y; val1.w += val1.z; val2.y += val2.x; val2.z += val2.y; val2.w += val2.z; val3.y += val3.x; val3.z += val3.y; val3.w += val3.z; int4 coda=make_int4(val0.w,val1.w,val2.w,val3.w); scan_delle_code(coda); int4 correzione_dai_thread_precedenti = make_int4(0,0,0,0); if (threadIdx.x > 0) correzione_dai_thread_precedenti = shared[threadIdx.x-1]; int correzione_totale = correzione_dal_blocco_precedente.x + correzione_dai_thread_precedenti.x; val0.x += correzione_totale; val0.y += correzione_totale; val0.z += correzione_totale; val0.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.y + correzione_dai_thread_precedenti.y; val1.x += correzione_totale; val1.y += correzione_totale; val1.z += correzione_totale; val1.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.z + correzione_dai_thread_precedenti.z; val2.x += correzione_totale; val2.y += correzione_totale; val2.z += correzione_totale; val2.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.w + correzione_dai_thread_precedenti.w; val3.x += correzione_totale; val3.y += correzione_totale; val3.z += correzione_totale; val3.w += correzione_totale; correzione_dal_blocco_precedente.x += shared[blockDim.x-1].x; //correzione di 00 (0) correzione_dal_blocco_precedente.y += shared[blockDim.x-1].y; //correzione di 01 (1) correzione_dal_blocco_precedente.z += shared[blockDim.x-1].z; //correzione di 10 (2) correzione_dal_blocco_precedente.w += shared[blockDim.x-1].w; //correzione di 11 (3) if (idx < nels) { out0[idx] = val0; out1[idx] = val1; out2[idx] = val2; out3[idx] = val3; } } if (gridDim.x > 1 && threadIdx.x == blockDim.x - 1) { code0[blockIdx.x] = val0.w; code1[blockIdx.x] = val1.w; code2[blockIdx.x] = val2.w; code3[blockIdx.x] = val3.w; } } //secondo scan fatto sulo sui vettori di code __global__ void scan_step2(int4 * restrict code0, int4 * restrict code1, int4 * restrict code2, int4 * restrict code3, int nels /* numero di quartine */) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 val0,val1,val2,val3; int4 correzione_dal_blocco_precedente = make_int4(0,0,0,0); int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { val0 = (idx < elemento_limite ? code0[idx] : make_int4(0, 0, 0, 0)); val1 = (idx < elemento_limite ? code1[idx] : make_int4(0, 0, 0, 0)); val2 = (idx < elemento_limite ? code2[idx] : make_int4(0, 0, 0, 0)); val3 = (idx < elemento_limite ? code3[idx] : make_int4(0, 0, 0, 0)); /* scan delle componenti di val */ val0.y += val0.x; val0.z += val0.y; val0.w += val0.z; val1.y += val1.x; val1.z += val1.y; val1.w += val1.z; val2.y += val2.x; val2.z += val2.y; val2.w += val2.z; val3.y += val3.x; val3.z += val3.y; val3.w += val3.z; //da modificare anche scan_delle_code int4 coda=make_int4(val0.w,val1.w,val2.w,val3.w); scan_delle_code(coda); int4 correzione_dai_thread_precedenti = make_int4(0,0,0,0); if (threadIdx.x > 0) correzione_dai_thread_precedenti = shared[threadIdx.x-1]; int correzione_totale = correzione_dal_blocco_precedente.x + correzione_dai_thread_precedenti.x; val0.x += correzione_totale; val0.y += correzione_totale; val0.z += correzione_totale; val0.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.y + correzione_dai_thread_precedenti.y; val1.x += correzione_totale; val1.y += correzione_totale; val1.z += correzione_totale; val1.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.z + correzione_dai_thread_precedenti.z; val2.x += correzione_totale; val2.y += correzione_totale; val2.z += correzione_totale; val2.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.w + correzione_dai_thread_precedenti.w; val3.x += correzione_totale; val3.y += correzione_totale; val3.z += correzione_totale; val3.w += correzione_totale; correzione_dal_blocco_precedente.x += shared[blockDim.x-1].x; //correzione di 00 (0) correzione_dal_blocco_precedente.y += shared[blockDim.x-1].y; //correzione di 01 (1) correzione_dal_blocco_precedente.z += shared[blockDim.x-1].z; //correzione di 10 (2) correzione_dal_blocco_precedente.w += shared[blockDim.x-1].w; //correzione di 11 (3) if (idx < nels) { code0[idx] = val0; code1[idx] = val1; code2[idx] = val2; code3[idx] = val3; } } } __global__ void fixup(int4 * restrict scan0,int4 * restrict scan1,int4 * restrict scan2,int4 * restrict scan3,int nels ,const int * restrict code0,const int * restrict code1,const int * restrict code2,const int * restrict code3,const int4* restrict in,int nbit,int4* max) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 correzione_dal_blocco_precedente =make_int4(0,0,0,0); if(blockIdx.x>0) { correzione_dal_blocco_precedente.x = code0[blockIdx.x - 1]; correzione_dal_blocco_precedente.y = code1[blockIdx.x - 1]; correzione_dal_blocco_precedente.z = code2[blockIdx.x - 1]; correzione_dal_blocco_precedente.w = code3[blockIdx.x - 1]; } int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { if (idx < elemento_limite) { int4 val0 = scan0[idx], val1=scan1[idx], val2=scan2[idx], val3=scan3[idx]; int4 val_in=in[idx]; if(idx==nels-1) { //salvataggio in memoria globale degli ultimi elementi degli scan inclusivi (*max).x=val0.w+correzione_dal_blocco_precedente.x; (*max).y=val1.w+correzione_dal_blocco_precedente.y; (*max).z=val2.w+correzione_dal_blocco_precedente.z; (*max).w=val3.w+correzione_dal_blocco_precedente.w; } //trasformazione degli scan da inclusivi ad esclusivi val0.x += correzione_dal_blocco_precedente.x - ((((val_in.x>>nbit)&3)==0)?1:0); val0.y += correzione_dal_blocco_precedente.x - ((((val_in.y>>nbit)&3)==0)?1:0); val0.z += correzione_dal_blocco_precedente.x - ((((val_in.z>>nbit)&3)==0)?1:0); val0.w += correzione_dal_blocco_precedente.x - ((((val_in.w>>nbit)&3)==0)?1:0); val1.x += correzione_dal_blocco_precedente.y - ((((val_in.x>>nbit)&3)==1)?1:0); val1.y += correzione_dal_blocco_precedente.y - ((((val_in.y>>nbit)&3)==1)?1:0); val1.z += correzione_dal_blocco_precedente.y - ((((val_in.z>>nbit)&3)==1)?1:0); val1.w += correzione_dal_blocco_precedente.y - ((((val_in.w>>nbit)&3)==1)?1:0); val2.x += correzione_dal_blocco_precedente.z - ((((val_in.x>>nbit)&3)==2)?1:0); val2.y += correzione_dal_blocco_precedente.z - ((((val_in.y>>nbit)&3)==2)?1:0); val2.z += correzione_dal_blocco_precedente.z - ((((val_in.z>>nbit)&3)==2)?1:0); val2.w += correzione_dal_blocco_precedente.z - ((((val_in.w>>nbit)&3)==2)?1:0); val3.x += correzione_dal_blocco_precedente.w - ((((val_in.x>>nbit)&3)==3)?1:0); val3.y += correzione_dal_blocco_precedente.w - ((((val_in.y>>nbit)&3)==3)?1:0); val3.z += correzione_dal_blocco_precedente.w - ((((val_in.z>>nbit)&3)==3)?1:0); val3.w += correzione_dal_blocco_precedente.w - ((((val_in.w>>nbit)&3)==3)?1:0); scan0[idx] = val0; scan1[idx] = val1; scan2[idx] = val2; scan3[idx] = val3; } } } //kernel adibito al riordino dei vettori utilizzando 2 bit __global__ void reorder(const int4* restrict scan0,const int4* restrict scan1,const int4* restrict scan2,const int4* restrict scan3,const int4* restrict in, int* restrict out,int nels,int4* max) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int numero_cicli = (els_per_sezione +blockDim.x -1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); int4 offset_max=make_int4((*max).x,(*max).y,(*max).z,(*max).w); int4 val_scan_succ; //scan dei max int4 max_scan=make_int4(offset_max.x,offset_max.y,offset_max.z,offset_max.w); offset_max.x=0; //offset di 00 (0) offset_max.y=max_scan.x; //offset di 01 (1) offset_max.z=offset_max.y + max_scan.y; //offset di 10 (2) offset_max.w=offset_max.z + max_scan.z; //offset di 11 (3) //inizio ciclo di reorder for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { if (idx < elemento_limite) { int4 val_num=in[idx], val_scan0=scan0[idx], val_scan1=scan1[idx], val_scan2=scan2[idx], val_scan3=scan3[idx]; // confronto 1° elemento con 2° elemento della quartina if(val_scan0.x!=val_scan0.y) out[val_scan0.x+offset_max.x]=val_num.x; else if(val_scan1.x!=val_scan1.y) out[val_scan1.x+offset_max.y]=val_num.x; else if(val_scan2.x!=val_scan2.y) out[val_scan2.x+offset_max.z]=val_num.x; else out[val_scan3.x+offset_max.w]=val_num.x; // confronto 2° elemento con 3° elemento della quartina if(val_scan0.y!=val_scan0.z) out[val_scan0.y+offset_max.x]=val_num.y; else if(val_scan1.y!=val_scan1.z) out[val_scan1.y+offset_max.y]=val_num.y; else if(val_scan2.y!=val_scan2.z) out[val_scan2.y+offset_max.z]=val_num.y; else out[val_scan3.y+offset_max.w]=val_num.y; // confronto 3° elemento con 4° elemento della quartina if(val_scan0.z!=val_scan0.w) out[val_scan0.z+offset_max.x]=val_num.z; else if(val_scan1.z!=val_scan1.w) out[val_scan1.z+offset_max.y]=val_num.z; else if(val_scan2.z!=val_scan2.w) out[val_scan2.z+offset_max.z]=val_num.z; else out[val_scan3.z+offset_max.w]=val_num.z; //confronto 4° elemento con 1° elemento della quartina dello scan successivo if(idx!=nels-1) { //scan3[idx+1] puo non essere preso, visto che non viene mai usato val_scan_succ=make_int4(scan0[idx+1].x,scan1[idx+1].x,scan2[idx+1].x,0); // primi valori delle quartine successive degli scan if(val_scan0.w!=val_scan_succ.x) out[val_scan0.w + offset_max.x]=val_num.w; else if(val_scan1.w!=val_scan_succ.y) out[val_scan1.w + offset_max.y]=val_num.w; else if(val_scan2.w!=val_scan_succ.z) out[val_scan2.w + offset_max.z]=val_num.w; else out[val_scan3.w + offset_max.w]=val_num.w; } else { if(val_scan0.w!=max_scan.x) out[val_scan0.w + offset_max.x]=val_num.w; else if(val_scan1.w!=max_scan.y) out[val_scan1.w + offset_max.y]=val_num.w; else if(val_scan2.w!=max_scan.z) out[val_scan2.w + offset_max.z]=val_num.w; else out[val_scan3.w + offset_max.w]=val_num.w; } } } } int main(int argc, char *argv[]) { if (argc < 4) error("sintassi radix_sort: numels thread_per_blocco numero_blocchi_scan valore_massimo"); int nels = atoi(argv[1]); /* numero di elementi */ if (nels <= 0) error("il numero di elementi deve essere positivo"); if (nels & 3) error("il numero di elementi deve essere multiplo di 4"); int numThreads = atoi(argv[2]); /* local work size */ if (numThreads <= 0) error("il numero di thread per blocco deve essere positivo"); int numBlocksScan = atoi(argv[3]); /* numero blocchi scan */ if (numBlocksScan <= 0) error("il numero di blocchi deve essere positivo"); int numMax = atoi(argv[4]); /* numero blocchi scan */ if (numMax <= 0) error("il valore massimo deve essere positivo"); //inizializzazione dei vettori const size_t memsize = sizeof(int)*nels; int4 *d_v1, *d_scan0,*d_scan1,*d_scan2,*d_scan3, *d_code0,*d_code1,*d_code2,*d_code3,*d_out,*tmp; int numbit; int4 *d_max; //calolo dei cicli da fare avendo il massimo int cicli=int(log(numMax)/log(2)) + 1; printf("numero cicli da fare=%d\n",cicli/2); //allocazione dei vettori su GPU cudaError_t err = cudaMalloc(&d_v1, memsize); cuda_check(err, "malloc v1"); err= cudaMalloc(&d_max,sizeof(int4)); cuda_check(err,"malloc max"); err = cudaMalloc(&d_scan0, memsize); cuda_check(err, "malloc scan0"); err = cudaMalloc(&d_scan1, memsize); cuda_check(err, "malloc scan1"); err = cudaMalloc(&d_scan2, memsize); cuda_check(err, "malloc scan2"); err = cudaMalloc(&d_scan3, memsize); cuda_check(err, "malloc scan3"); err = cudaMalloc(&d_code0, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code0"); err = cudaMalloc(&d_code1, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code1"); err = cudaMalloc(&d_code2, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code2"); err = cudaMalloc(&d_code3, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code3"); err = cudaMalloc(&d_out, memsize); cuda_check(err, "malloc out"); //allocazione su CPU int *vout = (int*)malloc(memsize); if (!vout) error("alloc vscan"); //inizializzazione su CPU di numeri random con massimo possibile numMax init_random(vout,nels,numMax); /* //inizializzazione su GPU di numeri decrescenti a partire da nels int numBlocks = (nels + numThreads - 1)/numThreads; init<<<numBlocks, numThreads>>>((int*)d_v1, nels); */ //prova ad otimizzare la cache cudaFuncSetCacheConfig(scan_step1, cudaFuncCachePreferL1); cudaFuncSetCacheConfig(scan_step2, cudaFuncCachePreferL1); cudaFuncSetCacheConfig(fixup, cudaFuncCachePreferL1); cudaFuncSetCacheConfig(reorder, cudaFuncCachePreferL1); err = cudaMemcpy(d_v1,vout,memsize, cudaMemcpyHostToDevice); cuda_check(err, "memcpy vett su GPU"); //creazione eventi cudaEvent_t before_scan, after_scan; err = cudaEventCreate(&before_scan); cuda_check(err, "create event before"); err = cudaEventCreate(&after_scan); cuda_check(err, "create event after"); cudaEventRecord(before_scan); for(numbit=0;numbit<cicli;numbit+=2) { //pulizia dei vettori di code err= cudaMemset (d_code0,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset0"); err= cudaMemset (d_code1,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset1"); err= cudaMemset (d_code2,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset2"); err= cudaMemset (d_code3,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset3"); scan_step1<<<numBlocksScan, numThreads, numThreads*sizeof(int)*4>>> (d_scan0,d_scan1,d_scan2,d_scan3, d_v1, nels/4, (int*)d_code0,(int*)d_code1,(int*)d_code2,(int*)d_code3, numbit); scan_step2<<<1, numThreads, numThreads*sizeof(int)*4>>> (d_code0, d_code1,d_code2,d_code3, numBlocksScan/4); fixup<<<numBlocksScan, numThreads>>>(d_scan0,d_scan1,d_scan2,d_scan3, nels/4, (int*)d_code0,(int*)d_code1,(int*)d_code2,(int*)d_code3,d_v1,numbit,d_max); reorder<<<numBlocksScan,numThreads>>>(d_scan0,d_scan1,d_scan2,d_scan3,d_v1,(int*)d_out,nels/4,d_max); //scambio deii puntatori d_out e d_v1 if(numbit+2 <cicli) { tmp=d_v1; d_v1=d_out; d_out=tmp; } } cudaEventRecord(after_scan); err = cudaEventSynchronize(after_scan); cuda_check(err, "after scan sznc"); float runtime_ms; cudaEventElapsedTime(&runtime_ms, before_scan, after_scan); printf("scan runtime: %.4g ms\n", runtime_ms); err = cudaMemcpy(vout, d_out, memsize, cudaMemcpyDeviceToHost); cuda_check(err, "memcpy"); printf("\n\n"); verify_random(vout,nels)?printf("Ordinamento riuscito!\n"):printf("Ordinamento non riuscito!\n"); if(nels <=32) for(int i=0;i<nels;++i) printf("%d ",vout[i]); printf("\n"); }
#include <hip/hip_runtime.h> #include <cstdio> #include <math.h> #include <stdlib.h> #include <time.h> #include <hip/hip_runtime_api.h> #define restrict __restrict__ using namespace std; int error(const char *msg) { fprintf(stderr, "%s\n", msg); exit(1); } void cuda_check(hipError_t err, const char *msg) { if (err != hipSuccess) { fprintf(stderr, "%s: errore %d - %s\n", msg, err, hipGetErrorString(err)); exit(1); } } //inizializzazione su CPU con numeri random void init_random(int vett[],int nels,int max) { srand(time(NULL)); for(int i=0;i<nels;++i) vett[i]=rand()% max+ 1; } //verifica con numeri random bool verify_random(const int* scan_out, int nels) { for(int i=0;i<nels-1;++i) if(scan_out[i]>scan_out[i+1]) { fprintf(stderr, "errore tra le posizioni %d e %d \n", i,i+1); return false; } return true; } //verifica con numeri ordinati al contrario partendo da nels fino ad arrivare ad 1 bool verify(const int* scan_out, int nels) { int err=0; for (int i = 0; i < nels; ++i) { if(i+1!=scan_out[i]) { fprintf(stderr, "verify,idx=%d: val_scan:%d \n", i,scan_out[i]); err=1; } } if(err) return false; return true; } //inizializzazione su GPU __global__ void init(int *vec, int nels) { int idx = threadIdx.x + blockDim.x*blockIdx.x; if (idx < nels) vec[idx] = nels-idx; } extern __shared__ int4 shared[]; __device__ void scan_delle_code(int4 coda) { __syncthreads(); shared[threadIdx.x] = coda; for (int offset = 1; offset < blockDim.x; offset *= 2) { __syncthreads(); if (threadIdx.x >= offset) { coda.x += shared[threadIdx.x - offset].x; coda.y += shared[threadIdx.x - offset].y; coda.z += shared[threadIdx.x - offset].z; coda.w += shared[threadIdx.x - offset].w; } __syncthreads(); shared[threadIdx.x] = coda; } __syncthreads(); } //primo scan __global__ void scan_step1(int4 * restrict out0,int4 * restrict out1,int4 * restrict out2,int4 * restrict out3, const int4 * restrict in, int nels /* numero di quartine */, int * restrict code0,int * restrict code1,int * restrict code2,int * restrict code3, int nbit) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 val,val0,val1,val2,val3; int4 correzione_dal_blocco_precedente = make_int4(0,0,0,0); int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { val = (idx < elemento_limite ? in[idx] : make_int4(0, 0, 0, 0)); val0= make_int4(0,0,0,0); val1= make_int4(0,0,0,0); val2= make_int4(0,0,0,0); val3= make_int4(0,0,0,0); //basta fare solo 3 controlli sui bit,il quarto cofronto (00 per comodità) è complementare! //controllo sul primo valore della quartina (x) if(((val.x>>nbit)&3)==1) val1.x=1; else if(((val.x>>nbit)&3)==2) val2.x=1; else if(((val.x>>nbit)&3)==3) val3.x=1; else val0.x=1; //controllo sulla seconda componente della quartina (y) if(((val.y>>nbit)&3)==1) val1.y=1; else if(((val.y>>nbit)&3)==2) val2.y=1; else if(((val.y>>nbit)&3)==3) val3.y=1; else val0.y=1; //controllo sulla terza componente della quartina (z) if(((val.z>>nbit)&3)==1) val1.z=1; else if(((val.z>>nbit)&3)==2) val2.z=1; else if(((val.z>>nbit)&3)==3) val3.z=1; else val0.z=1; //controllo sulla quarta componente della quartina (w) if(((val.w>>nbit)&3)==1) val1.w=1; else if(((val.w>>nbit)&3)==2) val2.w=1; else if(((val.w>>nbit)&3)==3) val3.w=1; else val0.w=1; /* scan delle componenti dei val */ val0.y += val0.x; val0.z += val0.y; val0.w += val0.z; val1.y += val1.x; val1.z += val1.y; val1.w += val1.z; val2.y += val2.x; val2.z += val2.y; val2.w += val2.z; val3.y += val3.x; val3.z += val3.y; val3.w += val3.z; int4 coda=make_int4(val0.w,val1.w,val2.w,val3.w); scan_delle_code(coda); int4 correzione_dai_thread_precedenti = make_int4(0,0,0,0); if (threadIdx.x > 0) correzione_dai_thread_precedenti = shared[threadIdx.x-1]; int correzione_totale = correzione_dal_blocco_precedente.x + correzione_dai_thread_precedenti.x; val0.x += correzione_totale; val0.y += correzione_totale; val0.z += correzione_totale; val0.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.y + correzione_dai_thread_precedenti.y; val1.x += correzione_totale; val1.y += correzione_totale; val1.z += correzione_totale; val1.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.z + correzione_dai_thread_precedenti.z; val2.x += correzione_totale; val2.y += correzione_totale; val2.z += correzione_totale; val2.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.w + correzione_dai_thread_precedenti.w; val3.x += correzione_totale; val3.y += correzione_totale; val3.z += correzione_totale; val3.w += correzione_totale; correzione_dal_blocco_precedente.x += shared[blockDim.x-1].x; //correzione di 00 (0) correzione_dal_blocco_precedente.y += shared[blockDim.x-1].y; //correzione di 01 (1) correzione_dal_blocco_precedente.z += shared[blockDim.x-1].z; //correzione di 10 (2) correzione_dal_blocco_precedente.w += shared[blockDim.x-1].w; //correzione di 11 (3) if (idx < nels) { out0[idx] = val0; out1[idx] = val1; out2[idx] = val2; out3[idx] = val3; } } if (gridDim.x > 1 && threadIdx.x == blockDim.x - 1) { code0[blockIdx.x] = val0.w; code1[blockIdx.x] = val1.w; code2[blockIdx.x] = val2.w; code3[blockIdx.x] = val3.w; } } //secondo scan fatto sulo sui vettori di code __global__ void scan_step2(int4 * restrict code0, int4 * restrict code1, int4 * restrict code2, int4 * restrict code3, int nels /* numero di quartine */) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 val0,val1,val2,val3; int4 correzione_dal_blocco_precedente = make_int4(0,0,0,0); int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { val0 = (idx < elemento_limite ? code0[idx] : make_int4(0, 0, 0, 0)); val1 = (idx < elemento_limite ? code1[idx] : make_int4(0, 0, 0, 0)); val2 = (idx < elemento_limite ? code2[idx] : make_int4(0, 0, 0, 0)); val3 = (idx < elemento_limite ? code3[idx] : make_int4(0, 0, 0, 0)); /* scan delle componenti di val */ val0.y += val0.x; val0.z += val0.y; val0.w += val0.z; val1.y += val1.x; val1.z += val1.y; val1.w += val1.z; val2.y += val2.x; val2.z += val2.y; val2.w += val2.z; val3.y += val3.x; val3.z += val3.y; val3.w += val3.z; //da modificare anche scan_delle_code int4 coda=make_int4(val0.w,val1.w,val2.w,val3.w); scan_delle_code(coda); int4 correzione_dai_thread_precedenti = make_int4(0,0,0,0); if (threadIdx.x > 0) correzione_dai_thread_precedenti = shared[threadIdx.x-1]; int correzione_totale = correzione_dal_blocco_precedente.x + correzione_dai_thread_precedenti.x; val0.x += correzione_totale; val0.y += correzione_totale; val0.z += correzione_totale; val0.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.y + correzione_dai_thread_precedenti.y; val1.x += correzione_totale; val1.y += correzione_totale; val1.z += correzione_totale; val1.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.z + correzione_dai_thread_precedenti.z; val2.x += correzione_totale; val2.y += correzione_totale; val2.z += correzione_totale; val2.w += correzione_totale; correzione_totale = correzione_dal_blocco_precedente.w + correzione_dai_thread_precedenti.w; val3.x += correzione_totale; val3.y += correzione_totale; val3.z += correzione_totale; val3.w += correzione_totale; correzione_dal_blocco_precedente.x += shared[blockDim.x-1].x; //correzione di 00 (0) correzione_dal_blocco_precedente.y += shared[blockDim.x-1].y; //correzione di 01 (1) correzione_dal_blocco_precedente.z += shared[blockDim.x-1].z; //correzione di 10 (2) correzione_dal_blocco_precedente.w += shared[blockDim.x-1].w; //correzione di 11 (3) if (idx < nels) { code0[idx] = val0; code1[idx] = val1; code2[idx] = val2; code3[idx] = val3; } } } __global__ void fixup(int4 * restrict scan0,int4 * restrict scan1,int4 * restrict scan2,int4 * restrict scan3,int nels ,const int * restrict code0,const int * restrict code1,const int * restrict code2,const int * restrict code3,const int4* restrict in,int nbit,int4* max) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int4 correzione_dal_blocco_precedente =make_int4(0,0,0,0); if(blockIdx.x>0) { correzione_dal_blocco_precedente.x = code0[blockIdx.x - 1]; correzione_dal_blocco_precedente.y = code1[blockIdx.x - 1]; correzione_dal_blocco_precedente.z = code2[blockIdx.x - 1]; correzione_dal_blocco_precedente.w = code3[blockIdx.x - 1]; } int numero_cicli = (els_per_sezione + blockDim.x - 1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { if (idx < elemento_limite) { int4 val0 = scan0[idx], val1=scan1[idx], val2=scan2[idx], val3=scan3[idx]; int4 val_in=in[idx]; if(idx==nels-1) { //salvataggio in memoria globale degli ultimi elementi degli scan inclusivi (*max).x=val0.w+correzione_dal_blocco_precedente.x; (*max).y=val1.w+correzione_dal_blocco_precedente.y; (*max).z=val2.w+correzione_dal_blocco_precedente.z; (*max).w=val3.w+correzione_dal_blocco_precedente.w; } //trasformazione degli scan da inclusivi ad esclusivi val0.x += correzione_dal_blocco_precedente.x - ((((val_in.x>>nbit)&3)==0)?1:0); val0.y += correzione_dal_blocco_precedente.x - ((((val_in.y>>nbit)&3)==0)?1:0); val0.z += correzione_dal_blocco_precedente.x - ((((val_in.z>>nbit)&3)==0)?1:0); val0.w += correzione_dal_blocco_precedente.x - ((((val_in.w>>nbit)&3)==0)?1:0); val1.x += correzione_dal_blocco_precedente.y - ((((val_in.x>>nbit)&3)==1)?1:0); val1.y += correzione_dal_blocco_precedente.y - ((((val_in.y>>nbit)&3)==1)?1:0); val1.z += correzione_dal_blocco_precedente.y - ((((val_in.z>>nbit)&3)==1)?1:0); val1.w += correzione_dal_blocco_precedente.y - ((((val_in.w>>nbit)&3)==1)?1:0); val2.x += correzione_dal_blocco_precedente.z - ((((val_in.x>>nbit)&3)==2)?1:0); val2.y += correzione_dal_blocco_precedente.z - ((((val_in.y>>nbit)&3)==2)?1:0); val2.z += correzione_dal_blocco_precedente.z - ((((val_in.z>>nbit)&3)==2)?1:0); val2.w += correzione_dal_blocco_precedente.z - ((((val_in.w>>nbit)&3)==2)?1:0); val3.x += correzione_dal_blocco_precedente.w - ((((val_in.x>>nbit)&3)==3)?1:0); val3.y += correzione_dal_blocco_precedente.w - ((((val_in.y>>nbit)&3)==3)?1:0); val3.z += correzione_dal_blocco_precedente.w - ((((val_in.z>>nbit)&3)==3)?1:0); val3.w += correzione_dal_blocco_precedente.w - ((((val_in.w>>nbit)&3)==3)?1:0); scan0[idx] = val0; scan1[idx] = val1; scan2[idx] = val2; scan3[idx] = val3; } } } //kernel adibito al riordino dei vettori utilizzando 2 bit __global__ void reorder(const int4* restrict scan0,const int4* restrict scan1,const int4* restrict scan2,const int4* restrict scan3,const int4* restrict in, int* restrict out,int nels,int4* max) { int els_per_sezione = (nels + gridDim.x - 1)/gridDim.x; int idx = threadIdx.x + blockIdx.x*els_per_sezione; int numero_cicli = (els_per_sezione +blockDim.x -1)/blockDim.x; int elemento_limite = min(els_per_sezione*(blockIdx.x+1), nels); int4 offset_max=make_int4((*max).x,(*max).y,(*max).z,(*max).w); int4 val_scan_succ; //scan dei max int4 max_scan=make_int4(offset_max.x,offset_max.y,offset_max.z,offset_max.w); offset_max.x=0; //offset di 00 (0) offset_max.y=max_scan.x; //offset di 01 (1) offset_max.z=offset_max.y + max_scan.y; //offset di 10 (2) offset_max.w=offset_max.z + max_scan.z; //offset di 11 (3) //inizio ciclo di reorder for (int ciclo = 0; ciclo < numero_cicli;++ciclo, idx += blockDim.x) { if (idx < elemento_limite) { int4 val_num=in[idx], val_scan0=scan0[idx], val_scan1=scan1[idx], val_scan2=scan2[idx], val_scan3=scan3[idx]; // confronto 1° elemento con 2° elemento della quartina if(val_scan0.x!=val_scan0.y) out[val_scan0.x+offset_max.x]=val_num.x; else if(val_scan1.x!=val_scan1.y) out[val_scan1.x+offset_max.y]=val_num.x; else if(val_scan2.x!=val_scan2.y) out[val_scan2.x+offset_max.z]=val_num.x; else out[val_scan3.x+offset_max.w]=val_num.x; // confronto 2° elemento con 3° elemento della quartina if(val_scan0.y!=val_scan0.z) out[val_scan0.y+offset_max.x]=val_num.y; else if(val_scan1.y!=val_scan1.z) out[val_scan1.y+offset_max.y]=val_num.y; else if(val_scan2.y!=val_scan2.z) out[val_scan2.y+offset_max.z]=val_num.y; else out[val_scan3.y+offset_max.w]=val_num.y; // confronto 3° elemento con 4° elemento della quartina if(val_scan0.z!=val_scan0.w) out[val_scan0.z+offset_max.x]=val_num.z; else if(val_scan1.z!=val_scan1.w) out[val_scan1.z+offset_max.y]=val_num.z; else if(val_scan2.z!=val_scan2.w) out[val_scan2.z+offset_max.z]=val_num.z; else out[val_scan3.z+offset_max.w]=val_num.z; //confronto 4° elemento con 1° elemento della quartina dello scan successivo if(idx!=nels-1) { //scan3[idx+1] puo non essere preso, visto che non viene mai usato val_scan_succ=make_int4(scan0[idx+1].x,scan1[idx+1].x,scan2[idx+1].x,0); // primi valori delle quartine successive degli scan if(val_scan0.w!=val_scan_succ.x) out[val_scan0.w + offset_max.x]=val_num.w; else if(val_scan1.w!=val_scan_succ.y) out[val_scan1.w + offset_max.y]=val_num.w; else if(val_scan2.w!=val_scan_succ.z) out[val_scan2.w + offset_max.z]=val_num.w; else out[val_scan3.w + offset_max.w]=val_num.w; } else { if(val_scan0.w!=max_scan.x) out[val_scan0.w + offset_max.x]=val_num.w; else if(val_scan1.w!=max_scan.y) out[val_scan1.w + offset_max.y]=val_num.w; else if(val_scan2.w!=max_scan.z) out[val_scan2.w + offset_max.z]=val_num.w; else out[val_scan3.w + offset_max.w]=val_num.w; } } } } int main(int argc, char *argv[]) { if (argc < 4) error("sintassi radix_sort: numels thread_per_blocco numero_blocchi_scan valore_massimo"); int nels = atoi(argv[1]); /* numero di elementi */ if (nels <= 0) error("il numero di elementi deve essere positivo"); if (nels & 3) error("il numero di elementi deve essere multiplo di 4"); int numThreads = atoi(argv[2]); /* local work size */ if (numThreads <= 0) error("il numero di thread per blocco deve essere positivo"); int numBlocksScan = atoi(argv[3]); /* numero blocchi scan */ if (numBlocksScan <= 0) error("il numero di blocchi deve essere positivo"); int numMax = atoi(argv[4]); /* numero blocchi scan */ if (numMax <= 0) error("il valore massimo deve essere positivo"); //inizializzazione dei vettori const size_t memsize = sizeof(int)*nels; int4 *d_v1, *d_scan0,*d_scan1,*d_scan2,*d_scan3, *d_code0,*d_code1,*d_code2,*d_code3,*d_out,*tmp; int numbit; int4 *d_max; //calolo dei cicli da fare avendo il massimo int cicli=int(log(numMax)/log(2)) + 1; printf("numero cicli da fare=%d\n",cicli/2); //allocazione dei vettori su GPU hipError_t err = hipMalloc(&d_v1, memsize); cuda_check(err, "malloc v1"); err= hipMalloc(&d_max,sizeof(int4)); cuda_check(err,"malloc max"); err = hipMalloc(&d_scan0, memsize); cuda_check(err, "malloc scan0"); err = hipMalloc(&d_scan1, memsize); cuda_check(err, "malloc scan1"); err = hipMalloc(&d_scan2, memsize); cuda_check(err, "malloc scan2"); err = hipMalloc(&d_scan3, memsize); cuda_check(err, "malloc scan3"); err = hipMalloc(&d_code0, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code0"); err = hipMalloc(&d_code1, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code1"); err = hipMalloc(&d_code2, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code2"); err = hipMalloc(&d_code3, numBlocksScan*sizeof(int)); cuda_check(err, "malloc code3"); err = hipMalloc(&d_out, memsize); cuda_check(err, "malloc out"); //allocazione su CPU int *vout = (int*)malloc(memsize); if (!vout) error("alloc vscan"); //inizializzazione su CPU di numeri random con massimo possibile numMax init_random(vout,nels,numMax); /* //inizializzazione su GPU di numeri decrescenti a partire da nels int numBlocks = (nels + numThreads - 1)/numThreads; init<<<numBlocks, numThreads>>>((int*)d_v1, nels); */ //prova ad otimizzare la cache hipFuncSetCacheConfig(reinterpret_cast<const void*>(scan_step1), hipFuncCachePreferL1); hipFuncSetCacheConfig(reinterpret_cast<const void*>(scan_step2), hipFuncCachePreferL1); hipFuncSetCacheConfig(reinterpret_cast<const void*>(fixup), hipFuncCachePreferL1); hipFuncSetCacheConfig(reinterpret_cast<const void*>(reorder), hipFuncCachePreferL1); err = hipMemcpy(d_v1,vout,memsize, hipMemcpyHostToDevice); cuda_check(err, "memcpy vett su GPU"); //creazione eventi hipEvent_t before_scan, after_scan; err = hipEventCreate(&before_scan); cuda_check(err, "create event before"); err = hipEventCreate(&after_scan); cuda_check(err, "create event after"); hipEventRecord(before_scan); for(numbit=0;numbit<cicli;numbit+=2) { //pulizia dei vettori di code err= hipMemset (d_code0,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset0"); err= hipMemset (d_code1,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset1"); err= hipMemset (d_code2,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset2"); err= hipMemset (d_code3,0, numBlocksScan*sizeof(int)); cuda_check(err, "memset3"); scan_step1<<<numBlocksScan, numThreads, numThreads*sizeof(int)*4>>> (d_scan0,d_scan1,d_scan2,d_scan3, d_v1, nels/4, (int*)d_code0,(int*)d_code1,(int*)d_code2,(int*)d_code3, numbit); scan_step2<<<1, numThreads, numThreads*sizeof(int)*4>>> (d_code0, d_code1,d_code2,d_code3, numBlocksScan/4); fixup<<<numBlocksScan, numThreads>>>(d_scan0,d_scan1,d_scan2,d_scan3, nels/4, (int*)d_code0,(int*)d_code1,(int*)d_code2,(int*)d_code3,d_v1,numbit,d_max); reorder<<<numBlocksScan,numThreads>>>(d_scan0,d_scan1,d_scan2,d_scan3,d_v1,(int*)d_out,nels/4,d_max); //scambio deii puntatori d_out e d_v1 if(numbit+2 <cicli) { tmp=d_v1; d_v1=d_out; d_out=tmp; } } hipEventRecord(after_scan); err = hipEventSynchronize(after_scan); cuda_check(err, "after scan sznc"); float runtime_ms; hipEventElapsedTime(&runtime_ms, before_scan, after_scan); printf("scan runtime: %.4g ms\n", runtime_ms); err = hipMemcpy(vout, d_out, memsize, hipMemcpyDeviceToHost); cuda_check(err, "memcpy"); printf("\n\n"); verify_random(vout,nels)?printf("Ordinamento riuscito!\n"):printf("Ordinamento non riuscito!\n"); if(nels <=32) for(int i=0;i<nels;++i) printf("%d ",vout[i]); printf("\n"); }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include <stdio.h> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; cudaDeviceProp prop; cudaGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; cudaGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } cudaMalloc((void**)(&dev_x), size); cudaMalloc((void**)(&dev_y), size); cudaMemcpy(dev_x, x, size, cudaMemcpyHostToDevice); cudaMemcpy(dev_y, y, size, cudaMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); cudaMemcpy(yy, dev_y, size, cudaMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); cudaFree(dev_x); cudaFree(dev_y); exit(0); }
code for sm_80 Function : _Z5saxpyjfPfS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R4, R4, c[0x0][0x0], R3 ; /* 0x0000000004047a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R4, c[0x0][0x160], PT ; /* 0x0000580004007a0c */ /* 0x000fda0003f06070 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE.U32 R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x000fc800078e0005 */ /*0090*/ IMAD.WIDE.U32 R4, R4, R5, c[0x0][0x170] ; /* 0x00005c0004047625 */ /* 0x000fe400078e0005 */ /*00a0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*00c0*/ FFMA R7, R2, c[0x0][0x164], R7 ; /* 0x0000590002077a23 */ /* 0x004fca0000000007 */ /*00d0*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00f0*/ BRA 0xf0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; cudaDeviceProp prop; cudaGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; cudaGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } cudaMalloc((void**)(&dev_x), size); cudaMalloc((void**)(&dev_y), size); cudaMemcpy(dev_x, x, size, cudaMemcpyHostToDevice); cudaMemcpy(dev_y, y, size, cudaMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); cudaMemcpy(yy, dev_y, size, cudaMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); cudaFree(dev_x); cudaFree(dev_y); exit(0); }
.file "tmpxft_00020df7_00000000-6_saxpy.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 "" .LC1: .string " " .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "This machine has %d CUDA devices.\n" .section .rodata.str1.1 .LC3: .string "Device %d:\n" .LC4: .string "%sdevice.name = %s\n" .section .rodata.str1.8 .align 8 .LC5: .string "%sdevice.maxThreadsPerBlock = %d\n" .text .globl _Z4miscv .type _Z4miscv, @function _Z4miscv: .LFB2057: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $1064, %rsp .cfi_def_cfa_offset 1120 movq %fs:40, %rax movq %rax, 1048(%rsp) xorl %eax, %eax leaq 12(%rsp), %rdi call cudaGetDeviceCount@PLT movl 12(%rsp), %edx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %ebp leaq .LC1(%rip), %rbx leaq 16(%rsp), %r12 leaq .LC3(%rip), %r15 leaq .LC4(%rip), %r14 leaq .LC5(%rip), %r13 cmpl $0, 12(%rsp) jg .L4 .L3: movq 1048(%rsp), %rax subq %fs:40, %rax jne .L14 addq $1064, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state movq %r12, %rcx movq %rbx, %rdx movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 336(%rsp), %ecx movq %rbx, %rdx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %ebp movl 12(%rsp), %eax cmpl %ebp, %eax jle .L3 testl %eax, %eax leaq .LC1(%rip), %rbx leaq .LC0(%rip), %rax cmove %rax, %rbx .L4: movl %ebp, %esi movq %r12, %rdi call cudaGetDeviceProperties_v2@PLT cmpl $0, 12(%rsp) jle .L7 movl %ebp, %edx movq %r15, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L7 .L14: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z4miscv, .-_Z4miscv .globl _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ .type _Z28__device_stub__Z5saxpyjfPfS_jfPfS_, @function _Z28__device_stub__Z5saxpyjfPfS_jfPfS_: .LFB2083: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 28(%rsp) movss %xmm0, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%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 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 .L19 .L15: movq 136(%rsp), %rax subq %fs:40, %rax jne .L20 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L19: .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 _Z5saxpyjfPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L15 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z28__device_stub__Z5saxpyjfPfS_jfPfS_, .-_Z28__device_stub__Z5saxpyjfPfS_jfPfS_ .globl _Z5saxpyjfPfS_ .type _Z5saxpyjfPfS_, @function _Z5saxpyjfPfS_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z5saxpyjfPfS_, .-_Z5saxpyjfPfS_ .section .rodata.str1.8 .align 8 .LC12: .string "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n" .section .rodata.str1.1 .LC13: .string "The results match!\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $56, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r12 movl %eax, %r15d call _Z4miscv leal 0(,%r12,4), %r14d movslq %r14d, %r14 movq %r14, %rdi call malloc@PLT movq %rax, %rbp movq %r14, %rdi call malloc@PLT movq %rax, %rbx movq %r14, %rdi call malloc@PLT movq %rax, %r13 testl %r12d, %r12d je .L24 movl %r12d, %ecx movl $0, %eax .L25: pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 movss %xmm0, 0(%rbp,%rax,4) movl %eax, %edx imull %eax, %edx pxor %xmm0, %xmm0 cvtsi2ssl %edx, %xmm0 movss %xmm0, (%rbx,%rax,4) addq $1, %rax cmpq %rcx, %rax jne .L25 .L24: movq %rsp, %rdi movq %r14, %rsi call cudaMalloc@PLT leaq 8(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r14, %rdx movq %rbp, %rsi movq (%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %r14, %rdx movq %rbx, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $256, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %r12d, %eax pxor %xmm0, %xmm0 cvtsi2sdq %rax, %xmm0 mulsd .LC6(%rip), %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L28 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L28: cvttsd2siq %xmm3, %rax movl %eax, 16(%rsp) movl $1, 20(%rsp) movl 36(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L39 .L29: movl $2, %ecx movq %r14, %rdx movq 8(%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT testl %r12d, %r12d je .L30 movl %r12d, %r12d movl $0, %ecx movss .LC10(%rip), %xmm4 .L33: movss 0(%r13,%rcx,4), %xmm3 movss 0(%rbp,%rcx,4), %xmm1 movss (%rbx,%rcx,4), %xmm2 movaps %xmm1, %xmm0 mulss %xmm4, %xmm0 addss %xmm2, %xmm0 ucomiss %xmm0, %xmm3 jp .L35 jne .L35 addq $1, %rcx cmpq %r12, %rcx jne .L33 .L30: leaq .LC13(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq %r13, %rdi call free@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movl $0, %edi call exit@PLT .L39: movq 8(%rsp), %rdx movq (%rsp), %rsi movss .LC10(%rip), %xmm0 movl %r15d, %edi call _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ jmp .L29 .L35: cvtss2sd %xmm3, %xmm3 cvtss2sd %xmm2, %xmm2 cvtss2sd %xmm1, %xmm1 movsd .LC11(%rip), %xmm0 leaq .LC12(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $4, %eax call __fprintf_chk@PLT movl $-1, %edi call exit@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC15: .string "_Z5saxpyjfPfS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _Z5saxpyjfPfS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC6: .long 0 .long 1064304640 .align 8 .LC7: .long 0 .long 1127219200 .align 8 .LC9: .long 0 .long 1072693248 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC10: .long 1077936128 .section .rodata.cst8 .align 8 .LC11: .long 0 .long 1074266112 .align 8 .LC14: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; cudaDeviceProp prop; cudaGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; cudaGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } cudaMalloc((void**)(&dev_x), size); cudaMalloc((void**)(&dev_y), size); cudaMemcpy(dev_x, x, size, cudaMemcpyHostToDevice); cudaMemcpy(dev_y, y, size, cudaMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); cudaMemcpy(yy, dev_y, size, cudaMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); cudaFree(dev_x); cudaFree(dev_y); exit(0); }
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; hipDeviceProp_t prop; hipGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; hipGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } hipMalloc((void**)(&dev_x), size); hipMalloc((void**)(&dev_y), size); hipMemcpy(dev_x, x, size, hipMemcpyHostToDevice); hipMemcpy(dev_y, y, size, hipMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); hipMemcpy(yy, dev_y, size, hipMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); hipFree(dev_x); hipFree(dev_y); exit(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> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; hipDeviceProp_t prop; hipGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; hipGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } hipMalloc((void**)(&dev_x), size); hipMalloc((void**)(&dev_y), size); hipMemcpy(dev_x, x, size, hipMemcpyHostToDevice); hipMemcpy(dev_y, y, size, hipMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); hipMemcpy(yy, dev_y, size, hipMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); hipFree(dev_x); hipFree(dev_y); exit(0); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z5saxpyjfPfS_ .globl _Z5saxpyjfPfS_ .p2align 8 .type _Z5saxpyjfPfS_,@function _Z5saxpyjfPfS_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b32 s3, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e64 s3, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x8 v_mov_b32_e32 v2, 0 s_load_b32 s0, s[0:1], 0x4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v1, vcc_lo global_load_b32 v2, v[2:3], off global_load_b32 v3, v[0:1], off s_waitcnt vmcnt(0) v_fmac_f32_e32 v3, s0, v2 global_store_b32 v[0:1], v3, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5saxpyjfPfS_ .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 _Z5saxpyjfPfS_, .Lfunc_end0-_Z5saxpyjfPfS_ .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 - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5saxpyjfPfS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5saxpyjfPfS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 4 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> __global__ void saxpy(uint n, float a, float *x, float *y) { uint i = blockIdx.x*blockDim.x + threadIdx.x; // nvcc built-ins if(i < n) y[i] = a*x[i] + y[i]; } void misc(void) { int ndev; hipDeviceProp_t prop; hipGetDeviceCount(&ndev); printf("This machine has %d CUDA devices.\n", ndev); for(int i = 0; i < ndev; i++) { const char *indent = (ndev == 0) ? "" : " "; hipGetDeviceProperties(&prop, i); if(ndev > 0) printf("Device %d:\n", i); printf("%sdevice.name = %s\n", indent, prop.name); printf("%sdevice.maxThreadsPerBlock = %d\n", indent, prop.maxThreadsPerBlock); } } int main(int argc, char **argv) { uint n = atoi(argv[1]); int size = n*sizeof(float); float *x, *y, *yy; float *dev_x, *dev_y; misc(); x = (float *)malloc(size); y = (float *)malloc(size); yy = (float *)malloc(size); for(int i = 0; i < n; i++) { x[i] = i; y[i] = i*i; } hipMalloc((void**)(&dev_x), size); hipMalloc((void**)(&dev_y), size); hipMemcpy(dev_x, x, size, hipMemcpyHostToDevice); hipMemcpy(dev_y, y, size, hipMemcpyHostToDevice); float a = 3.0; saxpy<<<ceil(n/256.0),256>>>(n, a, dev_x, dev_y); hipMemcpy(yy, dev_y, size, hipMemcpyDeviceToHost); for(int i = 0; i < n; i++) { // check the result if(yy[i] != a*x[i] + y[i]) { fprintf(stderr, "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n", i, a, x[i], y[i], yy[i]); exit(-1); } } printf("The results match!\n"); free(x); free(y); free(yy); hipFree(dev_x); hipFree(dev_y); exit(0); }
.text .file "saxpy.hip" .globl _Z20__device_stub__saxpyjfPfS_ # -- Begin function _Z20__device_stub__saxpyjfPfS_ .p2align 4, 0x90 .type _Z20__device_stub__saxpyjfPfS_,@function _Z20__device_stub__saxpyjfPfS_: # @_Z20__device_stub__saxpyjfPfS_ .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movss %xmm0, 8(%rsp) movq %rsi, 72(%rsp) movq %rdx, 64(%rsp) leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) leaq 72(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%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 $_Z5saxpyjfPfS_, %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 _Z20__device_stub__saxpyjfPfS_, .Lfunc_end0-_Z20__device_stub__saxpyjfPfS_ .cfi_endproc # -- End function .globl _Z4miscv # -- Begin function _Z4miscv .p2align 4, 0x90 .type _Z4miscv,@function _Z4miscv: # @_Z4miscv .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 1520 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 leaq 12(%rsp), %rdi callq hipGetDeviceCount movl 12(%rsp), %esi movl $.L.str, %edi xorl %eax, %eax callq printf cmpl $0, 12(%rsp) jle .LBB1_7 # %bb.1: # %.lr.ph movl $1, %eax leaq 16(%rsp), %rbx xorl %ebp, %ebp jmp .LBB1_2 .p2align 4, 0x90 .LBB1_6: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf movl 336(%rsp), %edx movl $.L.str.5, %edi movq %r14, %rsi xorl %eax, %eax callq printf incl %ebp movl 12(%rsp), %eax cmpl %eax, %ebp jge .LBB1_7 .LBB1_2: # =>This Inner Loop Header: Depth=1 movl $.L.str.1, %r14d testl %eax, %eax je .LBB1_4 # %bb.3: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.2, %r14d .LBB1_4: # in Loop: Header=BB1_2 Depth=1 movq %rbx, %rdi movl %ebp, %esi callq hipGetDevicePropertiesR0600 cmpl $0, 12(%rsp) jle .LBB1_6 # %bb.5: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.3, %edi movl %ebp, %esi xorl %eax, %eax callq printf jmp .LBB1_6 .LBB1_7: # %._crit_edge addq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z4miscv, .Lfunc_end1-_Z4miscv .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI2_0: .quad 0x3f70000000000000 # double 0.00390625 .LCPI2_2: .quad 0x4008000000000000 # double 3 .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 .LCPI2_1: .long 0x40400000 # float 3 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $128, %rsp .cfi_def_cfa_offset 176 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r12 leal (,%rax,4), %ebx callq _Z4miscv movslq %ebx, %r13 movq %r13, %rdi callq malloc movq %rax, %rbx movq %r13, %rdi callq malloc movq %rax, %r14 movq %r13, %rdi callq malloc movq %rax, %r15 testl %r12d, %r12d je .LBB2_3 # %bb.1: # %.lr.ph.preheader movl %r12d, %eax xorl %ecx, %ecx .p2align 4, 0x90 .LBB2_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 xorps %xmm0, %xmm0 cvtsi2ss %ecx, %xmm0 movl %ecx, %edx imull %ecx, %edx xorps %xmm1, %xmm1 cvtsi2ss %edx, %xmm1 movss %xmm0, (%rbx,%rcx,4) movss %xmm1, (%r14,%rcx,4) incq %rcx cmpq %rcx, %rax jne .LBB2_2 .LBB2_3: # %._crit_edge leaq 16(%rsp), %rdi movq %r13, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %rbx, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq %r14, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movl %r12d, %eax xorps %xmm0, %xmm0 cvtsi2sd %rax, %xmm0 mulsd .LCPI2_0(%rip), %xmm0 callq ceil@PLT cvttsd2si %xmm0, %rax movl %eax, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $256, %rdx # imm = 0x100 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_5 # %bb.4: movq 16(%rsp), %rax movq 8(%rsp), %rcx movl %r12d, 28(%rsp) movl $1077936128, 24(%rsp) # imm = 0x40400000 movq %rax, 88(%rsp) movq %rcx, 80(%rsp) leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 24(%rsp), %rax movq %rax, 104(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 64(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z5saxpyjfPfS_, %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 .LBB2_5: movq 8(%rsp), %rsi movq %r15, %rdi movq %r13, %rdx movl $2, %ecx callq hipMemcpy testl %r12d, %r12d je .LBB2_9 # %bb.6: # %.lr.ph60.preheader movl %r12d, %eax xorl %edx, %edx movss .LCPI2_1(%rip), %xmm0 # xmm0 = mem[0],zero,zero,zero .p2align 4, 0x90 .LBB2_7: # %.lr.ph60 # =>This Inner Loop Header: Depth=1 movss (%r15,%rdx,4), %xmm3 # xmm3 = mem[0],zero,zero,zero movss (%rbx,%rdx,4), %xmm1 # xmm1 = mem[0],zero,zero,zero movaps %xmm1, %xmm4 mulss %xmm0, %xmm4 movss (%r14,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero addss %xmm2, %xmm4 ucomiss %xmm4, %xmm3 jne .LBB2_10 jp .LBB2_10 # %bb.8: # in Loop: Header=BB2_7 Depth=1 incq %rdx cmpq %rdx, %rax jne .LBB2_7 .LBB2_9: # %._crit_edge61 movl $.Lstr, %edi callq puts@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree xorl %edi, %edi callq exit .LBB2_10: movq stderr(%rip), %rdi cvtss2sd %xmm1, %xmm1 cvtss2sd %xmm2, %xmm2 cvtss2sd %xmm3, %xmm3 movsd .LCPI2_2(%rip), %xmm0 # xmm0 = mem[0],zero movl $.L.str.6, %esi # kill: def $edx killed $edx killed $rdx movb $4, %al callq fprintf movl $-1, %edi callq exit .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5saxpyjfPfS_, %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 _Z5saxpyjfPfS_,@object # @_Z5saxpyjfPfS_ .section .rodata,"a",@progbits .globl _Z5saxpyjfPfS_ .p2align 3, 0x0 _Z5saxpyjfPfS_: .quad _Z20__device_stub__saxpyjfPfS_ .size _Z5saxpyjfPfS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "This machine has %d CUDA devices.\n" .size .L.str, 35 .type .L.str.1,@object # @.str.1 .L.str.1: .zero 1 .size .L.str.1, 1 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz " " .size .L.str.2, 3 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Device %d:\n" .size .L.str.3, 12 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%sdevice.name = %s\n" .size .L.str.4, 20 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "%sdevice.maxThreadsPerBlock = %d\n" .size .L.str.5, 34 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n" .size .L.str.6, 49 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5saxpyjfPfS_" .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 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "The results match!" .size .Lstr, 19 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__saxpyjfPfS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5saxpyjfPfS_ .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 : _Z5saxpyjfPfS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R4, R4, c[0x0][0x0], R3 ; /* 0x0000000004047a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.U32.AND P0, PT, R4, c[0x0][0x160], PT ; /* 0x0000580004007a0c */ /* 0x000fda0003f06070 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE.U32 R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x000fc800078e0005 */ /*0090*/ IMAD.WIDE.U32 R4, R4, R5, c[0x0][0x170] ; /* 0x00005c0004047625 */ /* 0x000fe400078e0005 */ /*00a0*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*00c0*/ FFMA R7, R2, c[0x0][0x164], R7 ; /* 0x0000590002077a23 */ /* 0x004fca0000000007 */ /*00d0*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00e0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00f0*/ BRA 0xf0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 _Z5saxpyjfPfS_ .globl _Z5saxpyjfPfS_ .p2align 8 .type _Z5saxpyjfPfS_,@function _Z5saxpyjfPfS_: s_clause 0x1 s_load_b32 s2, s[0:1], 0x24 s_load_b32 s3, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_u32_e64 s3, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x8 v_mov_b32_e32 v2, 0 s_load_b32 s0, s[0:1], 0x4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s7, v1, vcc_lo global_load_b32 v2, v[2:3], off global_load_b32 v3, v[0:1], off s_waitcnt vmcnt(0) v_fmac_f32_e32 v3, s0, v2 global_store_b32 v[0:1], v3, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z5saxpyjfPfS_ .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 _Z5saxpyjfPfS_, .Lfunc_end0-_Z5saxpyjfPfS_ .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 - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: hidden_block_count_x - .offset: 28 .size: 4 .value_kind: hidden_block_count_y - .offset: 32 .size: 4 .value_kind: hidden_block_count_z - .offset: 36 .size: 2 .value_kind: hidden_group_size_x - .offset: 38 .size: 2 .value_kind: hidden_group_size_y - .offset: 40 .size: 2 .value_kind: hidden_group_size_z - .offset: 42 .size: 2 .value_kind: hidden_remainder_x - .offset: 44 .size: 2 .value_kind: hidden_remainder_y - .offset: 46 .size: 2 .value_kind: hidden_remainder_z - .offset: 64 .size: 8 .value_kind: hidden_global_offset_x - .offset: 72 .size: 8 .value_kind: hidden_global_offset_y - .offset: 80 .size: 8 .value_kind: hidden_global_offset_z - .offset: 88 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 280 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z5saxpyjfPfS_ .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z5saxpyjfPfS_.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_00020df7_00000000-6_saxpy.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 "" .LC1: .string " " .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC2: .string "This machine has %d CUDA devices.\n" .section .rodata.str1.1 .LC3: .string "Device %d:\n" .LC4: .string "%sdevice.name = %s\n" .section .rodata.str1.8 .align 8 .LC5: .string "%sdevice.maxThreadsPerBlock = %d\n" .text .globl _Z4miscv .type _Z4miscv, @function _Z4miscv: .LFB2057: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $1064, %rsp .cfi_def_cfa_offset 1120 movq %fs:40, %rax movq %rax, 1048(%rsp) xorl %eax, %eax leaq 12(%rsp), %rdi call cudaGetDeviceCount@PLT movl 12(%rsp), %edx leaq .LC2(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $0, %ebp leaq .LC1(%rip), %rbx leaq 16(%rsp), %r12 leaq .LC3(%rip), %r15 leaq .LC4(%rip), %r14 leaq .LC5(%rip), %r13 cmpl $0, 12(%rsp) jg .L4 .L3: movq 1048(%rsp), %rax subq %fs:40, %rax jne .L14 addq $1064, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state movq %r12, %rcx movq %rbx, %rdx movq %r14, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl 336(%rsp), %ecx movq %rbx, %rdx movq %r13, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT addl $1, %ebp movl 12(%rsp), %eax cmpl %ebp, %eax jle .L3 testl %eax, %eax leaq .LC1(%rip), %rbx leaq .LC0(%rip), %rax cmove %rax, %rbx .L4: movl %ebp, %esi movq %r12, %rdi call cudaGetDeviceProperties_v2@PLT cmpl $0, 12(%rsp) jle .L7 movl %ebp, %edx movq %r15, %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L7 .L14: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z4miscv, .-_Z4miscv .globl _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ .type _Z28__device_stub__Z5saxpyjfPfS_jfPfS_, @function _Z28__device_stub__Z5saxpyjfPfS_jfPfS_: .LFB2083: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movl %edi, 28(%rsp) movss %xmm0, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%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 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 .L19 .L15: movq 136(%rsp), %rax subq %fs:40, %rax jne .L20 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L19: .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 _Z5saxpyjfPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L15 .L20: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z28__device_stub__Z5saxpyjfPfS_jfPfS_, .-_Z28__device_stub__Z5saxpyjfPfS_jfPfS_ .globl _Z5saxpyjfPfS_ .type _Z5saxpyjfPfS_, @function _Z5saxpyjfPfS_: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z5saxpyjfPfS_, .-_Z5saxpyjfPfS_ .section .rodata.str1.8 .align 8 .LC12: .string "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n" .section .rodata.str1.1 .LC13: .string "The results match!\n" .text .globl main .type main, @function main: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $56, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movq 8(%rsi), %rdi movl $10, %edx movl $0, %esi call __isoc23_strtol@PLT movq %rax, %r12 movl %eax, %r15d call _Z4miscv leal 0(,%r12,4), %r14d movslq %r14d, %r14 movq %r14, %rdi call malloc@PLT movq %rax, %rbp movq %r14, %rdi call malloc@PLT movq %rax, %rbx movq %r14, %rdi call malloc@PLT movq %rax, %r13 testl %r12d, %r12d je .L24 movl %r12d, %ecx movl $0, %eax .L25: pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 movss %xmm0, 0(%rbp,%rax,4) movl %eax, %edx imull %eax, %edx pxor %xmm0, %xmm0 cvtsi2ssl %edx, %xmm0 movss %xmm0, (%rbx,%rax,4) addq $1, %rax cmpq %rcx, %rax jne .L25 .L24: movq %rsp, %rdi movq %r14, %rsi call cudaMalloc@PLT leaq 8(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r14, %rdx movq %rbp, %rsi movq (%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %r14, %rdx movq %rbx, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $256, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl %r12d, %eax pxor %xmm0, %xmm0 cvtsi2sdq %rax, %xmm0 mulsd .LC6(%rip), %xmm0 movapd %xmm0, %xmm3 movsd .LC14(%rip), %xmm2 movapd %xmm0, %xmm1 andpd %xmm2, %xmm1 movsd .LC7(%rip), %xmm4 ucomisd %xmm1, %xmm4 jbe .L28 cvttsd2siq %xmm0, %rax pxor %xmm1, %xmm1 cvtsi2sdq %rax, %xmm1 cmpnlesd %xmm1, %xmm3 movsd .LC9(%rip), %xmm4 andpd %xmm4, %xmm3 addsd %xmm1, %xmm3 andnpd %xmm0, %xmm2 orpd %xmm2, %xmm3 .L28: cvttsd2siq %xmm3, %rax movl %eax, 16(%rsp) movl $1, 20(%rsp) movl 36(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L39 .L29: movl $2, %ecx movq %r14, %rdx movq 8(%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT testl %r12d, %r12d je .L30 movl %r12d, %r12d movl $0, %ecx movss .LC10(%rip), %xmm4 .L33: movss 0(%r13,%rcx,4), %xmm3 movss 0(%rbp,%rcx,4), %xmm1 movss (%rbx,%rcx,4), %xmm2 movaps %xmm1, %xmm0 mulss %xmm4, %xmm0 addss %xmm2, %xmm0 ucomiss %xmm0, %xmm3 jp .L35 jne .L35 addq $1, %rcx cmpq %r12, %rcx jne .L33 .L30: leaq .LC13(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movq %rbp, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq %r13, %rdi call free@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movl $0, %edi call exit@PLT .L39: movq 8(%rsp), %rdx movq (%rsp), %rsi movss .LC10(%rip), %xmm0 movl %r15d, %edi call _Z28__device_stub__Z5saxpyjfPfS_jfPfS_ jmp .L29 .L35: cvtss2sd %xmm3, %xmm3 cvtss2sd %xmm2, %xmm2 cvtss2sd %xmm1, %xmm1 movsd .LC11(%rip), %xmm0 leaq .LC12(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $4, %eax call __fprintf_chk@PLT movl $-1, %edi call exit@PLT .cfi_endproc .LFE2058: .size main, .-main .section .rodata.str1.1 .LC15: .string "_Z5saxpyjfPfS_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _Z5saxpyjfPfS_(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2086: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC6: .long 0 .long 1064304640 .align 8 .LC7: .long 0 .long 1127219200 .align 8 .LC9: .long 0 .long 1072693248 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC10: .long 1077936128 .section .rodata.cst8 .align 8 .LC11: .long 0 .long 1074266112 .align 8 .LC14: .long -1 .long 2147483647 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "saxpy.hip" .globl _Z20__device_stub__saxpyjfPfS_ # -- Begin function _Z20__device_stub__saxpyjfPfS_ .p2align 4, 0x90 .type _Z20__device_stub__saxpyjfPfS_,@function _Z20__device_stub__saxpyjfPfS_: # @_Z20__device_stub__saxpyjfPfS_ .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movss %xmm0, 8(%rsp) movq %rsi, 72(%rsp) movq %rdx, 64(%rsp) leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) leaq 72(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%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 $_Z5saxpyjfPfS_, %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 _Z20__device_stub__saxpyjfPfS_, .Lfunc_end0-_Z20__device_stub__saxpyjfPfS_ .cfi_endproc # -- End function .globl _Z4miscv # -- Begin function _Z4miscv .p2align 4, 0x90 .type _Z4miscv,@function _Z4miscv: # @_Z4miscv .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 1520 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 leaq 12(%rsp), %rdi callq hipGetDeviceCount movl 12(%rsp), %esi movl $.L.str, %edi xorl %eax, %eax callq printf cmpl $0, 12(%rsp) jle .LBB1_7 # %bb.1: # %.lr.ph movl $1, %eax leaq 16(%rsp), %rbx xorl %ebp, %ebp jmp .LBB1_2 .p2align 4, 0x90 .LBB1_6: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.4, %edi movq %r14, %rsi movq %rbx, %rdx xorl %eax, %eax callq printf movl 336(%rsp), %edx movl $.L.str.5, %edi movq %r14, %rsi xorl %eax, %eax callq printf incl %ebp movl 12(%rsp), %eax cmpl %eax, %ebp jge .LBB1_7 .LBB1_2: # =>This Inner Loop Header: Depth=1 movl $.L.str.1, %r14d testl %eax, %eax je .LBB1_4 # %bb.3: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.2, %r14d .LBB1_4: # in Loop: Header=BB1_2 Depth=1 movq %rbx, %rdi movl %ebp, %esi callq hipGetDevicePropertiesR0600 cmpl $0, 12(%rsp) jle .LBB1_6 # %bb.5: # in Loop: Header=BB1_2 Depth=1 movl $.L.str.3, %edi movl %ebp, %esi xorl %eax, %eax callq printf jmp .LBB1_6 .LBB1_7: # %._crit_edge addq $1488, %rsp # imm = 0x5D0 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z4miscv, .Lfunc_end1-_Z4miscv .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI2_0: .quad 0x3f70000000000000 # double 0.00390625 .LCPI2_2: .quad 0x4008000000000000 # double 3 .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 .LCPI2_1: .long 0x40400000 # float 3 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $128, %rsp .cfi_def_cfa_offset 176 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq 8(%rsi), %rdi xorl %esi, %esi movl $10, %edx callq __isoc23_strtol movq %rax, %r12 leal (,%rax,4), %ebx callq _Z4miscv movslq %ebx, %r13 movq %r13, %rdi callq malloc movq %rax, %rbx movq %r13, %rdi callq malloc movq %rax, %r14 movq %r13, %rdi callq malloc movq %rax, %r15 testl %r12d, %r12d je .LBB2_3 # %bb.1: # %.lr.ph.preheader movl %r12d, %eax xorl %ecx, %ecx .p2align 4, 0x90 .LBB2_2: # %.lr.ph # =>This Inner Loop Header: Depth=1 xorps %xmm0, %xmm0 cvtsi2ss %ecx, %xmm0 movl %ecx, %edx imull %ecx, %edx xorps %xmm1, %xmm1 cvtsi2ss %edx, %xmm1 movss %xmm0, (%rbx,%rcx,4) movss %xmm1, (%r14,%rcx,4) incq %rcx cmpq %rcx, %rax jne .LBB2_2 .LBB2_3: # %._crit_edge leaq 16(%rsp), %rdi movq %r13, %rsi callq hipMalloc leaq 8(%rsp), %rdi movq %r13, %rsi callq hipMalloc movq 16(%rsp), %rdi movq %rbx, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi movq %r14, %rsi movq %r13, %rdx movl $1, %ecx callq hipMemcpy movl %r12d, %eax xorps %xmm0, %xmm0 cvtsi2sd %rax, %xmm0 mulsd .LCPI2_0(%rip), %xmm0 callq ceil@PLT cvttsd2si %xmm0, %rax movl %eax, %edi movabsq $4294967296, %rdx # imm = 0x100000000 orq %rdx, %rdi orq $256, %rdx # imm = 0x100 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB2_5 # %bb.4: movq 16(%rsp), %rax movq 8(%rsp), %rcx movl %r12d, 28(%rsp) movl $1077936128, 24(%rsp) # imm = 0x40400000 movq %rax, 88(%rsp) movq %rcx, 80(%rsp) leaq 28(%rsp), %rax movq %rax, 96(%rsp) leaq 24(%rsp), %rax movq %rax, 104(%rsp) leaq 88(%rsp), %rax movq %rax, 112(%rsp) leaq 80(%rsp), %rax movq %rax, 120(%rsp) leaq 64(%rsp), %rdi leaq 48(%rsp), %rsi leaq 40(%rsp), %rdx leaq 32(%rsp), %rcx callq __hipPopCallConfiguration movq 64(%rsp), %rsi movl 72(%rsp), %edx movq 48(%rsp), %rcx movl 56(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z5saxpyjfPfS_, %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 .LBB2_5: movq 8(%rsp), %rsi movq %r15, %rdi movq %r13, %rdx movl $2, %ecx callq hipMemcpy testl %r12d, %r12d je .LBB2_9 # %bb.6: # %.lr.ph60.preheader movl %r12d, %eax xorl %edx, %edx movss .LCPI2_1(%rip), %xmm0 # xmm0 = mem[0],zero,zero,zero .p2align 4, 0x90 .LBB2_7: # %.lr.ph60 # =>This Inner Loop Header: Depth=1 movss (%r15,%rdx,4), %xmm3 # xmm3 = mem[0],zero,zero,zero movss (%rbx,%rdx,4), %xmm1 # xmm1 = mem[0],zero,zero,zero movaps %xmm1, %xmm4 mulss %xmm0, %xmm4 movss (%r14,%rdx,4), %xmm2 # xmm2 = mem[0],zero,zero,zero addss %xmm2, %xmm4 ucomiss %xmm4, %xmm3 jne .LBB2_10 jp .LBB2_10 # %bb.8: # in Loop: Header=BB2_7 Depth=1 incq %rdx cmpq %rdx, %rax jne .LBB2_7 .LBB2_9: # %._crit_edge61 movl $.Lstr, %edi callq puts@PLT movq %rbx, %rdi callq free movq %r14, %rdi callq free movq %r15, %rdi callq free movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree xorl %edi, %edi callq exit .LBB2_10: movq stderr(%rip), %rdi cvtss2sd %xmm1, %xmm1 cvtss2sd %xmm2, %xmm2 cvtss2sd %xmm3, %xmm3 movsd .LCPI2_2(%rip), %xmm0 # xmm0 = mem[0],zero movl $.L.str.6, %esi # kill: def $edx killed $edx killed $rdx movb $4, %al callq fprintf movl $-1, %edi callq exit .Lfunc_end2: .size main, .Lfunc_end2-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z5saxpyjfPfS_, %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 _Z5saxpyjfPfS_,@object # @_Z5saxpyjfPfS_ .section .rodata,"a",@progbits .globl _Z5saxpyjfPfS_ .p2align 3, 0x0 _Z5saxpyjfPfS_: .quad _Z20__device_stub__saxpyjfPfS_ .size _Z5saxpyjfPfS_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "This machine has %d CUDA devices.\n" .size .L.str, 35 .type .L.str.1,@object # @.str.1 .L.str.1: .zero 1 .size .L.str.1, 1 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz " " .size .L.str.2, 3 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "Device %d:\n" .size .L.str.3, 12 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "%sdevice.name = %s\n" .size .L.str.4, 20 .type .L.str.5,@object # @.str.5 .L.str.5: .asciz "%sdevice.maxThreadsPerBlock = %d\n" .size .L.str.5, 34 .type .L.str.6,@object # @.str.6 .L.str.6: .asciz "ERROR: i=%d, a = %s, x[i]=%f, y[i]=%f, yy[i]=%f\n" .size .L.str.6, 49 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z5saxpyjfPfS_" .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 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "The results match!" .size .Lstr, 19 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z20__device_stub__saxpyjfPfS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z5saxpyjfPfS_ .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.
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <stdio.h> #include <stdlib.h> #define CUERR { cudaError_t err; \ if ((err = cudaGetLastError()) != cudaSuccess) { \ printf("CUDA error: %s, line %d\n", cudaGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } cudaMemcpyToSymbol(atominfo, atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
code for sm_80 Function : _Z7cenergyifPf .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2UR UR4, SR_CTAID.X ; /* 0x00000000000479c3 */ /* 0x000e220000002500 */ /*0020*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000e620000002200 */ /*0030*/ MOV R5, 0xffffff ; /* 0x00ffffff00057802 */ /* 0x000fe20000000f00 */ /*0040*/ ULDC.64 UR10, c[0x0][0x118] ; /* 0x00004600000a7ab9 */ /* 0x000fe20000000a00 */ /*0050*/ MOV R0, c[0x0][0x0] ; /* 0x0000000000007a02 */ /* 0x000fe20000000f00 */ /*0060*/ S2R R7, SR_TID.X ; /* 0x0000000000077919 */ /* 0x000ea20000002100 */ /*0070*/ MOV R11, c[0x0][0x160] ; /* 0x00005800000b7a02 */ /* 0x000fe20000000f00 */ /*0080*/ HFMA2.MMA R18, -RZ, RZ, 0, 0 ; /* 0x00000000ff127435 */ /* 0x000fe200000001ff */ /*0090*/ S2UR UR5, SR_CTAID.Y ; /* 0x00000000000579c3 */ /* 0x000ee20000002600 */ /*00a0*/ LOP3.LUT R2, R5.reuse, c[0x0][0x4], RZ, 0xc0, !PT ; /* 0x0000010005027a12 */ /* 0x040fe200078ec0ff */ /*00b0*/ CS2R R14, SRZ ; /* 0x00000000000e7805 */ /* 0x000fe2000001ff00 */ /*00c0*/ LOP3.LUT R5, R5, c[0x0][0xc], RZ, 0xc0, !PT ; /* 0x0000030005057a12 */ /* 0x000fe200078ec0ff */ /*00d0*/ CS2R R12, SRZ ; /* 0x00000000000c7805 */ /* 0x000fe2000001ff00 */ /*00e0*/ LOP3.LUT R0, R0, 0xffffff, RZ, 0xc0, !PT ; /* 0x00ffffff00007812 */ /* 0x000fc400078ec0ff */ /*00f0*/ ISETP.GE.AND P0, PT, R11, 0x1, PT ; /* 0x000000010b00780c */ /* 0x000fe40003f06270 */ /*0100*/ MOV R17, RZ ; /* 0x000000ff00117202 */ /* 0x000fe20000000f00 */ /*0110*/ IMAD R5, R0, R5, RZ ; /* 0x0000000500057224 */ /* 0x000fca00078e02ff */ /*0120*/ SHF.L.U32 R5, R5, 0x3, RZ ; /* 0x0000000305057819 */ /* 0x000fe200000006ff */ /*0130*/ USHF.L.U32 UR4, UR4, 0x3, URZ ; /* 0x0000000304047899 */ /* 0x001fc8000800063f */ /*0140*/ ULOP3.LUT UR4, UR4, 0x7fffff8, URZ, 0xc0, !UPT ; /* 0x07fffff804047892 */ /* 0x000fe4000f8ec03f */ /*0150*/ ULOP3.LUT UR5, UR5, 0xffffff, URZ, 0xc0, !UPT ; /* 0x00ffffff05057892 */ /* 0x008fc8000f8ec03f */ /*0160*/ IMAD R0, R0, UR4, R7 ; /* 0x0000000400007c24 */ /* 0x004fe4000f8e0207 */ /*0170*/ CS2R R6, SRZ ; /* 0x0000000000067805 */ /* 0x000fe2000001ff00 */ /*0180*/ IMAD R3, R2, UR5, R3 ; /* 0x0000000502037c24 */ /* 0x002fc8000f8e0203 */ /*0190*/ IMAD R8, R3, R5, R0 ; /* 0x0000000503087224 */ /* 0x000fe200078e0200 */ /*01a0*/ @!P0 BRA 0xea0 ; /* 0x00000cf000008947 */ /* 0x000fea0003800000 */ /*01b0*/ ISETP.NE.AND P0, PT, R11.reuse, 0x1, PT ; /* 0x000000010b00780c */ /* 0x040fe20003f05270 */ /*01c0*/ I2F.U32 R10, R3 ; /* 0x00000003000a7306 */ /* 0x0000620000201000 */ /*01d0*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe20008000000 */ /*01e0*/ LOP3.LUT R11, R11, 0x1, RZ, 0xc0, !PT ; /* 0x000000010b0b7812 */ /* 0x000fe200078ec0ff */ /*01f0*/ CS2R R6, SRZ ; /* 0x0000000000067805 */ /* 0x000fe2000001ff00 */ /*0200*/ CS2R R12, SRZ ; /* 0x00000000000c7805 */ /* 0x000fe2000001ff00 */ /*0210*/ CS2R R14, SRZ ; /* 0x00000000000e7805 */ /* 0x000fe2000001ff00 */ /*0220*/ MOV R17, RZ ; /* 0x000000ff00117202 */ /* 0x000fe40000000f00 */ /*0230*/ I2F.U32 R9, R0 ; /* 0x0000000000097306 */ /* 0x0000a20000201000 */ /*0240*/ MOV R18, RZ ; /* 0x000000ff00127202 */ /* 0x000fc60000000f00 */ /*0250*/ @!P0 BRA 0xa90 ; /* 0x0000083000008947 */ /* 0x000fea0003800000 */ /*0260*/ HFMA2.MMA R7, -RZ, RZ, 0, 0 ; /* 0x00000000ff077435 */ /* 0x000fe200000001ff */ /*0270*/ IADD3 R0, -R11, c[0x0][0x160], RZ ; /* 0x000058000b007a10 */ /* 0x001fe20007ffe1ff */ /*0280*/ UMOV UR4, URZ ; /* 0x0000003f00047c82 */ /* 0x000fe40008000000 */ /*0290*/ USHF.L.U32 UR5, UR4, 0x4, URZ ; /* 0x0000000404057899 */ /* 0x000fe2000800063f */ /*02a0*/ MOV R16, c[0x0][0x164] ; /* 0x0000590000107a02 */ /* 0x000fe20000000f00 */ /*02b0*/ UIADD3 UR4, UR4, 0x2, URZ ; /* 0x0000000204047890 */ /* 0x000fe2000fffe03f */ /*02c0*/ IADD3 R0, R0, -0x2, RZ ; /* 0xfffffffe00007810 */ /* 0x000fc60007ffe0ff */ /*02d0*/ MOV R20, UR5 ; /* 0x0000000500147c02 */ /* 0x000fe40008000f00 */ /*02e0*/ MOV R22, UR5 ; /* 0x0000000500167c02 */ /* 0x000fe40008000f00 */ /*02f0*/ LDC.64 R4, c[0x3][R20] ; /* 0x00c0000014047b82 */ /* 0x000e240000000a00 */ /*0300*/ ULDC.64 UR8, c[0x3][UR5+0x8] ; /* 0x00c0020005087abb */ /* 0x000fe40008000a00 */ /*0310*/ ULDC.64 UR6, c[0x3][UR5+0x18] ; /* 0x00c0060005067abb */ /* 0x000fc80008000a00 */ /*0320*/ LDC.64 R2, c[0x3][R22+0x10] ; /* 0x00c0040016027b82 */ /* 0x000ee20000000a00 */ /*0330*/ FFMA R5, R10, c[0x0][0x164], -R5 ; /* 0x000059000a057a23 */ /* 0x003fe40000000805 */ /*0340*/ FFMA R4, R9, c[0x0][0x164], -R4 ; /* 0x0000590009047a23 */ /* 0x004fe40000000804 */ /*0350*/ FFMA R21, R5, R5, UR8 ; /* 0x0000000805157e23 */ /* 0x000fc80008000005 */ /*0360*/ FFMA R5, R4, R4, R21 ; /* 0x0000000404057223 */ /* 0x000fe40000000015 */ /*0370*/ FFMA R4, R16, 16, R4 ; /* 0x4180000010047823 */ /* 0x000fe40000000004 */ /*0380*/ FFMA R3, R10, c[0x0][0x164], -R3 ; /* 0x000059000a037a23 */ /* 0x008fe20000000803 */ /*0390*/ FSETP.GEU.AND P0, PT, |R5|, 1.175494350822287508e-38, PT ; /* 0x008000000500780b */ /* 0x000fe20003f0e200 */ /*03a0*/ FFMA R26, R4, R4, R21 ; /* 0x00000004041a7223 */ /* 0x000fe40000000015 */ /*03b0*/ FFMA R4, R16, 16, R4 ; /* 0x4180000010047823 */ /* 0x000fe40000000004 */ /*03c0*/ FFMA R19, R3, R3, UR6 ; /* 0x0000000603137e23 */ /* 0x000fe20008000003 */ /*03d0*/ FSETP.GEU.AND P3, PT, |R26|, 1.175494350822287508e-38, PT ; /* 0x008000001a00780b */ /* 0x000fe20003f6e200 */ /*03e0*/ FFMA R3, R16, 16, R4 ; /* 0x4180000010037823 */ /* 0x000fc40000000004 */ /*03f0*/ FFMA R23, R4, R4, R21.reuse ; /* 0x0000000404177223 */ /* 0x100fe40000000015 */ /*0400*/ FFMA R24, R3, R3, R21 ; /* 0x0000000303187223 */ /* 0x000fe40000000015 */ /*0410*/ @!P0 FMUL R5, R5, 16777216 ; /* 0x4b80000005058820 */ /* 0x000fe20000400000 */ /*0420*/ FSETP.GEU.AND P4, PT, |R23|, 1.175494350822287508e-38, PT ; /* 0x008000001700780b */ /* 0x000fe20003f8e200 */ /*0430*/ FFMA R3, R16, 16, R3 ; /* 0x4180000010037823 */ /* 0x000fe20000000003 */ /*0440*/ FSETP.GEU.AND P2, PT, |R24|, 1.175494350822287508e-38, PT ; /* 0x008000001800780b */ /* 0x000fe20003f4e200 */ /*0450*/ MUFU.RSQ R25, R5 ; /* 0x0000000500197308 */ /* 0x000e220000001400 */ /*0460*/ FFMA R2, R9, c[0x0][0x164], -R2 ; /* 0x0000590009027a23 */ /* 0x000fe40000000802 */ /*0470*/ @!P3 FMUL R26, R26, 16777216 ; /* 0x4b8000001a1ab820 */ /* 0x000fc40000400000 */ /*0480*/ FFMA R20, R16, 16, R3 ; /* 0x4180000010147823 */ /* 0x000fe40000000003 */ /*0490*/ FFMA R27, R2, R2, R19.reuse ; /* 0x00000002021b7223 */ /* 0x100fe40000000013 */ /*04a0*/ MUFU.RSQ R26, R26 ; /* 0x0000001a001a7308 */ /* 0x000e620000001400 */ /*04b0*/ @!P4 FMUL R23, R23, 16777216 ; /* 0x4b8000001717c820 */ /* 0x000fe40000400000 */ /*04c0*/ @!P2 FMUL R24, R24, 16777216 ; /* 0x4b8000001818a820 */ /* 0x000fe20000400000 */ /*04d0*/ FSETP.GEU.AND P1, PT, |R27|, 1.175494350822287508e-38, PT ; /* 0x008000001b00780b */ /* 0x000fe20003f2e200 */ /*04e0*/ FFMA R22, R16, 16, R2 ; /* 0x4180000010167823 */ /* 0x000fe40000000002 */ /*04f0*/ @!P0 FMUL R25, R25, 4096 ; /* 0x4580000019198820 */ /* 0x001fe20000400000 */ /*0500*/ MUFU.RSQ R23, R23 ; /* 0x0000001700177308 */ /* 0x000fe20000001400 */ /*0510*/ FFMA R4, R22, R22, R19 ; /* 0x0000001616047223 */ /* 0x000fc40000000013 */ /*0520*/ FFMA R5, R25, UR9, R18 ; /* 0x0000000919057c23 */ /* 0x000fe40008000012 */ /*0530*/ FFMA R25, R16, 16, R20 ; /* 0x4180000010197823 */ /* 0x000fe20000000014 */ /*0540*/ FSETP.GEU.AND P0, PT, |R4|, 1.175494350822287508e-38, PT ; /* 0x008000000400780b */ /* 0x000fe20003f0e200 */ /*0550*/ FFMA R18, R3, R3, R21 ; /* 0x0000000303127223 */ /* 0x000fe20000000015 */ /*0560*/ MUFU.RSQ R24, R24 ; /* 0x0000001800187308 */ /* 0x000e220000001400 */ /*0570*/ @!P3 FMUL R26, R26, 4096 ; /* 0x458000001a1ab820 */ /* 0x002fe40000400000 */ /*0580*/ @!P1 FMUL R27, R27, 16777216 ; /* 0x4b8000001b1b9820 */ /* 0x000fe20000400000 */ /*0590*/ FSETP.GEU.AND P3, PT, |R18|, 1.175494350822287508e-38, PT ; /* 0x008000001200780b */ /* 0x000fe20003f6e200 */ /*05a0*/ FFMA R3, R26, UR9, R17 ; /* 0x000000091a037c23 */ /* 0x000fc40008000011 */ /*05b0*/ FFMA R26, R16, 16, R25 ; /* 0x41800000101a7823 */ /* 0x000fe20000000019 */ /*05c0*/ MUFU.RSQ R2, R27 ; /* 0x0000001b00027308 */ /* 0x000e620000001400 */ /*05d0*/ FFMA R17, R20, R20, R21.reuse ; /* 0x0000001414117223 */ /* 0x100fe40000000015 */ /*05e0*/ FFMA R20, R25, R25, R21.reuse ; /* 0x0000001919147223 */ /* 0x100fe40000000015 */ /*05f0*/ FFMA R21, R26, R26, R21 ; /* 0x0000001a1a157223 */ /* 0x000fe20000000015 */ /*0600*/ FSETP.GEU.AND P6, PT, |R17|, 1.175494350822287508e-38, PT ; /* 0x008000001100780b */ /* 0x000fe20003fce200 */ /*0610*/ FFMA R25, R16, 16, R22 ; /* 0x4180000010197823 */ /* 0x000fe20000000016 */ /*0620*/ FSETP.GEU.AND P5, PT, |R20|, 1.175494350822287508e-38, PT ; /* 0x008000001400780b */ /* 0x000fe20003fae200 */ /*0630*/ @!P2 FMUL R24, R24, 4096 ; /* 0x458000001818a820 */ /* 0x001fe20000400000 */ /*0640*/ FSETP.GEU.AND P2, PT, |R21|, 1.175494350822287508e-38, PT ; /* 0x008000001500780b */ /* 0x000fe20003f4e200 */ /*0650*/ FFMA R22, R25, R25, R19 ; /* 0x0000001919167223 */ /* 0x000fc40000000013 */ /*0660*/ FFMA R25, R16, 16, R25 ; /* 0x4180000010197823 */ /* 0x000fe40000000019 */ /*0670*/ FFMA R15, R24, UR9, R15 ; /* 0x00000009180f7c23 */ /* 0x000fe4000800000f */ /*0680*/ @!P3 FMUL R18, R18, 16777216 ; /* 0x4b8000001212b820 */ /* 0x000fe40000400000 */ /*0690*/ FFMA R24, R16, 16, R25 ; /* 0x4180000010187823 */ /* 0x000fe40000000019 */ /*06a0*/ @!P4 FMUL R23, R23, 4096 ; /* 0x458000001717c820 */ /* 0x000fe20000400000 */ /*06b0*/ FSETP.GEU.AND P4, PT, |R22|, 1.175494350822287508e-38, PT ; /* 0x008000001600780b */ /* 0x000fe20003f8e200 */ /*06c0*/ @!P6 FMUL R17, R17, 16777216 ; /* 0x4b8000001111e820 */ /* 0x000fe20000400000 */ /*06d0*/ MUFU.RSQ R18, R18 ; /* 0x0000001200127308 */ /* 0x000e220000001400 */ /*06e0*/ @!P2 FMUL R21, R21, 16777216 ; /* 0x4b8000001515a820 */ /* 0x000fc40000400000 */ /*06f0*/ @!P5 FMUL R20, R20, 16777216 ; /* 0x4b8000001414d820 */ /* 0x000fe40000400000 */ /*0700*/ FFMA R28, R24, R24, R19.reuse ; /* 0x00000018181c7223 */ /* 0x100fe40000000013 */ /*0710*/ FFMA R29, R16, 16, R24 ; /* 0x41800000101d7823 */ /* 0x000fe20000000018 */ /*0720*/ MUFU.RSQ R17, R17 ; /* 0x0000001100117308 */ /* 0x000ea20000001400 */ /*0730*/ FFMA R14, R23, UR9, R14 ; /* 0x00000009170e7c23 */ /* 0x000fe4000800000e */ /*0740*/ FFMA R26, R25, R25, R19 ; /* 0x00000019191a7223 */ /* 0x000fe40000000013 */ /*0750*/ @!P1 FMUL R2, R2, 4096 ; /* 0x4580000002029820 */ /* 0x002fc40000400000 */ /*0760*/ @!P0 FMUL R4, R4, 16777216 ; /* 0x4b80000004048820 */ /* 0x000fe20000400000 */ /*0770*/ MUFU.RSQ R24, R21 ; /* 0x0000001500187308 */ /* 0x000e620000001400 */ /*0780*/ @!P3 FMUL R18, R18, 4096 ; /* 0x458000001212b820 */ /* 0x001fe20000400000 */ /*0790*/ FSETP.GEU.AND P3, PT, |R26|, 1.175494350822287508e-38, PT ; /* 0x008000001a00780b */ /* 0x000fe20003f6e200 */ /*07a0*/ @!P4 FMUL R22, R22, 16777216 ; /* 0x4b8000001616c820 */ /* 0x000fe40000400000 */ /*07b0*/ FFMA R13, R18, UR9, R13 ; /* 0x00000009120d7c23 */ /* 0x000fe4000800000d */ /*07c0*/ FFMA R18, R2, UR7, R5 ; /* 0x0000000702127c23 */ /* 0x000fe20008000005 */ /*07d0*/ MUFU.RSQ R23, R20 ; /* 0x0000001400177308 */ /* 0x0000e20000001400 */ /*07e0*/ @!P6 FMUL R17, R17, 4096 ; /* 0x458000001111e820 */ /* 0x004fe20000400000 */ /*07f0*/ FSETP.GEU.AND P6, PT, |R28|, 1.175494350822287508e-38, PT ; /* 0x008000001c00780b */ /* 0x000fc60003fce200 */ /*0800*/ FFMA R6, R17, UR9, R6 ; /* 0x0000000911067c23 */ /* 0x000fe40008000006 */ /*0810*/ @!P3 FMUL R26, R26, 16777216 ; /* 0x4b8000001a1ab820 */ /* 0x000fe20000400000 */ /*0820*/ MUFU.RSQ R4, R4 ; /* 0x0000000400047308 */ /* 0x000ea20000001400 */ /*0830*/ FFMA R20, R16.reuse, 16, R29 ; /* 0x4180000010147823 */ /* 0x041fe4000000001d */ /*0840*/ FFMA R29, R29, R29, R19.reuse ; /* 0x0000001d1d1d7223 */ /* 0x100fe40000000013 */ /*0850*/ FFMA R16, R16, 16, R20 ; /* 0x4180000010107823 */ /* 0x000fe40000000014 */ /*0860*/ FFMA R20, R20, R20, R19.reuse ; /* 0x0000001414147223 */ /* 0x100fe20000000013 */ /*0870*/ MUFU.RSQ R25, R22 ; /* 0x0000001600197308 */ /* 0x000e220000001400 */ /*0880*/ FFMA R27, R16, R16, R19 ; /* 0x00000010101b7223 */ /* 0x000fc40000000013 */ /*0890*/ @!P2 FMUL R24, R24, 4096 ; /* 0x458000001818a820 */ /* 0x002fe20000400000 */ /*08a0*/ FSETP.GEU.AND P2, PT, |R20|, 1.175494350822287508e-38, PT ; /* 0x008000001400780b */ /* 0x000fe20003f4e200 */ /*08b0*/ @!P5 FMUL R23, R23, 4096 ; /* 0x458000001717d820 */ /* 0x008fe20000400000 */ /*08c0*/ FSETP.GEU.AND P1, PT, |R27|, 1.175494350822287508e-38, PT ; /* 0x008000001b00780b */ /* 0x000fe20003f2e200 */ /*08d0*/ @!P6 FMUL R28, R28, 16777216 ; /* 0x4b8000001c1ce820 */ /* 0x000fe20000400000 */ /*08e0*/ FSETP.GEU.AND P5, PT, |R29|, 1.175494350822287508e-38, PT ; /* 0x008000001d00780b */ /* 0x000fe20003fae200 */ /*08f0*/ MUFU.RSQ R26, R26 ; /* 0x0000001a001a7308 */ /* 0x000e620000001400 */ /*0900*/ @!P0 FMUL R4, R4, 4096 ; /* 0x4580000004048820 */ /* 0x004fe20000400000 */ /*0910*/ ISETP.NE.AND P0, PT, R0, RZ, PT ; /* 0x000000ff0000720c */ /* 0x000fe20003f05270 */ /*0920*/ FFMA R23, R23, UR9, R12 ; /* 0x0000000917177c23 */ /* 0x000fe4000800000c */ /*0930*/ FFMA R24, R24, UR9, R7 ; /* 0x0000000918187c23 */ /* 0x000fc40008000007 */ /*0940*/ @!P4 FMUL R25, R25, 4096 ; /* 0x458000001919c820 */ /* 0x001fe20000400000 */ /*0950*/ MUFU.RSQ R16, R28 ; /* 0x0000001c00107308 */ /* 0x000e220000001400 */ /*0960*/ @!P2 FMUL R20, R20, 16777216 ; /* 0x4b8000001414a820 */ /* 0x000fe40000400000 */ /*0970*/ @!P1 FMUL R27, R27, 16777216 ; /* 0x4b8000001b1b9820 */ /* 0x000fe40000400000 */ /*0980*/ @!P5 FMUL R29, R29, 16777216 ; /* 0x4b8000001d1dd820 */ /* 0x000fe40000400000 */ /*0990*/ FFMA R17, R4, UR7, R3 ; /* 0x0000000704117c23 */ /* 0x000fe20008000003 */ /*09a0*/ MUFU.RSQ R20, R20 ; /* 0x0000001400147308 */ /* 0x000ea20000001400 */ /*09b0*/ @!P3 FMUL R26, R26, 4096 ; /* 0x458000001a1ab820 */ /* 0x002fc40000400000 */ /*09c0*/ FFMA R14, R25, UR7, R14 ; /* 0x00000007190e7c23 */ /* 0x000fe4000800000e */ /*09d0*/ FFMA R15, R26, UR7, R15 ; /* 0x000000071a0f7c23 */ /* 0x000fc6000800000f */ /*09e0*/ MUFU.RSQ R19, R29 ; /* 0x0000001d00137308 */ /* 0x000e620000001400 */ /*09f0*/ @!P6 FMUL R16, R16, 4096 ; /* 0x458000001010e820 */ /* 0x001fc80000400000 */ /*0a00*/ FFMA R13, R16, UR7, R13 ; /* 0x00000007100d7c23 */ /* 0x000fc6000800000d */ /*0a10*/ MUFU.RSQ R27, R27 ; /* 0x0000001b001b7308 */ /* 0x000e220000001400 */ /*0a20*/ @!P2 FMUL R20, R20, 4096 ; /* 0x458000001414a820 */ /* 0x004fc80000400000 */ /*0a30*/ FFMA R12, R20, UR7, R23 ; /* 0x00000007140c7c23 */ /* 0x000fe40008000017 */ /*0a40*/ @!P5 FMUL R19, R19, 4096 ; /* 0x458000001313d820 */ /* 0x002fc80000400000 */ /*0a50*/ FFMA R6, R19, UR7, R6 ; /* 0x0000000713067c23 */ /* 0x000fe40008000006 */ /*0a60*/ @!P1 FMUL R27, R27, 4096 ; /* 0x458000001b1b9820 */ /* 0x001fc80000400000 */ /*0a70*/ FFMA R7, R27, UR7, R24 ; /* 0x000000071b077c23 */ /* 0x000fe20008000018 */ /*0a80*/ @P0 BRA 0x290 ; /* 0xfffff80000000947 */ /* 0x000fea000383ffff */ /*0a90*/ ISETP.NE.AND P0, PT, R11, RZ, PT ; /* 0x000000ff0b00720c */ /* 0x000fda0003f05270 */ /*0aa0*/ @!P0 BRA 0xea0 ; /* 0x000003f000008947 */ /* 0x000fea0003800000 */ /*0ab0*/ USHF.L.U32 UR4, UR4, 0x4, URZ ; /* 0x0000000404047899 */ /* 0x000fe2000800063f */ /*0ac0*/ MOV R20, c[0x0][0x164] ; /* 0x0000590000147a02 */ /* 0x000fca0000000f00 */ /*0ad0*/ MOV R5, UR4 ; /* 0x0000000400057c02 */ /* 0x000fc80008000f00 */ /*0ae0*/ LDC.64 R2, c[0x3][R5] ; /* 0x00c0000005027b82 */ /* 0x001e240000000a00 */ /*0af0*/ ULDC.64 UR4, c[0x3][UR4+0x8] ; /* 0x00c0020004047abb */ /* 0x000fe20008000a00 */ /*0b00*/ FFMA R3, R10, c[0x0][0x164], -R3 ; /* 0x000059000a037a23 */ /* 0x003fe40000000803 */ /*0b10*/ FFMA R9, R9, c[0x0][0x164], -R2 ; /* 0x0000590009097a23 */ /* 0x004fe40000000802 */ /*0b20*/ FFMA R0, R3, R3, UR4 ; /* 0x0000000403007e23 */ /* 0x000fc80008000003 */ /*0b30*/ FFMA R3, R9, R9, R0 ; /* 0x0000000909037223 */ /* 0x000fe40000000000 */ /*0b40*/ FFMA R9, R20, 16, R9 ; /* 0x4180000014097823 */ /* 0x000fc60000000009 */ /*0b50*/ FSETP.GEU.AND P6, PT, |R3|, 1.175494350822287508e-38, PT ; /* 0x008000000300780b */ /* 0x000fe20003fce200 */ /*0b60*/ FFMA R2, R9, R9, R0 ; /* 0x0000000909027223 */ /* 0x000fe40000000000 */ /*0b70*/ FFMA R9, R20, 16, R9 ; /* 0x4180000014097823 */ /* 0x000fc60000000009 */ /*0b80*/ FSETP.GEU.AND P5, PT, |R2|, 1.175494350822287508e-38, PT ; /* 0x008000000200780b */ /* 0x000fe20003fae200 */ /*0b90*/ FFMA R4, R9, R9, R0 ; /* 0x0000000909047223 */ /* 0x000fe40000000000 */ /*0ba0*/ FFMA R9, R20, 16, R9 ; /* 0x4180000014097823 */ /* 0x000fc60000000009 */ /*0bb0*/ FSETP.GEU.AND P4, PT, |R4|, 1.175494350822287508e-38, PT ; /* 0x008000000400780b */ /* 0x000fe20003f8e200 */ /*0bc0*/ @!P6 FMUL R3, R3, 16777216 ; /* 0x4b8000000303e820 */ /* 0x000fe40000400000 */ /*0bd0*/ FFMA R10, R9, R9, R0.reuse ; /* 0x00000009090a7223 */ /* 0x100fe40000000000 */ /*0be0*/ FFMA R9, R20, 16, R9 ; /* 0x4180000014097823 */ /* 0x000fe40000000009 */ /*0bf0*/ MUFU.RSQ R3, R3 ; /* 0x0000000300037308 */ /* 0x000e220000001400 */ /*0c00*/ FSETP.GEU.AND P3, PT, |R10|, 1.175494350822287508e-38, PT ; /* 0x008000000a00780b */ /* 0x000fe20003f6e200 */ /*0c10*/ FFMA R16, R9, R9, R0 ; /* 0x0000000909107223 */ /* 0x000fe40000000000 */ /*0c20*/ FFMA R9, R20, 16, R9 ; /* 0x4180000014097823 */ /* 0x000fc40000000009 */ /*0c30*/ @!P5 FMUL R2, R2, 16777216 ; /* 0x4b8000000202d820 */ /* 0x000fe20000400000 */ /*0c40*/ FSETP.GEU.AND P2, PT, |R16|, 1.175494350822287508e-38, PT ; /* 0x008000001000780b */ /* 0x000fe20003f4e200 */ /*0c50*/ FFMA R5, R20.reuse, 16, R9 ; /* 0x4180000014057823 */ /* 0x040fe40000000009 */ /*0c60*/ FFMA R19, R9, R9, R0.reuse ; /* 0x0000000909137223 */ /* 0x100fe40000000000 */ /*0c70*/ FFMA R9, R20, 16, R5 ; /* 0x4180000014097823 */ /* 0x000fe20000000005 */ /*0c80*/ MUFU.RSQ R2, R2 ; /* 0x0000000200027308 */ /* 0x000e620000001400 */ /*0c90*/ FFMA R5, R5, R5, R0.reuse ; /* 0x0000000505057223 */ /* 0x100fe20000000000 */ /*0ca0*/ FSETP.GEU.AND P1, PT, |R19|, 1.175494350822287508e-38, PT ; /* 0x008000001300780b */ /* 0x000fe20003f2e200 */ /*0cb0*/ FFMA R0, R9, R9, R0 ; /* 0x0000000909007223 */ /* 0x000fc40000000000 */ /*0cc0*/ @!P6 FMUL R3, R3, 4096 ; /* 0x458000000303e820 */ /* 0x001fe20000400000 */ /*0cd0*/ FSETP.GEU.AND P0, PT, |R5|, 1.175494350822287508e-38, PT ; /* 0x008000000500780b */ /* 0x000fe20003f0e200 */ /*0ce0*/ @!P3 FMUL R10, R10, 16777216 ; /* 0x4b8000000a0ab820 */ /* 0x000fe20000400000 */ /*0cf0*/ FSETP.GEU.AND P6, PT, |R0|, 1.175494350822287508e-38, PT ; /* 0x008000000000780b */ /* 0x000fe20003fce200 */ /*0d00*/ @!P2 FMUL R16, R16, 16777216 ; /* 0x4b8000001010a820 */ /* 0x000fe40000400000 */ /*0d10*/ @!P4 FMUL R4, R4, 16777216 ; /* 0x4b8000000404c820 */ /* 0x000fe40000400000 */ /*0d20*/ MUFU.RSQ R10, R10 ; /* 0x0000000a000a7308 */ /* 0x000e220000001400 */ /*0d30*/ FFMA R18, R3, UR5, R18 ; /* 0x0000000503127c23 */ /* 0x000fe40008000012 */ /*0d40*/ @!P1 FMUL R19, R19, 16777216 ; /* 0x4b80000013139820 */ /* 0x000fc40000400000 */ /*0d50*/ @!P5 FMUL R2, R2, 4096 ; /* 0x458000000202d820 */ /* 0x002fe40000400000 */ /*0d60*/ @!P0 FMUL R5, R5, 16777216 ; /* 0x4b80000005058820 */ /* 0x000fe20000400000 */ /*0d70*/ MUFU.RSQ R11, R4 ; /* 0x00000004000b7308 */ /* 0x000e620000001400 */ /*0d80*/ @!P6 FMUL R0, R0, 16777216 ; /* 0x4b8000000000e820 */ /* 0x000fe40000400000 */ /*0d90*/ FFMA R17, R2, UR5, R17 ; /* 0x0000000502117c23 */ /* 0x000fca0008000011 */ /*0da0*/ MUFU.RSQ R16, R16 ; /* 0x0000001000107308 */ /* 0x000ea20000001400 */ /*0db0*/ @!P3 FMUL R10, R10, 4096 ; /* 0x458000000a0ab820 */ /* 0x001fc80000400000 */ /*0dc0*/ FFMA R15, R10, UR5, R15 ; /* 0x000000050a0f7c23 */ /* 0x000fc6000800000f */ /*0dd0*/ MUFU.RSQ R19, R19 ; /* 0x0000001300137308 */ /* 0x000e220000001400 */ /*0de0*/ @!P4 FMUL R11, R11, 4096 ; /* 0x458000000b0bc820 */ /* 0x002fc80000400000 */ /*0df0*/ FFMA R14, R11, UR5, R14 ; /* 0x000000050b0e7c23 */ /* 0x000fc6000800000e */ /*0e00*/ MUFU.RSQ R5, R5 ; /* 0x0000000500057308 */ /* 0x000e620000001400 */ /*0e10*/ @!P2 FMUL R16, R16, 4096 ; /* 0x458000001010a820 */ /* 0x004fc80000400000 */ /*0e20*/ FFMA R13, R16, UR5, R13 ; /* 0x00000005100d7c23 */ /* 0x000fc6000800000d */ /*0e30*/ MUFU.RSQ R0, R0 ; /* 0x0000000000007308 */ /* 0x000ea20000001400 */ /*0e40*/ @!P1 FMUL R19, R19, 4096 ; /* 0x4580000013139820 */ /* 0x001fc80000400000 */ /*0e50*/ FFMA R6, R19, UR5, R6 ; /* 0x0000000513067c23 */ /* 0x000fe40008000006 */ /*0e60*/ @!P0 FMUL R5, R5, 4096 ; /* 0x4580000005058820 */ /* 0x002fc80000400000 */ /*0e70*/ FFMA R12, R5, UR5, R12 ; /* 0x00000005050c7c23 */ /* 0x000fe4000800000c */ /*0e80*/ @!P6 FMUL R0, R0, 4096 ; /* 0x458000000000e820 */ /* 0x004fc80000400000 */ /*0e90*/ FFMA R7, R0, UR5, R7 ; /* 0x0000000500077c23 */ /* 0x000fe40008000007 */ /*0ea0*/ MOV R9, 0x4 ; /* 0x0000000400097802 */ /* 0x004fca0000000f00 */ /*0eb0*/ IMAD.WIDE.U32 R2, R8, R9, c[0x0][0x168] ; /* 0x00005a0008027625 */ /* 0x001fca00078e0009 */ /*0ec0*/ LDG.E R5, [R2.64] ; /* 0x0000000a02057981 */ /* 0x000ea2000c1e1900 */ /*0ed0*/ IADD3 R4, R8, 0x10, RZ ; /* 0x0000001008047810 */ /* 0x000fe20007ffe0ff */ /*0ee0*/ FADD R19, R5, R18 ; /* 0x0000001205137221 */ /* 0x004fc80000000000 */ /*0ef0*/ IMAD.WIDE.U32 R4, R4, R9, c[0x0][0x168] ; /* 0x00005a0004047625 */ /* 0x000fe200078e0009 */ /*0f00*/ STG.E [R2.64], R19 ; /* 0x0000001302007986 */ /* 0x0001e8000c10190a */ /*0f10*/ LDG.E R0, [R4.64] ; /* 0x0000000a04007981 */ /* 0x000ea2000c1e1900 */ /*0f20*/ IADD3 R10, R8, 0x20, RZ ; /* 0x00000020080a7810 */ /* 0x002fca0007ffe0ff */ /*0f30*/ IMAD.WIDE.U32 R10, R10, R9, c[0x0][0x168] ; /* 0x00005a000a0a7625 */ /* 0x000fc800078e0009 */ /*0f40*/ FADD R21, R0, R17 ; /* 0x0000001100157221 */ /* 0x004fca0000000000 */ /*0f50*/ STG.E [R4.64], R21 ; /* 0x0000001504007986 */ /* 0x0003e8000c10190a */ /*0f60*/ LDG.E R17, [R10.64] ; /* 0x0000000a0a117981 */ /* 0x000ea2000c1e1900 */ /*0f70*/ IADD3 R16, R8, 0x30, RZ ; /* 0x0000003008107810 */ /* 0x000fe20007ffe0ff */ /*0f80*/ FADD R23, R17, R14 ; /* 0x0000000e11177221 */ /* 0x004fc80000000000 */ /*0f90*/ IMAD.WIDE.U32 R16, R16, R9, c[0x0][0x168] ; /* 0x00005a0010107625 */ /* 0x000fe200078e0009 */ /*0fa0*/ STG.E [R10.64], R23 ; /* 0x000000170a007986 */ /* 0x0005e8000c10190a */ /*0fb0*/ LDG.E R0, [R16.64] ; /* 0x0000000a10007981 */ /* 0x000ee2000c1e1900 */ /*0fc0*/ IADD3 R2, R8, 0x40, RZ ; /* 0x0000004008027810 */ /* 0x001fca0007ffe0ff */ /*0fd0*/ IMAD.WIDE.U32 R2, R2, R9, c[0x0][0x168] ; /* 0x00005a0002027625 */ /* 0x000fc800078e0009 */ /*0fe0*/ FADD R15, R0, R15 ; /* 0x0000000f000f7221 */ /* 0x008fca0000000000 */ /*0ff0*/ STG.E [R16.64], R15 ; /* 0x0000000f10007986 */ /* 0x0001e8000c10190a */ /*1000*/ LDG.E R0, [R2.64] ; /* 0x0000000a02007981 */ /* 0x000ee2000c1e1900 */ /*1010*/ IADD3 R4, R8, 0x50, RZ ; /* 0x0000005008047810 */ /* 0x002fca0007ffe0ff */ /*1020*/ IMAD.WIDE.U32 R4, R4, R9, c[0x0][0x168] ; /* 0x00005a0004047625 */ /* 0x000fc800078e0009 */ /*1030*/ FADD R13, R0, R13 ; /* 0x0000000d000d7221 */ /* 0x008fca0000000000 */ /*1040*/ STG.E [R2.64], R13 ; /* 0x0000000d02007986 */ /* 0x000fe8000c10190a */ /*1050*/ LDG.E R11, [R4.64] ; /* 0x0000000a040b7981 */ /* 0x004ea2000c1e1900 */ /*1060*/ IADD3 R10, R8, 0x60, RZ ; /* 0x00000060080a7810 */ /* 0x000fe20007ffe0ff */ /*1070*/ FADD R19, R11, R6 ; /* 0x000000060b137221 */ /* 0x004fc80000000000 */ /*1080*/ IMAD.WIDE.U32 R10, R10, R9, c[0x0][0x168] ; /* 0x00005a000a0a7625 */ /* 0x000fe200078e0009 */ /*1090*/ STG.E [R4.64], R19 ; /* 0x0000001304007986 */ /* 0x000fe8000c10190a */ /*10a0*/ LDG.E R15, [R10.64] ; /* 0x0000000a0a0f7981 */ /* 0x001ea2000c1e1900 */ /*10b0*/ IADD3 R8, R8, 0x70, RZ ; /* 0x0000007008087810 */ /* 0x000fca0007ffe0ff */ /*10c0*/ IMAD.WIDE.U32 R8, R8, R9, c[0x0][0x168] ; /* 0x00005a0008087625 */ /* 0x000fc800078e0009 */ /*10d0*/ FADD R15, R15, R12 ; /* 0x0000000c0f0f7221 */ /* 0x004fca0000000000 */ /*10e0*/ STG.E [R10.64], R15 ; /* 0x0000000f0a007986 */ /* 0x000fe8000c10190a */ /*10f0*/ LDG.E R0, [R8.64] ; /* 0x0000000a08007981 */ /* 0x000ea4000c1e1900 */ /*1100*/ FADD R7, R0, R7 ; /* 0x0000000700077221 */ /* 0x004fca0000000000 */ /*1110*/ STG.E [R8.64], R7 ; /* 0x0000000708007986 */ /* 0x000fe2000c10190a */ /*1120*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*1130*/ BRA 0x1130; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*1140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*1150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*1160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*1170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*1180*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*1190*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*11f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <stdio.h> #include <stdlib.h> #define CUERR { cudaError_t err; \ if ((err = cudaGetLastError()) != cudaSuccess) { \ printf("CUDA error: %s, line %d\n", cudaGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } cudaMemcpyToSymbol(atominfo, atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
.file "tmpxft_00025b6d_00000000-6_cuenergy.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, line %d\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC1: .string "Atom count exceeds constant buffer storage capacity\n" .text .globl _Z19copyatomstoconstbufPfif .type _Z19copyatomstoconstbufPfif, @function _Z19copyatomstoconstbufPfif: .LFB2057: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 leaq -61440(%rsp), %r11 .cfi_def_cfa 11, 61464 .LPSRL0: subq $4096, %rsp orq $0, (%rsp) cmpq %r11, %rsp jne .LPSRL0 .cfi_def_cfa_register 7 subq $2600, %rsp .cfi_def_cfa_offset 64064 movq %rdi, %rbx movl %esi, %ebp movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 64024(%rsp) xorl %eax, %eax call cudaGetLastError@PLT testl %eax, %eax jne .L15 cmpl $4000, %ebp jg .L6 leal 0(,%rbp,4), %edx leaq 16(%rsp), %rcx movl $0, %eax testl %edx, %edx jle .L8 .L9: movss (%rbx,%rax,4), %xmm0 movss %xmm0, (%rcx) movss 4(%rbx,%rax,4), %xmm0 movss %xmm0, 4(%rcx) movss 12(%rsp), %xmm0 subss 8(%rbx,%rax,4), %xmm0 mulss %xmm0, %xmm0 movss %xmm0, 8(%rcx) movss 12(%rbx,%rax,4), %xmm0 movss %xmm0, 12(%rcx) addq $4, %rax addq $16, %rcx cmpl %eax, %edx jg .L9 .L8: movslq %edx, %rdx salq $2, %rdx leaq 16(%rsp), %rsi movl $1, %r8d movl $0, %ecx leaq _ZL8atominfo(%rip), %rdi call cudaMemcpyToSymbol@PLT call cudaGetLastError@PLT testl %eax, %eax jne .L16 movl $0, %eax .L3: movq 64024(%rsp), %rdx subq %fs:40, %rdx jne .L17 addq $64040, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $111, %ecx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L6: leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L16: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $129, %ecx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L17: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z19copyatomstoconstbufPfif, .-_Z19copyatomstoconstbufPfif .globl _Z9initatomsPPfi4dim3f .type _Z9initatomsPPfi4dim3f, @function _Z9initatomsPPfi4dim3f: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $40, %rsp .cfi_def_cfa_offset 96 movq %rdi, %rbp movl %esi, %ebx movss %xmm0, 12(%rsp) movq %rdx, 16(%rsp) movl %ecx, 24(%rsp) leal 0(,%rsi,4), %r15d movslq %r15d, %r15 salq $2, %r15 movq %r15, %rdi call malloc@PLT movq %rax, 0(%rbp) cvttss2siq 12(%rsp), %r12 movl %r12d, %r13d imull 16(%rsp), %r13d movl %r12d, %r14d imull 20(%rsp), %r14d imull 24(%rsp), %r12d testl %ebx, %ebx jle .L19 movq %rax, %rbx addq %rax, %r15 .L26: call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r13d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, (%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r14d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, 4(%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r12d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, 8(%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 cvtss2sd %xmm0, %xmm0 addsd %xmm0, %xmm0 subsd .LC3(%rip), %xmm0 cvtsd2ss %xmm0, %xmm0 movss %xmm0, 12(%rbx) addq $16, %rbx cmpq %r15, %rbx jne .L26 .L19: movl $0, %eax addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z9initatomsPPfi4dim3f, .-_Z9initatomsPPfi4dim3f .globl _Z28__device_stub__Z7cenergyifPfifPf .type _Z28__device_stub__Z7cenergyifPfifPf, @function _Z28__device_stub__Z7cenergyifPfifPf: .LFB2083: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movss %xmm0, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L33 .L29: movq 104(%rsp), %rax subq %fs:40, %rax jne .L34 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L33: .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 _Z7cenergyifPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L29 .L34: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z28__device_stub__Z7cenergyifPfifPf, .-_Z28__device_stub__Z7cenergyifPfifPf .globl _Z7cenergyifPf .type _Z7cenergyifPf, @function _Z7cenergyifPf: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z7cenergyifPfifPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7cenergyifPf, .-_Z7cenergyifPf .section .rodata.str1.1 .LC4: .string "_Z7cenergyifPf" .LC5: .string "atominfo" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z7cenergyifPf(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $64000, %r9d movl $0, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _ZL8atominfo(%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 .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 .local _ZL8atominfo .comm _ZL8atominfo,64000,32 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC2: .long 805306368 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC3: .long 0 .long 1072693248 .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.
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <stdio.h> #include <stdlib.h> #define CUERR { cudaError_t err; \ if ((err = cudaGetLastError()) != cudaSuccess) { \ printf("CUDA error: %s, line %d\n", cudaGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } cudaMemcpyToSymbol(atominfo, atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define CUERR { hipError_t err; \ if ((err = hipGetLastError()) != hipSuccess) { \ printf("CUDA error: %s, line %d\n", hipGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } hipMemcpyToSymbol(HIP_SYMBOL(atominfo), atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define CUERR { hipError_t err; \ if ((err = hipGetLastError()) != hipSuccess) { \ printf("CUDA error: %s, line %d\n", hipGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } hipMemcpyToSymbol(HIP_SYMBOL(atominfo), atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z7cenergyifPf .globl _Z7cenergyifPf .p2align 8 .type _Z7cenergyifPf,@function _Z7cenergyifPf: s_clause 0x1 s_load_b32 s2, s[0:1], 0x1c s_load_b32 s10, s[0:1], 0x0 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_lshl_b32 s3, s14, 3 s_and_b32 s4, s15, 0xffffff s_and_b32 s3, s3, 0x7fffff8 s_waitcnt lgkmcnt(0) s_and_b32 s7, s2, 0xffff s_lshr_b32 s2, s2, 16 v_mad_u64_u32 v[0:1], null, s3, s7, v[2:3] v_mad_u64_u32 v[1:2], null, s4, s2, v[3:4] s_cmp_lt_i32 s10, 1 s_cbranch_scc1 .LBB0_3 s_load_b32 s2, s[0:1], 0x4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_3) v_cvt_f32_u32_e32 v9, v1 v_cvt_f32_u32_e32 v11, v0 v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0 v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v5, 0 v_dual_mov_b32 v6, 0 :: v_dual_mov_b32 v7, 0 v_mov_b32_e32 v8, 0 s_getpc_b64 s[8:9] s_add_u32 s8, s8, atominfo@rel32@lo+16 s_addc_u32 s9, s9, atominfo@rel32@hi+24 s_waitcnt lgkmcnt(0) v_dual_mul_f32 v11, s2, v11 :: v_dual_mul_f32 v10, s2, v9 v_mov_b32_e32 v9, 0 v_mul_f32_e64 v12, 0x41800000, s2 .LBB0_2: s_add_u32 s2, s8, -12 s_addc_u32 s3, s9, -1 s_add_u32 s4, s8, -8 s_addc_u32 s5, s9, -1 s_add_u32 s12, s8, -4 s_addc_u32 s13, s9, -1 s_clause 0x3 s_load_b32 s4, s[4:5], 0x0 s_load_b32 s5, s[12:13], 0x0 s_load_b32 s11, s[8:9], 0x0 s_load_b32 s2, s[2:3], 0x0 s_add_i32 s10, s10, -1 s_add_u32 s8, s8, 16 s_addc_u32 s9, s9, 0 s_cmp_lg_u32 s10, 0 s_waitcnt lgkmcnt(0) v_dual_subrev_f32 v13, s4, v10 :: v_dual_subrev_f32 v14, s2, v11 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_fma_f32 v13, v13, v13, s5 v_add_f32_e32 v15, v12, v14 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f32 v14, v14, v14, v13 v_mul_f32_e32 v17, 0x4b800000, v14 v_cmp_gt_f32_e32 vcc_lo, 0x800000, v14 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_cndmask_b32_e32 v14, v14, v17, vcc_lo v_add_f32_e32 v16, v12, v15 v_fma_f32 v15, v15, v15, v13 v_rsq_f32_e32 v14, v14 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_f32_e32 v18, v12, v16 v_cmp_gt_f32_e64 s2, 0x800000, v15 v_fma_f32 v16, v16, v16, v13 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_2) | instid1(VALU_DEP_4) v_add_f32_e32 v17, v12, v18 v_mul_f32_e32 v19, 0x4b800000, v15 v_fma_f32 v18, v18, v18, v13 v_cmp_gt_f32_e64 s3, 0x800000, v16 s_delay_alu instid0(TRANS32_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_4) v_mul_f32_e32 v22, 0x45800000, v14 v_add_f32_e32 v20, v12, v17 v_cndmask_b32_e64 v15, v15, v19, s2 v_mul_f32_e32 v19, 0x4b800000, v16 v_cmp_gt_f32_e64 s4, 0x800000, v18 v_dual_cndmask_b32 v14, v14, v22 :: v_dual_add_f32 v21, v12, v20 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_rsq_f32_e32 v15, v15 v_cndmask_b32_e64 v16, v16, v19, s3 v_mul_f32_e32 v19, 0x4b800000, v18 v_fma_f32 v20, v20, v20, v13 v_dual_add_f32 v23, v12, v21 :: v_dual_fmac_f32 v2, s11, v14 v_fma_f32 v21, v21, v21, v13 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_2) | instid1(TRANS32_DEP_2) v_cndmask_b32_e64 v18, v18, v19, s4 v_rsq_f32_e32 v16, v16 v_cmp_gt_f32_e64 s6, 0x800000, v20 v_mul_f32_e32 v24, 0x45800000, v15 v_fma_f32 v17, v17, v17, v13 v_dual_fmac_f32 v13, v23, v23 :: v_dual_mul_f32 v22, 0x4b800000, v21 v_cmp_gt_f32_e32 vcc_lo, 0x800000, v21 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_cndmask_b32_e64 v15, v15, v24, s2 v_mul_f32_e32 v19, 0x4b800000, v17 v_cmp_gt_f32_e64 s5, 0x800000, v17 v_mul_f32_e32 v24, 0x4b800000, v13 v_cmp_gt_f32_e64 s2, 0x800000, v13 v_rsq_f32_e32 v18, v18 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v17, v17, v19, s5 v_mul_f32_e32 v19, 0x4b800000, v20 v_cndmask_b32_e64 v13, v13, v24, s2 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_rsq_f32_e32 v17, v17 v_cndmask_b32_e64 v19, v20, v19, s6 v_dual_cndmask_b32 v20, v21, v22 :: v_dual_mul_f32 v23, 0x45800000, v16 s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2) v_rsq_f32_e32 v13, v13 v_rsq_f32_e32 v14, v19 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_rsq_f32_e32 v19, v20 v_cndmask_b32_e64 v16, v16, v23, s3 v_fmac_f32_e32 v3, s11, v15 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(TRANS32_DEP_3) v_dual_mul_f32 v15, 0x45800000, v18 :: v_dual_fmac_f32 v4, s11, v16 v_mul_f32_e32 v22, 0x45800000, v13 v_mul_f32_e32 v16, 0x45800000, v17 s_waitcnt_depctr 0xfff v_dual_mul_f32 v20, 0x45800000, v14 :: v_dual_mul_f32 v21, 0x45800000, v19 v_cndmask_b32_e64 v15, v18, v15, s4 v_cndmask_b32_e64 v13, v13, v22, s2 v_cndmask_b32_e64 v16, v17, v16, s5 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_cndmask_b32_e64 v14, v14, v20, s6 v_cndmask_b32_e32 v17, v19, v21, vcc_lo v_dual_fmac_f32 v9, s11, v13 :: v_dual_fmac_f32 v6, s11, v16 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_4) v_dual_fmac_f32 v5, s11, v15 :: v_dual_fmac_f32 v8, s11, v17 v_fmac_f32_e32 v7, s11, v14 s_cbranch_scc1 .LBB0_2 s_branch .LBB0_4 .LBB0_3: v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0 v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v5, 0 v_dual_mov_b32 v6, 0 :: v_dual_mov_b32 v7, 0 v_dual_mov_b32 v8, 0 :: v_dual_mov_b32 v9, 0 .LBB0_4: s_clause 0x1 s_load_b32 s2, s[0:1], 0x10 s_load_b64 s[0:1], s[0:1], 0x8 s_waitcnt lgkmcnt(0) s_lshl_b32 s2, s2, 3 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, s2, 0x7fffff8 s_mul_i32 s2, s2, s7 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mad_u64_u32 v[10:11], null, s2, v1, v[0:1] v_dual_mov_b32 v11, 0 :: v_dual_add_nc_u32 v0, 16, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_dual_mov_b32 v1, v11 :: v_dual_add_nc_u32 v12, 32, v10 v_lshlrev_b64 v[16:17], 2, v[10:11] v_dual_mov_b32 v13, v11 :: v_dual_add_nc_u32 v14, 48, v10 v_lshlrev_b64 v[0:1], 2, v[0:1] v_dual_mov_b32 v15, v11 :: v_dual_add_nc_u32 v18, 64, v10 s_delay_alu instid0(VALU_DEP_3) v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_u32 v16, vcc_lo, s0, v16 v_dual_mov_b32 v19, v11 :: v_dual_add_nc_u32 v20, 0x50, v10 v_add_co_ci_u32_e32 v17, vcc_lo, s1, v17, vcc_lo v_lshlrev_b64 v[14:15], 2, v[14:15] v_add_co_u32 v0, vcc_lo, s0, v0 v_dual_mov_b32 v21, v11 :: v_dual_add_nc_u32 v22, 0x60, v10 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo v_add_co_u32 v12, vcc_lo, s0, v12 v_lshlrev_b64 v[18:19], 2, v[18:19] v_dual_mov_b32 v23, v11 :: v_dual_add_nc_u32 v10, 0x70, v10 v_add_co_ci_u32_e32 v13, vcc_lo, s1, v13, vcc_lo v_add_co_u32 v14, vcc_lo, s0, v14 v_lshlrev_b64 v[20:21], 2, v[20:21] v_add_co_ci_u32_e32 v15, vcc_lo, s1, v15, vcc_lo v_add_co_u32 v18, vcc_lo, s0, v18 v_lshlrev_b64 v[22:23], 2, v[22:23] v_add_co_ci_u32_e32 v19, vcc_lo, s1, v19, vcc_lo v_add_co_u32 v20, vcc_lo, s0, v20 v_lshlrev_b64 v[10:11], 2, v[10:11] v_add_co_ci_u32_e32 v21, vcc_lo, s1, v21, vcc_lo v_add_co_u32 v22, vcc_lo, s0, v22 v_add_co_ci_u32_e32 v23, vcc_lo, s1, v23, vcc_lo s_delay_alu instid0(VALU_DEP_4) v_add_co_u32 v10, vcc_lo, s0, v10 v_add_co_ci_u32_e32 v11, vcc_lo, s1, v11, vcc_lo s_clause 0x7 global_load_b32 v24, v[16:17], off global_load_b32 v25, v[0:1], off global_load_b32 v26, v[12:13], off global_load_b32 v27, v[14:15], off global_load_b32 v28, v[18:19], off global_load_b32 v29, v[20:21], off global_load_b32 v30, v[22:23], off global_load_b32 v31, v[10:11], off s_waitcnt vmcnt(6) v_dual_add_f32 v2, v2, v24 :: v_dual_add_f32 v3, v3, v25 s_waitcnt vmcnt(4) v_dual_add_f32 v4, v4, v26 :: v_dual_add_f32 v5, v5, v27 s_waitcnt vmcnt(2) v_dual_add_f32 v6, v6, v28 :: v_dual_add_f32 v7, v7, v29 s_waitcnt vmcnt(0) v_dual_add_f32 v8, v8, v30 :: v_dual_add_f32 v9, v9, v31 s_clause 0x7 global_store_b32 v[16:17], v2, off global_store_b32 v[0:1], v3, off global_store_b32 v[12:13], v4, off global_store_b32 v[14:15], v5, off global_store_b32 v[18:19], v6, off global_store_b32 v[20:21], v7, off global_store_b32 v[22:23], v8, off global_store_b32 v[10:11], v9, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z7cenergyifPf .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 32 .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 _Z7cenergyifPf, .Lfunc_end0-_Z7cenergyifPf .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .protected atominfo .type atominfo,@object .section .bss,"aw",@nobits .globl atominfo .p2align 4, 0x0 atominfo: .zero 64000 .size atominfo, 64000 .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 atominfo .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: 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: _Z7cenergyifPf .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z7cenergyifPf.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 32 .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.
/*************************************************************************** *cr *cr (C) Copyright 2007 The Board of Trustees of the *cr University of Illinois *cr All Rights Reserved *cr ***************************************************************************/ /* * CUDA accelerated coulombic potential grid test code * John E. Stone and Chris Rodrigues * http://www.ks.uiuc.edu/~johns/ */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #define CUERR { hipError_t err; \ if ((err = hipGetLastError()) != hipSuccess) { \ printf("CUDA error: %s, line %d\n", hipGetErrorString(err), __LINE__); \ return -1; }} // max constant buffer size is 64KB, minus whatever // the CUDA runtime and compiler are using that we don't know about // At 16 bytes for atom, for this program 4070 atoms is about the max // we can store in the constant buffer. #define MAXATOMS 4000 __constant__ float4 atominfo[MAXATOMS]; #define UNROLLX 8 #define UNROLLY 1 #define BLOCKSIZEX 16 #define BLOCKSIZEY 16 #define BLOCKSIZE BLOCKSIZEX * BLOCKSIZEY // This kernel calculates coulombic potential at each grid point and // stores the results in the output array. // // This kernel was written by Chris Rodrigues of Wen-mei's group // __global__ void cenergy(int numatoms, float gridspacing, float * energygrid) { unsigned int xindex = __umul24(blockIdx.x, blockDim.x) * UNROLLX + threadIdx.x; unsigned int yindex = __umul24(blockIdx.y, blockDim.y) + threadIdx.y; unsigned int outaddr = (__umul24(gridDim.x, blockDim.x) * UNROLLX) * yindex + xindex; float coory = gridspacing * yindex; float coorx = gridspacing * xindex; float energyvalx1=0.0f; float energyvalx2=0.0f; float energyvalx3=0.0f; float energyvalx4=0.0f; #if UNROLLX == 8 float energyvalx5=0.0f; float energyvalx6=0.0f; float energyvalx7=0.0f; float energyvalx8=0.0f; #endif float gridspacing_u = gridspacing * BLOCKSIZEX; // // XXX 59/8 FLOPS per atom // int atomid; for (atomid=0; atomid<numatoms; atomid++) { float dy = coory - atominfo[atomid].y; float dyz2 = (dy * dy) + atominfo[atomid].z; float atomq=atominfo[atomid].w; float dx1 = coorx - atominfo[atomid].x; float dx2 = dx1 + gridspacing_u; float dx3 = dx2 + gridspacing_u; float dx4 = dx3 + gridspacing_u; #if UNROLLX == 8 float dx5 = dx4 + gridspacing_u; float dx6 = dx5 + gridspacing_u; float dx7 = dx6 + gridspacing_u; float dx8 = dx7 + gridspacing_u; #endif energyvalx1 += atomq * rsqrtf(dx1*dx1 + dyz2); energyvalx2 += atomq * rsqrtf(dx2*dx2 + dyz2); energyvalx3 += atomq * rsqrtf(dx3*dx3 + dyz2); energyvalx4 += atomq * rsqrtf(dx4*dx4 + dyz2); #if UNROLLX == 8 energyvalx5 += atomq * rsqrtf(dx5*dx5 + dyz2); energyvalx6 += atomq * rsqrtf(dx6*dx6 + dyz2); energyvalx7 += atomq * rsqrtf(dx7*dx7 + dyz2); energyvalx8 += atomq * rsqrtf(dx8*dx8 + dyz2); #endif } energygrid[outaddr ] += energyvalx1; energygrid[outaddr+1*BLOCKSIZEX] += energyvalx2; energygrid[outaddr+2*BLOCKSIZEX] += energyvalx3; energygrid[outaddr+3*BLOCKSIZEX] += energyvalx4; #if UNROLLX == 8 energygrid[outaddr+4*BLOCKSIZEX] += energyvalx5; energygrid[outaddr+5*BLOCKSIZEX] += energyvalx6; energygrid[outaddr+6*BLOCKSIZEX] += energyvalx7; energygrid[outaddr+7*BLOCKSIZEX] += energyvalx8; #endif } int copyatomstoconstbuf(float *atoms, int count, float zplane) { CUERR // check and clear any existing errors if (count > MAXATOMS) { printf("Atom count exceeds constant buffer storage capacity\n"); return -1; } float atompre[4*MAXATOMS]; int i; for (i=0; i<count*4; i+=4) { atompre[i ] = atoms[i ]; atompre[i + 1] = atoms[i + 1]; float dz = zplane - atoms[i + 2]; atompre[i + 2] = dz*dz; atompre[i + 3] = atoms[i + 3]; } hipMemcpyToSymbol(HIP_SYMBOL(atominfo), atompre, count * 4 * sizeof(float), 0); CUERR // check and clear any existing errors return 0; } int initatoms(float **atombuf, int count, dim3 volsize, float gridspacing) { dim3 size; int i; float *atoms; atoms = (float *) malloc(count * 4 * sizeof(float)); *atombuf = atoms; // compute grid dimensions in angstroms size.x = (unsigned int) gridspacing * volsize.x; size.y = (unsigned int) gridspacing * volsize.y; size.z = (unsigned int) gridspacing * volsize.z; for (i=0; i<count; i++) { int addr = i * 4; atoms[addr ] = (rand() / (float) RAND_MAX) * size.x; atoms[addr + 1] = (rand() / (float) RAND_MAX) * size.y; atoms[addr + 2] = (rand() / (float) RAND_MAX) * size.z; atoms[addr + 3] = ((rand() / (float) RAND_MAX) * 2.0) - 1.0; // charge } return 0; }
.text .file "cuenergy.hip" .globl _Z22__device_stub__cenergyifPf # -- Begin function _Z22__device_stub__cenergyifPf .p2align 4, 0x90 .type _Z22__device_stub__cenergyifPf,@function _Z22__device_stub__cenergyifPf: # @_Z22__device_stub__cenergyifPf .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movl %edi, 4(%rsp) movss %xmm0, (%rsp) movq %rsi, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 64(%rsp) movq %rsp, %rax movq %rax, 72(%rsp) leaq 56(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z7cenergyifPf, %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 _Z22__device_stub__cenergyifPf, .Lfunc_end0-_Z22__device_stub__cenergyifPf .cfi_endproc # -- End function .globl _Z19copyatomstoconstbufPfif # -- Begin function _Z19copyatomstoconstbufPfif .p2align 4, 0x90 .type _Z19copyatomstoconstbufPfif,@function _Z19copyatomstoconstbufPfif: # @_Z19copyatomstoconstbufPfif .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $64016, %rsp # imm = 0xFA10 .cfi_def_cfa_offset 64048 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 movss %xmm0, 12(%rsp) # 4-byte Spill movl %esi, %r14d movq %rdi, %rbx callq hipGetLastError testl %eax, %eax je .LBB1_2 # %bb.1: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movq %rax, %rsi movl $113, %edx xorl %eax, %eax callq printf movl $-1, %ebx jmp .LBB1_10 .LBB1_2: # %.critedge movss 12(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero cmpl $4001, %r14d # imm = 0xFA1 jl .LBB1_4 # %bb.3: movl $.Lstr, %edi callq puts@PLT movl $-1, %ebx jmp .LBB1_10 .LBB1_4: leal (,%r14,4), %eax testl %r14d, %r14d jle .LBB1_7 # %bb.5: # %.lr.ph.preheader movl %eax, %ecx xorl %edx, %edx .p2align 4, 0x90 .LBB1_6: # %.lr.ph # =>This Inner Loop Header: Depth=1 movss (%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 16(%rsp,%rdx,4) movss 4(%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 20(%rsp,%rdx,4) movaps %xmm1, %xmm0 subss 8(%rbx,%rdx,4), %xmm0 mulss %xmm0, %xmm0 movss %xmm0, 24(%rsp,%rdx,4) movss 12(%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 28(%rsp,%rdx,4) addq $4, %rdx cmpq %rcx, %rdx jb .LBB1_6 .LBB1_7: # %._crit_edge movslq %eax, %rdx shlq $2, %rdx xorl %ebx, %ebx leaq 16(%rsp), %rsi movl $atominfo, %edi xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol callq hipGetLastError movl %eax, %ebp testl %eax, %eax je .LBB1_9 # %bb.8: movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rax, %rsi movl $131, %edx xorl %eax, %eax callq printf .LBB1_9: negl %ebp sbbl %ebx, %ebx .LBB1_10: movl %ebx, %eax addq $64016, %rsp # imm = 0xFA10 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z19copyatomstoconstbufPfif, .Lfunc_end1-_Z19copyatomstoconstbufPfif .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z9initatomsPPfi4dim3f .LCPI2_0: .long 0x30000000 # float 4.65661287E-10 .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 .LCPI2_1: .quad 0xbff0000000000000 # double -1 .text .globl _Z9initatomsPPfi4dim3f .p2align 4, 0x90 .type _Z9initatomsPPfi4dim3f,@function _Z9initatomsPPfi4dim3f: # @_Z9initatomsPPfi4dim3f .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $16, %rsp .cfi_def_cfa_offset 64 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movss %xmm0, 4(%rsp) # 4-byte Spill movl %ecx, %r14d movq %rdx, %r12 movl %esi, %r15d movq %rdi, %r13 leal (,%r15,4), %eax movslq %eax, %rdi shlq $2, %rdi callq malloc movq %rax, (%r13) testl %r15d, %r15d jle .LBB2_3 # %bb.1: # %.lr.ph movq %rax, %rbx cvttss2si 4(%rsp), %rax # 4-byte Folded Reload imull %eax, %r14d movq %r12, %rcx shrq $32, %rcx imull %eax, %r12d xorps %xmm0, %xmm0 cvtsi2ss %r12, %xmm0 movss %xmm0, 4(%rsp) # 4-byte Spill imull %eax, %ecx xorps %xmm0, %xmm0 cvtsi2ss %rcx, %xmm0 movss %xmm0, 12(%rsp) # 4-byte Spill xorps %xmm0, %xmm0 cvtsi2ss %r14, %xmm0 movss %xmm0, 8(%rsp) # 4-byte Spill movl %r15d, %r14d shlq $2, %r14 xorl %r15d, %r15d .p2align 4, 0x90 .LBB2_2: # =>This Inner Loop Header: Depth=1 callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 movss .LCPI2_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero mulss %xmm1, %xmm0 mulss 4(%rsp), %xmm0 # 4-byte Folded Reload movl %r15d, %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 mulss 12(%rsp), %xmm0 # 4-byte Folded Reload leal 1(%r15), %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 mulss 8(%rsp), %xmm0 # 4-byte Folded Reload leal 2(%r15), %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 cvtss2sd %xmm0, %xmm0 addsd %xmm0, %xmm0 addsd .LCPI2_1(%rip), %xmm0 cvtsd2ss %xmm0, %xmm0 leal 3(%r15), %eax movss %xmm0, (%rbx,%rax,4) addq $4, %r15 cmpq %r15, %r14 jne .LBB2_2 .LBB2_3: # %._crit_edge xorl %eax, %eax addq $16, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9initatomsPPfi4dim3f, .Lfunc_end2-_Z9initatomsPPfi4dim3f .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7cenergyifPf, %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 $1, (%rsp) movl $atominfo, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $64000, %r9d # imm = 0xFA00 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_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 atominfo,@object # @atominfo .local atominfo .comm atominfo,64000,16 .type _Z7cenergyifPf,@object # @_Z7cenergyifPf .section .rodata,"a",@progbits .globl _Z7cenergyifPf .p2align 3, 0x0 _Z7cenergyifPf: .quad _Z22__device_stub__cenergyifPf .size _Z7cenergyifPf, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "CUDA error: %s, line %d\n" .size .L.str, 25 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7cenergyifPf" .size .L__unnamed_1, 15 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "atominfo" .size .L__unnamed_2, 9 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Atom count exceeds constant buffer storage capacity" .size .Lstr, 52 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__cenergyifPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym atominfo .addrsig_sym _Z7cenergyifPf .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_00025b6d_00000000-6_cuenergy.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, line %d\n" .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC1: .string "Atom count exceeds constant buffer storage capacity\n" .text .globl _Z19copyatomstoconstbufPfif .type _Z19copyatomstoconstbufPfif, @function _Z19copyatomstoconstbufPfif: .LFB2057: .cfi_startproc endbr64 pushq %rbp .cfi_def_cfa_offset 16 .cfi_offset 6, -16 pushq %rbx .cfi_def_cfa_offset 24 .cfi_offset 3, -24 leaq -61440(%rsp), %r11 .cfi_def_cfa 11, 61464 .LPSRL0: subq $4096, %rsp orq $0, (%rsp) cmpq %r11, %rsp jne .LPSRL0 .cfi_def_cfa_register 7 subq $2600, %rsp .cfi_def_cfa_offset 64064 movq %rdi, %rbx movl %esi, %ebp movss %xmm0, 12(%rsp) movq %fs:40, %rax movq %rax, 64024(%rsp) xorl %eax, %eax call cudaGetLastError@PLT testl %eax, %eax jne .L15 cmpl $4000, %ebp jg .L6 leal 0(,%rbp,4), %edx leaq 16(%rsp), %rcx movl $0, %eax testl %edx, %edx jle .L8 .L9: movss (%rbx,%rax,4), %xmm0 movss %xmm0, (%rcx) movss 4(%rbx,%rax,4), %xmm0 movss %xmm0, 4(%rcx) movss 12(%rsp), %xmm0 subss 8(%rbx,%rax,4), %xmm0 mulss %xmm0, %xmm0 movss %xmm0, 8(%rcx) movss 12(%rbx,%rax,4), %xmm0 movss %xmm0, 12(%rcx) addq $4, %rax addq $16, %rcx cmpl %eax, %edx jg .L9 .L8: movslq %edx, %rdx salq $2, %rdx leaq 16(%rsp), %rsi movl $1, %r8d movl $0, %ecx leaq _ZL8atominfo(%rip), %rdi call cudaMemcpyToSymbol@PLT call cudaGetLastError@PLT testl %eax, %eax jne .L16 movl $0, %eax .L3: movq 64024(%rsp), %rdx subq %fs:40, %rdx jne .L17 addq $64040, %rsp .cfi_remember_state .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 ret .L15: .cfi_restore_state movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $111, %ecx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L6: leaq .LC1(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L16: movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rdx movl $129, %ecx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl $-1, %eax jmp .L3 .L17: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size _Z19copyatomstoconstbufPfif, .-_Z19copyatomstoconstbufPfif .globl _Z9initatomsPPfi4dim3f .type _Z9initatomsPPfi4dim3f, @function _Z9initatomsPPfi4dim3f: .LFB2058: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $40, %rsp .cfi_def_cfa_offset 96 movq %rdi, %rbp movl %esi, %ebx movss %xmm0, 12(%rsp) movq %rdx, 16(%rsp) movl %ecx, 24(%rsp) leal 0(,%rsi,4), %r15d movslq %r15d, %r15 salq $2, %r15 movq %r15, %rdi call malloc@PLT movq %rax, 0(%rbp) cvttss2siq 12(%rsp), %r12 movl %r12d, %r13d imull 16(%rsp), %r13d movl %r12d, %r14d imull 20(%rsp), %r14d imull 24(%rsp), %r12d testl %ebx, %ebx jle .L19 movq %rax, %rbx addq %rax, %r15 .L26: call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r13d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, (%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r14d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, 4(%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 movl %r12d, %eax pxor %xmm1, %xmm1 cvtsi2ssq %rax, %xmm1 mulss %xmm1, %xmm0 movss %xmm0, 8(%rbx) call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC2(%rip), %xmm0 cvtss2sd %xmm0, %xmm0 addsd %xmm0, %xmm0 subsd .LC3(%rip), %xmm0 cvtsd2ss %xmm0, %xmm0 movss %xmm0, 12(%rbx) addq $16, %rbx cmpq %r15, %rbx jne .L26 .L19: movl $0, %eax addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2058: .size _Z9initatomsPPfi4dim3f, .-_Z9initatomsPPfi4dim3f .globl _Z28__device_stub__Z7cenergyifPfifPf .type _Z28__device_stub__Z7cenergyifPfifPf, @function _Z28__device_stub__Z7cenergyifPfifPf: .LFB2083: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movl %edi, 12(%rsp) movss %xmm0, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 12(%rsp), %rax movq %rax, 80(%rsp) leaq 8(%rsp), %rax movq %rax, 88(%rsp) movq %rsp, %rax movq %rax, 96(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L33 .L29: movq 104(%rsp), %rax subq %fs:40, %rax jne .L34 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L33: .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 _Z7cenergyifPf(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L29 .L34: call __stack_chk_fail@PLT .cfi_endproc .LFE2083: .size _Z28__device_stub__Z7cenergyifPfifPf, .-_Z28__device_stub__Z7cenergyifPfifPf .globl _Z7cenergyifPf .type _Z7cenergyifPf, @function _Z7cenergyifPf: .LFB2084: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z28__device_stub__Z7cenergyifPfifPf addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2084: .size _Z7cenergyifPf, .-_Z7cenergyifPf .section .rodata.str1.1 .LC4: .string "_Z7cenergyifPf" .LC5: .string "atominfo" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2086: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC4(%rip), %rdx movq %rdx, %rcx leaq _Z7cenergyifPf(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $64000, %r9d movl $0, %r8d leaq .LC5(%rip), %rdx movq %rdx, %rcx leaq _ZL8atominfo(%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 .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 .local _ZL8atominfo .comm _ZL8atominfo,64000,32 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC2: .long 805306368 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC3: .long 0 .long 1072693248 .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 "cuenergy.hip" .globl _Z22__device_stub__cenergyifPf # -- Begin function _Z22__device_stub__cenergyifPf .p2align 4, 0x90 .type _Z22__device_stub__cenergyifPf,@function _Z22__device_stub__cenergyifPf: # @_Z22__device_stub__cenergyifPf .cfi_startproc # %bb.0: subq $88, %rsp .cfi_def_cfa_offset 96 movl %edi, 4(%rsp) movss %xmm0, (%rsp) movq %rsi, 56(%rsp) leaq 4(%rsp), %rax movq %rax, 64(%rsp) movq %rsp, %rax movq %rax, 72(%rsp) leaq 56(%rsp), %rax movq %rax, 80(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 64(%rsp), %r9 movl $_Z7cenergyifPf, %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 _Z22__device_stub__cenergyifPf, .Lfunc_end0-_Z22__device_stub__cenergyifPf .cfi_endproc # -- End function .globl _Z19copyatomstoconstbufPfif # -- Begin function _Z19copyatomstoconstbufPfif .p2align 4, 0x90 .type _Z19copyatomstoconstbufPfif,@function _Z19copyatomstoconstbufPfif: # @_Z19copyatomstoconstbufPfif .cfi_startproc # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 subq $64016, %rsp # imm = 0xFA10 .cfi_def_cfa_offset 64048 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %rbp, -16 movss %xmm0, 12(%rsp) # 4-byte Spill movl %esi, %r14d movq %rdi, %rbx callq hipGetLastError testl %eax, %eax je .LBB1_2 # %bb.1: movl %eax, %edi callq hipGetErrorString movl $.L.str, %edi movq %rax, %rsi movl $113, %edx xorl %eax, %eax callq printf movl $-1, %ebx jmp .LBB1_10 .LBB1_2: # %.critedge movss 12(%rsp), %xmm1 # 4-byte Reload # xmm1 = mem[0],zero,zero,zero cmpl $4001, %r14d # imm = 0xFA1 jl .LBB1_4 # %bb.3: movl $.Lstr, %edi callq puts@PLT movl $-1, %ebx jmp .LBB1_10 .LBB1_4: leal (,%r14,4), %eax testl %r14d, %r14d jle .LBB1_7 # %bb.5: # %.lr.ph.preheader movl %eax, %ecx xorl %edx, %edx .p2align 4, 0x90 .LBB1_6: # %.lr.ph # =>This Inner Loop Header: Depth=1 movss (%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 16(%rsp,%rdx,4) movss 4(%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 20(%rsp,%rdx,4) movaps %xmm1, %xmm0 subss 8(%rbx,%rdx,4), %xmm0 mulss %xmm0, %xmm0 movss %xmm0, 24(%rsp,%rdx,4) movss 12(%rbx,%rdx,4), %xmm0 # xmm0 = mem[0],zero,zero,zero movss %xmm0, 28(%rsp,%rdx,4) addq $4, %rdx cmpq %rcx, %rdx jb .LBB1_6 .LBB1_7: # %._crit_edge movslq %eax, %rdx shlq $2, %rdx xorl %ebx, %ebx leaq 16(%rsp), %rsi movl $atominfo, %edi xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol callq hipGetLastError movl %eax, %ebp testl %eax, %eax je .LBB1_9 # %bb.8: movl %ebp, %edi callq hipGetErrorString movl $.L.str, %edi movq %rax, %rsi movl $131, %edx xorl %eax, %eax callq printf .LBB1_9: negl %ebp sbbl %ebx, %ebx .LBB1_10: movl %ebx, %eax addq $64016, %rsp # imm = 0xFA10 .cfi_def_cfa_offset 32 popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .Lfunc_end1: .size _Z19copyatomstoconstbufPfif, .Lfunc_end1-_Z19copyatomstoconstbufPfif .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function _Z9initatomsPPfi4dim3f .LCPI2_0: .long 0x30000000 # float 4.65661287E-10 .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 .LCPI2_1: .quad 0xbff0000000000000 # double -1 .text .globl _Z9initatomsPPfi4dim3f .p2align 4, 0x90 .type _Z9initatomsPPfi4dim3f,@function _Z9initatomsPPfi4dim3f: # @_Z9initatomsPPfi4dim3f .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r13 .cfi_def_cfa_offset 32 pushq %r12 .cfi_def_cfa_offset 40 pushq %rbx .cfi_def_cfa_offset 48 subq $16, %rsp .cfi_def_cfa_offset 64 .cfi_offset %rbx, -48 .cfi_offset %r12, -40 .cfi_offset %r13, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movss %xmm0, 4(%rsp) # 4-byte Spill movl %ecx, %r14d movq %rdx, %r12 movl %esi, %r15d movq %rdi, %r13 leal (,%r15,4), %eax movslq %eax, %rdi shlq $2, %rdi callq malloc movq %rax, (%r13) testl %r15d, %r15d jle .LBB2_3 # %bb.1: # %.lr.ph movq %rax, %rbx cvttss2si 4(%rsp), %rax # 4-byte Folded Reload imull %eax, %r14d movq %r12, %rcx shrq $32, %rcx imull %eax, %r12d xorps %xmm0, %xmm0 cvtsi2ss %r12, %xmm0 movss %xmm0, 4(%rsp) # 4-byte Spill imull %eax, %ecx xorps %xmm0, %xmm0 cvtsi2ss %rcx, %xmm0 movss %xmm0, 12(%rsp) # 4-byte Spill xorps %xmm0, %xmm0 cvtsi2ss %r14, %xmm0 movss %xmm0, 8(%rsp) # 4-byte Spill movl %r15d, %r14d shlq $2, %r14 xorl %r15d, %r15d .p2align 4, 0x90 .LBB2_2: # =>This Inner Loop Header: Depth=1 callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 movss .LCPI2_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero mulss %xmm1, %xmm0 mulss 4(%rsp), %xmm0 # 4-byte Folded Reload movl %r15d, %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 mulss 12(%rsp), %xmm0 # 4-byte Folded Reload leal 1(%r15), %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 mulss 8(%rsp), %xmm0 # 4-byte Folded Reload leal 2(%r15), %eax movss %xmm0, (%rbx,%rax,4) callq rand xorps %xmm0, %xmm0 cvtsi2ss %eax, %xmm0 mulss .LCPI2_0(%rip), %xmm0 cvtss2sd %xmm0, %xmm0 addsd %xmm0, %xmm0 addsd .LCPI2_1(%rip), %xmm0 cvtsd2ss %xmm0, %xmm0 leal 3(%r15), %eax movss %xmm0, (%rbx,%rax,4) addq $4, %r15 cmpq %r15, %r14 jne .LBB2_2 .LBB2_3: # %._crit_edge xorl %eax, %eax addq $16, %rsp .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9initatomsPPfi4dim3f, .Lfunc_end2-_Z9initatomsPPfi4dim3f .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 subq $32, %rsp .cfi_def_cfa_offset 48 .cfi_offset %rbx, -16 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB3_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB3_2: movq __hip_gpubin_handle(%rip), %rbx xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z7cenergyifPf, %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 $1, (%rsp) movl $atominfo, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $64000, %r9d # imm = 0xFA00 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_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 atominfo,@object # @atominfo .local atominfo .comm atominfo,64000,16 .type _Z7cenergyifPf,@object # @_Z7cenergyifPf .section .rodata,"a",@progbits .globl _Z7cenergyifPf .p2align 3, 0x0 _Z7cenergyifPf: .quad _Z22__device_stub__cenergyifPf .size _Z7cenergyifPf, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "CUDA error: %s, line %d\n" .size .L.str, 25 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z7cenergyifPf" .size .L__unnamed_1, 15 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "atominfo" .size .L__unnamed_2, 9 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .type .Lstr,@object # @str .section .rodata.str1.1,"aMS",@progbits,1 .Lstr: .asciz "Atom count exceeds constant buffer storage capacity" .size .Lstr, 52 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z22__device_stub__cenergyifPf .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym atominfo .addrsig_sym _Z7cenergyifPf .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <iostream> #include <unistd.h> #include <cuda.h> #include <cuda_runtime.h> #define size 21 // Tamanho da matrix // Exibe os pontos na tela __host__ void print(bool grid[][size]){ std::cout << "\n\n\n\n\n"; for(unsigned int i = 1; i < size-1; i++) { for(unsigned int j = 1; j < size-1; j++) std::cout << (grid[i][j]?"#":"_"); std::cout << std::endl; } } __host__ bool someoneAlive(bool grid[][size]){ for(unsigned int i=0; i < size; i++) for(unsigned int j=0; j < size; j++) if(grid[i][j]==true) return true; return false; } // Calcula a simulacao __global__ void jogo(bool grid[][size]){ int m=blockIdx.x*blockDim.x+threadIdx.x; int n=blockIdx.y*blockDim.y+threadIdx.y; if (m<size && n<size){ // printf("m: %d n: %d\n",m,n); bool grid_tmp[size][size] = {}; for(unsigned int i=0; i < size; i++){ for(unsigned int j=0; j < size; j++){ grid_tmp[i][j] = grid[i][j]; // printf("%d",grid[i][j]); } // printf("\n"); } unsigned int count = 0; for(int k = -1; k <= 1; k++) for(int l = -1; l <= 1; l++) if(k != 0 || l != 0) if(grid_tmp[m+k][n+l]) ++count; if(count < 2 || count > 3){ grid[m][n] = false; // printf("m: %d n: %d MORREU\n",m,n); // printf("count = %d\n", count); } else { if(count == 3){ grid[m][n] = true; // printf("m: %d n: %d REVIVEU\n",m,n); } } } return; } int main(){ // bool grid[size][size] = {}; // dados iniciais bool grid[size][size] = {}; // dados iniciais grid[ 5][ 7] = true; grid[ 6][ 8] = true; grid[ 8][ 8] = true; grid[ 6][ 9] = true; grid[ 8][10] = true; grid[ 9][10] = true; grid[ 8][11] = true; grid[10][11] = true; grid[10][12] = true; bool d_grid[size][size]; int mem_size = size*size*sizeof(bool); cudaMalloc((void **) d_grid, mem_size); int nthreads = 7; dim3 blocks(size/nthreads+1,size/nthreads+1); dim3 threads(nthreads,nthreads); while(someoneAlive(grid)){ cudaMemcpy(d_grid, grid, mem_size, cudaMemcpyHostToDevice); jogo<<<blocks,threads>>>(d_grid); cudaDeviceSynchronize(); cudaMemcpy(grid, d_grid, mem_size, cudaMemcpyDeviceToHost); print(grid); usleep(100000); return 0; } cudaFree(d_grid); }
.file "tmpxft_0019ba00_00000000-6_q1try.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3687: .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 .LFE3687: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "#" .LC1: .string "_" .LC2: .string "\n\n\n\n\n" .text .globl _Z5printPA21_b .type _Z5printPA21_b, @function _Z5printPA21_b: .LFB3682: .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 $8, %rsp .cfi_def_cfa_offset 64 movq %rdi, %r15 movl $5, %edx leaq .LC2(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT leaq 41(%r15), %rbp addq $440, %r15 leaq .LC1(%rip), %r14 leaq .LC0(%rip), %r13 leaq _ZSt4cout(%rip), %r12 jmp .L4 .L14: call _ZSt16__throw_bad_castv@PLT .L15: movzbl 67(%rbx), %esi .L9: movsbl %sil, %esi movq %r12, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addq $21, %rbp cmpq %r15, %rbp je .L3 .L4: leaq -19(%rbp), %rbx .L6: cmpb $0, (%rbx) movq %r13, %rsi cmove %r14, %rsi movl $1, %edx movq %r12, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $1, %rbx cmpq %rbx, %rbp jne .L6 movq (%r12), %rax movq -24(%rax), %rax movq 240(%r12,%rax), %rbx testq %rbx, %rbx je .L14 cmpb $0, 56(%rbx) jne .L15 movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) movl %eax, %esi jmp .L9 .L3: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3682: .size _Z5printPA21_b, .-_Z5printPA21_b .globl _Z12someoneAlivePA21_b .type _Z12someoneAlivePA21_b, @function _Z12someoneAlivePA21_b: .LFB3683: .cfi_startproc endbr64 leaq 21(%rdi), %rcx addq $462, %rdi .L17: leaq -21(%rcx), %rax .L19: movzbl (%rax), %edx testb %dl, %dl jne .L16 addq $1, %rax cmpq %rcx, %rax jne .L19 addq $21, %rcx cmpq %rdi, %rcx jne .L17 .L16: movl %edx, %eax ret .cfi_endproc .LFE3683: .size _Z12someoneAlivePA21_b, .-_Z12someoneAlivePA21_b .globl _Z27__device_stub__Z4jogoPA21_bPA21_b .type _Z27__device_stub__Z4jogoPA21_bPA21_b, @function _Z27__device_stub__Z4jogoPA21_bPA21_b: .LFB3709: .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 .L25 .L21: movq 88(%rsp), %rax subq %fs:40, %rax jne .L26 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L25: .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 _Z4jogoPA21_b(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L21 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE3709: .size _Z27__device_stub__Z4jogoPA21_bPA21_b, .-_Z27__device_stub__Z4jogoPA21_bPA21_b .globl _Z4jogoPA21_b .type _Z4jogoPA21_b, @function _Z4jogoPA21_b: .LFB3710: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z4jogoPA21_bPA21_b addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3710: .size _Z4jogoPA21_b, .-_Z4jogoPA21_b .globl main .type main, @function main: .LFB3684: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $944, %rsp .cfi_def_cfa_offset 960 movq %fs:40, %rax movq %rax, 936(%rsp) xorl %eax, %eax leaq 32(%rsp), %rdi movl $55, %ecx rep stosq movb $0, (%rdi) movb $1, 144(%rsp) movb $1, 166(%rsp) movb $1, 208(%rsp) movb $1, 167(%rsp) movb $1, 210(%rsp) movb $1, 231(%rsp) movb $1, 211(%rsp) movb $1, 253(%rsp) movb $1, 254(%rsp) leaq 480(%rsp), %rdi movl $441, %esi call cudaMalloc@PLT movl $4, 8(%rsp) movl $4, 12(%rsp) movl $1, 16(%rsp) movl $7, 20(%rsp) movl $7, 24(%rsp) movl $1, 28(%rsp) leaq 32(%rsp), %rdi call _Z12someoneAlivePA21_b testb %al, %al je .L30 leaq 32(%rsp), %rsi leaq 480(%rsp), %rdi movl $1, %ecx movl $441, %edx call cudaMemcpy@PLT movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 8(%rsp), %rdi movl 16(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L35 .L31: call cudaDeviceSynchronize@PLT leaq 480(%rsp), %rsi leaq 32(%rsp), %rbx movl $2, %ecx movl $441, %edx movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rdi call _Z5printPA21_b movl $100000, %edi call usleep@PLT .L32: movq 936(%rsp), %rax subq %fs:40, %rax jne .L36 movl $0, %eax addq $944, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L35: .cfi_restore_state leaq 480(%rsp), %rdi call _Z27__device_stub__Z4jogoPA21_bPA21_b jmp .L31 .L30: leaq 480(%rsp), %rdi call cudaFree@PLT jmp .L32 .L36: call __stack_chk_fail@PLT .cfi_endproc .LFE3684: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z4jogoPA21_b" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3712: .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 _Z4jogoPA21_b(%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 .LFE3712: .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 <iostream> #include <unistd.h> #include <cuda.h> #include <cuda_runtime.h> #define size 21 // Tamanho da matrix // Exibe os pontos na tela __host__ void print(bool grid[][size]){ std::cout << "\n\n\n\n\n"; for(unsigned int i = 1; i < size-1; i++) { for(unsigned int j = 1; j < size-1; j++) std::cout << (grid[i][j]?"#":"_"); std::cout << std::endl; } } __host__ bool someoneAlive(bool grid[][size]){ for(unsigned int i=0; i < size; i++) for(unsigned int j=0; j < size; j++) if(grid[i][j]==true) return true; return false; } // Calcula a simulacao __global__ void jogo(bool grid[][size]){ int m=blockIdx.x*blockDim.x+threadIdx.x; int n=blockIdx.y*blockDim.y+threadIdx.y; if (m<size && n<size){ // printf("m: %d n: %d\n",m,n); bool grid_tmp[size][size] = {}; for(unsigned int i=0; i < size; i++){ for(unsigned int j=0; j < size; j++){ grid_tmp[i][j] = grid[i][j]; // printf("%d",grid[i][j]); } // printf("\n"); } unsigned int count = 0; for(int k = -1; k <= 1; k++) for(int l = -1; l <= 1; l++) if(k != 0 || l != 0) if(grid_tmp[m+k][n+l]) ++count; if(count < 2 || count > 3){ grid[m][n] = false; // printf("m: %d n: %d MORREU\n",m,n); // printf("count = %d\n", count); } else { if(count == 3){ grid[m][n] = true; // printf("m: %d n: %d REVIVEU\n",m,n); } } } return; } int main(){ // bool grid[size][size] = {}; // dados iniciais bool grid[size][size] = {}; // dados iniciais grid[ 5][ 7] = true; grid[ 6][ 8] = true; grid[ 8][ 8] = true; grid[ 6][ 9] = true; grid[ 8][10] = true; grid[ 9][10] = true; grid[ 8][11] = true; grid[10][11] = true; grid[10][12] = true; bool d_grid[size][size]; int mem_size = size*size*sizeof(bool); cudaMalloc((void **) d_grid, mem_size); int nthreads = 7; dim3 blocks(size/nthreads+1,size/nthreads+1); dim3 threads(nthreads,nthreads); while(someoneAlive(grid)){ cudaMemcpy(d_grid, grid, mem_size, cudaMemcpyHostToDevice); jogo<<<blocks,threads>>>(d_grid); cudaDeviceSynchronize(); cudaMemcpy(grid, d_grid, mem_size, cudaMemcpyDeviceToHost); print(grid); usleep(100000); return 0; } cudaFree(d_grid); }
#include <iostream> #include <unistd.h> #include <hip/hip_runtime.h> #define size 21 // Tamanho da matrix // Exibe os pontos na tela __host__ void print(bool grid[][size]){ std::cout << "\n\n\n\n\n"; for(unsigned int i = 1; i < size-1; i++) { for(unsigned int j = 1; j < size-1; j++) std::cout << (grid[i][j]?"#":"_"); std::cout << std::endl; } } __host__ bool someoneAlive(bool grid[][size]){ for(unsigned int i=0; i < size; i++) for(unsigned int j=0; j < size; j++) if(grid[i][j]==true) return true; return false; } // Calcula a simulacao __global__ void jogo(bool grid[][size]){ int m=blockIdx.x*blockDim.x+threadIdx.x; int n=blockIdx.y*blockDim.y+threadIdx.y; if (m<size && n<size){ // printf("m: %d n: %d\n",m,n); bool grid_tmp[size][size] = {}; for(unsigned int i=0; i < size; i++){ for(unsigned int j=0; j < size; j++){ grid_tmp[i][j] = grid[i][j]; // printf("%d",grid[i][j]); } // printf("\n"); } unsigned int count = 0; for(int k = -1; k <= 1; k++) for(int l = -1; l <= 1; l++) if(k != 0 || l != 0) if(grid_tmp[m+k][n+l]) ++count; if(count < 2 || count > 3){ grid[m][n] = false; // printf("m: %d n: %d MORREU\n",m,n); // printf("count = %d\n", count); } else { if(count == 3){ grid[m][n] = true; // printf("m: %d n: %d REVIVEU\n",m,n); } } } return; } int main(){ // bool grid[size][size] = {}; // dados iniciais bool grid[size][size] = {}; // dados iniciais grid[ 5][ 7] = true; grid[ 6][ 8] = true; grid[ 8][ 8] = true; grid[ 6][ 9] = true; grid[ 8][10] = true; grid[ 9][10] = true; grid[ 8][11] = true; grid[10][11] = true; grid[10][12] = true; bool d_grid[size][size]; int mem_size = size*size*sizeof(bool); hipMalloc((void **) d_grid, mem_size); int nthreads = 7; dim3 blocks(size/nthreads+1,size/nthreads+1); dim3 threads(nthreads,nthreads); while(someoneAlive(grid)){ hipMemcpy(d_grid, grid, mem_size, hipMemcpyHostToDevice); jogo<<<blocks,threads>>>(d_grid); hipDeviceSynchronize(); hipMemcpy(grid, d_grid, mem_size, hipMemcpyDeviceToHost); print(grid); usleep(100000); return 0; } hipFree(d_grid); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <iostream> #include <unistd.h> #include <hip/hip_runtime.h> #define size 21 // Tamanho da matrix // Exibe os pontos na tela __host__ void print(bool grid[][size]){ std::cout << "\n\n\n\n\n"; for(unsigned int i = 1; i < size-1; i++) { for(unsigned int j = 1; j < size-1; j++) std::cout << (grid[i][j]?"#":"_"); std::cout << std::endl; } } __host__ bool someoneAlive(bool grid[][size]){ for(unsigned int i=0; i < size; i++) for(unsigned int j=0; j < size; j++) if(grid[i][j]==true) return true; return false; } // Calcula a simulacao __global__ void jogo(bool grid[][size]){ int m=blockIdx.x*blockDim.x+threadIdx.x; int n=blockIdx.y*blockDim.y+threadIdx.y; if (m<size && n<size){ // printf("m: %d n: %d\n",m,n); bool grid_tmp[size][size] = {}; for(unsigned int i=0; i < size; i++){ for(unsigned int j=0; j < size; j++){ grid_tmp[i][j] = grid[i][j]; // printf("%d",grid[i][j]); } // printf("\n"); } unsigned int count = 0; for(int k = -1; k <= 1; k++) for(int l = -1; l <= 1; l++) if(k != 0 || l != 0) if(grid_tmp[m+k][n+l]) ++count; if(count < 2 || count > 3){ grid[m][n] = false; // printf("m: %d n: %d MORREU\n",m,n); // printf("count = %d\n", count); } else { if(count == 3){ grid[m][n] = true; // printf("m: %d n: %d REVIVEU\n",m,n); } } } return; } int main(){ // bool grid[size][size] = {}; // dados iniciais bool grid[size][size] = {}; // dados iniciais grid[ 5][ 7] = true; grid[ 6][ 8] = true; grid[ 8][ 8] = true; grid[ 6][ 9] = true; grid[ 8][10] = true; grid[ 9][10] = true; grid[ 8][11] = true; grid[10][11] = true; grid[10][12] = true; bool d_grid[size][size]; int mem_size = size*size*sizeof(bool); hipMalloc((void **) d_grid, mem_size); int nthreads = 7; dim3 blocks(size/nthreads+1,size/nthreads+1); dim3 threads(nthreads,nthreads); while(someoneAlive(grid)){ hipMemcpy(d_grid, grid, mem_size, hipMemcpyHostToDevice); jogo<<<blocks,threads>>>(d_grid); hipDeviceSynchronize(); hipMemcpy(grid, d_grid, mem_size, hipMemcpyDeviceToHost); print(grid); usleep(100000); return 0; } hipFree(d_grid); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z4jogoPA21_b .globl _Z4jogoPA21_b .p2align 8 .type _Z4jogoPA21_b,@function _Z4jogoPA21_b: s_load_b32 s2, s[0:1], 0x14 v_and_b32_e32 v2, 0x3ff, v0 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_and_b32 s8, s2, 0xffff s_lshr_b32 s2, s2, 16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_mad_u64_u32 v[0:1], null, s14, s8, v[2:3] s_mul_i32 s9, s15, s2 s_mov_b32 s2, exec_lo v_add_nc_u32_e32 v1, s9, v3 v_max_i32_e32 v4, v0, v1 s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e32 21, v4 s_cbranch_execz .LBB0_15 s_load_b64 s[0:1], s[0:1], 0x0 s_mov_b32 s4, 0 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(SALU_CYCLE_1) s_mov_b32 s5, s4 s_mov_b32 s6, s4 s_mov_b32 s7, s4 v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v9, s7 v_dual_mov_b32 v8, s6 :: v_dual_mov_b32 v7, s5 s_delay_alu instid0(VALU_DEP_2) v_dual_mov_b32 v6, s4 :: v_dual_mov_b32 v5, v4 s_clause 0x18 scratch_store_b8 off, v4, off offset:456 scratch_store_b64 off, v[4:5], off offset:448 scratch_store_b128 off, v[6:9], off offset:432 scratch_store_b128 off, v[6:9], off offset:416 scratch_store_b128 off, v[6:9], off offset:400 scratch_store_b128 off, v[6:9], off offset:384 scratch_store_b128 off, v[6:9], off offset:368 scratch_store_b128 off, v[6:9], off offset:352 scratch_store_b128 off, v[6:9], off offset:336 scratch_store_b128 off, v[6:9], off offset:320 scratch_store_b128 off, v[6:9], off offset:304 scratch_store_b128 off, v[6:9], off offset:288 scratch_store_b128 off, v[6:9], off offset:272 scratch_store_b128 off, v[6:9], off offset:256 scratch_store_b128 off, v[6:9], off offset:240 scratch_store_b128 off, v[6:9], off offset:224 scratch_store_b128 off, v[6:9], off offset:208 scratch_store_b128 off, v[6:9], off offset:192 scratch_store_b128 off, v[6:9], off offset:176 scratch_store_b128 off, v[6:9], off offset:160 scratch_store_b128 off, v[6:9], off offset:144 scratch_store_b128 off, v[6:9], off offset:128 scratch_store_b128 off, v[6:9], off offset:112 scratch_store_b128 off, v[6:9], off offset:96 scratch_store_b128 off, v[6:9], off offset:80 v_dual_mov_b32 v4, 16 :: v_dual_mov_b32 v5, 0 s_clause 0x1 scratch_store_b128 off, v[6:9], off offset:64 scratch_store_b128 off, v[6:9], off offset:48 s_waitcnt lgkmcnt(0) s_mov_b64 s[2:3], s[0:1] s_clause 0x1 scratch_store_b128 off, v[6:9], off offset:32 scratch_store_b128 off, v[6:9], off offset:16 .p2align 6 .LBB0_2: v_mov_b32_e32 v6, v4 s_mov_b64 s[6:7], 0 .LBB0_3: s_delay_alu instid0(SALU_CYCLE_1) s_add_u32 s10, s2, s6 s_addc_u32 s11, s3, s7 s_add_u32 s6, s6, 1 global_load_u8 v7, v5, s[10:11] s_addc_u32 s7, s7, 0 s_cmp_eq_u32 s6, 21 s_waitcnt vmcnt(0) scratch_store_b8 v6, v7, off v_add_nc_u32_e32 v6, 1, v6 s_cbranch_scc0 .LBB0_3 s_add_i32 s4, s4, 1 v_add_nc_u32_e32 v4, 21, v4 s_add_u32 s2, s2, 21 s_addc_u32 s3, s3, 0 s_cmp_lg_u32 s4, 21 s_cbranch_scc1 .LBB0_2 s_mul_i32 s2, s14, s8 v_mul_u32_u24_e32 v2, 21, v2 s_mul_i32 s2, s2, 21 v_mov_b32_e32 v4, 16 s_add_i32 s9, s9, s2 s_mov_b32 s2, -1 v_add3_u32 v2, s9, v3, v2 s_delay_alu instid0(VALU_DEP_1) v_add3_u32 v3, v2, v4, 0xffffffea v_mov_b32_e32 v2, 0 s_branch .LBB0_7 .p2align 6 .LBB0_6: v_add_nc_u32_e32 v3, 21, v3 s_add_i32 s2, s2, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s2, 2 s_cbranch_scc1 .LBB0_11 .LBB0_7: s_mov_b32 s3, 0 s_branch .LBB0_9 .LBB0_8: s_add_i32 s3, s3, 1 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u32 s3, 3 s_cbranch_scc1 .LBB0_6 .LBB0_9: s_add_i32 s4, s3, -1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_or_b32 s4, s4, s2 s_cmp_eq_u32 s4, 0 s_cbranch_scc1 .LBB0_8 v_add_nc_u32_e32 v4, s3, v3 scratch_load_u8 v4, v4, off s_waitcnt vmcnt(0) v_add_nc_u32_e32 v2, v2, v4 s_branch .LBB0_8 .LBB0_11: v_and_b32_e32 v3, -2, v2 s_delay_alu instid0(VALU_DEP_1) v_cmp_ne_u32_e64 s2, 2, v3 v_cmp_eq_u32_e32 vcc_lo, 2, v3 v_mov_b32_e32 v3, 0 s_and_saveexec_b32 s3, vcc_lo v_cmp_eq_u32_e32 vcc_lo, 3, v2 v_mov_b32_e32 v3, 1 s_and_not1_b32 s2, s2, exec_lo s_and_b32 s4, vcc_lo, exec_lo s_delay_alu instid0(SALU_CYCLE_1) s_or_b32 s2, s2, s4 s_or_b32 exec_lo, exec_lo, s3 s_delay_alu instid0(SALU_CYCLE_1) s_and_b32 exec_lo, exec_lo, s2 s_cbranch_execz .LBB0_15 v_mad_i64_i32 v[4:5], null, v0, 21, s[0:1] v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, v4, v1 v_add_co_ci_u32_e32 v1, vcc_lo, v5, v2, vcc_lo global_store_b8 v[0:1], v3, off .LBB0_15: s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z4jogoPA21_b .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 464 .amdhsa_kernarg_size 264 .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 1 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 10 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z4jogoPA21_b, .Lfunc_end0-_Z4jogoPA21_b .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: _Z4jogoPA21_b .private_segment_fixed_size: 464 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z4jogoPA21_b.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 10 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <iostream> #include <unistd.h> #include <hip/hip_runtime.h> #define size 21 // Tamanho da matrix // Exibe os pontos na tela __host__ void print(bool grid[][size]){ std::cout << "\n\n\n\n\n"; for(unsigned int i = 1; i < size-1; i++) { for(unsigned int j = 1; j < size-1; j++) std::cout << (grid[i][j]?"#":"_"); std::cout << std::endl; } } __host__ bool someoneAlive(bool grid[][size]){ for(unsigned int i=0; i < size; i++) for(unsigned int j=0; j < size; j++) if(grid[i][j]==true) return true; return false; } // Calcula a simulacao __global__ void jogo(bool grid[][size]){ int m=blockIdx.x*blockDim.x+threadIdx.x; int n=blockIdx.y*blockDim.y+threadIdx.y; if (m<size && n<size){ // printf("m: %d n: %d\n",m,n); bool grid_tmp[size][size] = {}; for(unsigned int i=0; i < size; i++){ for(unsigned int j=0; j < size; j++){ grid_tmp[i][j] = grid[i][j]; // printf("%d",grid[i][j]); } // printf("\n"); } unsigned int count = 0; for(int k = -1; k <= 1; k++) for(int l = -1; l <= 1; l++) if(k != 0 || l != 0) if(grid_tmp[m+k][n+l]) ++count; if(count < 2 || count > 3){ grid[m][n] = false; // printf("m: %d n: %d MORREU\n",m,n); // printf("count = %d\n", count); } else { if(count == 3){ grid[m][n] = true; // printf("m: %d n: %d REVIVEU\n",m,n); } } } return; } int main(){ // bool grid[size][size] = {}; // dados iniciais bool grid[size][size] = {}; // dados iniciais grid[ 5][ 7] = true; grid[ 6][ 8] = true; grid[ 8][ 8] = true; grid[ 6][ 9] = true; grid[ 8][10] = true; grid[ 9][10] = true; grid[ 8][11] = true; grid[10][11] = true; grid[10][12] = true; bool d_grid[size][size]; int mem_size = size*size*sizeof(bool); hipMalloc((void **) d_grid, mem_size); int nthreads = 7; dim3 blocks(size/nthreads+1,size/nthreads+1); dim3 threads(nthreads,nthreads); while(someoneAlive(grid)){ hipMemcpy(d_grid, grid, mem_size, hipMemcpyHostToDevice); jogo<<<blocks,threads>>>(d_grid); hipDeviceSynchronize(); hipMemcpy(grid, d_grid, mem_size, hipMemcpyDeviceToHost); print(grid); usleep(100000); return 0; } hipFree(d_grid); }
.text .file "q1try.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z5printPA21_b # -- Begin function _Z5printPA21_b .p2align 4, 0x90 .type _Z5printPA21_b,@function _Z5printPA21_b: # @_Z5printPA21_b .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdi, %rbx movl $_ZSt4cout, %edi movl $.L.str, %esi movl $5, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l addq $22, %rbx movl $1, %r15d jmp .LBB0_1 .p2align 4, 0x90 .LBB0_8: # in Loop: Header=BB0_1 Depth=1 movq %r14, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r14), %rax movq %r14, %rdi movl $10, %esi callq *48(%rax) .LBB0_9: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit # in Loop: Header=BB0_1 Depth=1 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r15 addq $21, %rbx cmpq $20, %r15 je .LBB0_10 .LBB0_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_2 Depth 2 xorl %r14d, %r14d jmp .LBB0_2 .p2align 4, 0x90 .LBB0_4: # in Loop: Header=BB0_2 Depth=2 movl $_ZSt4cout, %edi movl $1, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r14 cmpq $19, %r14 je .LBB0_5 .LBB0_2: # Parent Loop BB0_1 Depth=1 # => This Inner Loop Header: Depth=2 cmpb $0, (%rbx,%r14) movl $.L.str.2, %esi je .LBB0_4 # %bb.3: # in Loop: Header=BB0_2 Depth=2 movl $.L.str.1, %esi jmp .LBB0_4 .p2align 4, 0x90 .LBB0_5: # in Loop: Header=BB0_1 Depth=1 movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r14 testq %r14, %r14 je .LBB0_11 # %bb.6: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i # in Loop: Header=BB0_1 Depth=1 cmpb $0, 56(%r14) je .LBB0_8 # %bb.7: # in Loop: Header=BB0_1 Depth=1 movzbl 67(%r14), %eax jmp .LBB0_9 .LBB0_10: popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .LBB0_11: .cfi_def_cfa_offset 32 callq _ZSt16__throw_bad_castv .Lfunc_end0: .size _Z5printPA21_b, .Lfunc_end0-_Z5printPA21_b .cfi_endproc # -- End function .globl _Z12someoneAlivePA21_b # -- Begin function _Z12someoneAlivePA21_b .p2align 4, 0x90 .type _Z12someoneAlivePA21_b,@function _Z12someoneAlivePA21_b: # @_Z12someoneAlivePA21_b .cfi_startproc # %bb.0: leaq 1(%rdi), %rcx movb $1, %al xorl %edx, %edx jmp .LBB1_1 .p2align 4, 0x90 .LBB1_6: # %.critedge # in Loop: Header=BB1_1 Depth=1 cmpq $20, %rdx leaq 1(%rdx), %rsi setb %al addq $21, %rcx movq %rsi, %rdx cmpq $21, %rsi je .LBB1_7 .LBB1_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_3 Depth 2 leaq (%rdx,%rdx,4), %rsi leaq (%rdx,%rsi,4), %rsi cmpb $0, (%rdi,%rsi) jne .LBB1_7 # %bb.2: # %.lr.ph.preheader # in Loop: Header=BB1_1 Depth=1 movq $-1, %rsi .p2align 4, 0x90 .LBB1_3: # %.lr.ph # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 cmpq $19, %rsi je .LBB1_6 # %bb.4: # in Loop: Header=BB1_3 Depth=2 leaq 1(%rsi), %r8 cmpb $0, 1(%rcx,%rsi) movq %r8, %rsi je .LBB1_3 # %bb.5: # %._crit_edge # in Loop: Header=BB1_1 Depth=1 cmpq $20, %r8 jae .LBB1_6 .LBB1_7: # %.critedge26 andb $1, %al retq .Lfunc_end1: .size _Z12someoneAlivePA21_b, .Lfunc_end1-_Z12someoneAlivePA21_b .cfi_endproc # -- End function .globl _Z19__device_stub__jogoPA21_b # -- Begin function _Z19__device_stub__jogoPA21_b .p2align 4, 0x90 .type _Z19__device_stub__jogoPA21_b,@function _Z19__device_stub__jogoPA21_b: # @_Z19__device_stub__jogoPA21_b .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 $_Z4jogoPA21_b, %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_end2: .size _Z19__device_stub__jogoPA21_b, .Lfunc_end2-_Z19__device_stub__jogoPA21_b .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 $960, %rsp # imm = 0x3C0 .cfi_def_cfa_offset 976 .cfi_offset %rbx, -16 leaq 64(%rsp), %rdi movl $441, %edx # imm = 0x1B9 xorl %esi, %esi callq memset@PLT movb $1, 176(%rsp) movw $257, 198(%rsp) # imm = 0x101 movb $1, 240(%rsp) movw $257, 242(%rsp) # imm = 0x101 movb $1, 263(%rsp) movw $257, 285(%rsp) # imm = 0x101 leaq 512(%rsp), %rdi movl $441, %esi # imm = 0x1B9 callq hipMalloc cmpb $0, 64(%rsp) jne .LBB3_9 # %bb.1: # %.lr.ph.preheader.preheader movb $1, %al leaq 65(%rsp), %rcx xorl %edx, %edx .p2align 4, 0x90 .LBB3_3: # %.lr.ph.preheader # =>This Loop Header: Depth=1 # Child Loop BB3_4 Depth 2 movq %rdx, %rsi movq $-1, %rdx .p2align 4, 0x90 .LBB3_4: # %.lr.ph # Parent Loop BB3_3 Depth=1 # => This Inner Loop Header: Depth=2 cmpq $19, %rdx je .LBB3_7 # %bb.5: # in Loop: Header=BB3_4 Depth=2 leaq 1(%rdx), %rdi cmpb $0, 1(%rcx,%rdx) movq %rdi, %rdx je .LBB3_4 # %bb.6: # %._crit_edge # in Loop: Header=BB3_3 Depth=1 cmpq $20, %rdi jb .LBB3_8 .LBB3_7: # %.critedge.i # in Loop: Header=BB3_3 Depth=1 leaq 1(%rsi), %rdx cmpq $20, %rsi setb %al cmpq $21, %rdx je .LBB3_8 # %bb.2: # %.preheader.i # in Loop: Header=BB3_3 Depth=1 leaq (%rdx,%rdx,4), %rsi leaq (%rdx,%rsi,4), %rsi addq $21, %rcx cmpb $0, 64(%rsp,%rsi) je .LBB3_3 .LBB3_8: # %_Z12someoneAlivePA21_b.exit testb $1, %al je .LBB3_12 .LBB3_9: # %.critedge leaq 512(%rsp), %rbx leaq 64(%rsp), %rsi movl $441, %edx # imm = 0x1B9 movq %rbx, %rdi movl $1, %ecx callq hipMemcpy movabsq $17179869188, %rdi # imm = 0x400000004 movabsq $30064771079, %rdx # imm = 0x700000007 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_11 # %bb.10: movq %rbx, 56(%rsp) leaq 56(%rsp), %rax movq %rax, (%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 movq %rsp, %r9 movl $_Z4jogoPA21_b, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_11: callq hipDeviceSynchronize leaq 64(%rsp), %rbx leaq 512(%rsp), %rsi movl $441, %edx # imm = 0x1B9 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy movq %rbx, %rdi callq _Z5printPA21_b movl $100000, %edi # imm = 0x186A0 callq usleep .LBB3_13: xorl %eax, %eax addq $960, %rsp # imm = 0x3C0 .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB3_12: .cfi_def_cfa_offset 976 leaq 512(%rsp), %rdi callq hipFree jmp .LBB3_13 .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4jogoPA21_b, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "\n\n\n\n\n" .size .L.str, 6 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "#" .size .L.str.1, 2 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "_" .size .L.str.2, 2 .type _Z4jogoPA21_b,@object # @_Z4jogoPA21_b .section .rodata,"a",@progbits .globl _Z4jogoPA21_b .p2align 3, 0x0 _Z4jogoPA21_b: .quad _Z19__device_stub__jogoPA21_b .size _Z4jogoPA21_b, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4jogoPA21_b" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__jogoPA21_b .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _ZSt4cout .addrsig_sym _Z4jogoPA21_b .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_0019ba00_00000000-6_q1try.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3687: .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 .LFE3687: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "#" .LC1: .string "_" .LC2: .string "\n\n\n\n\n" .text .globl _Z5printPA21_b .type _Z5printPA21_b, @function _Z5printPA21_b: .LFB3682: .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 $8, %rsp .cfi_def_cfa_offset 64 movq %rdi, %r15 movl $5, %edx leaq .LC2(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT leaq 41(%r15), %rbp addq $440, %r15 leaq .LC1(%rip), %r14 leaq .LC0(%rip), %r13 leaq _ZSt4cout(%rip), %r12 jmp .L4 .L14: call _ZSt16__throw_bad_castv@PLT .L15: movzbl 67(%rbx), %esi .L9: movsbl %sil, %esi movq %r12, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addq $21, %rbp cmpq %r15, %rbp je .L3 .L4: leaq -19(%rbp), %rbx .L6: cmpb $0, (%rbx) movq %r13, %rsi cmove %r14, %rsi movl $1, %edx movq %r12, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $1, %rbx cmpq %rbx, %rbp jne .L6 movq (%r12), %rax movq -24(%rax), %rax movq 240(%r12,%rax), %rbx testq %rbx, %rbx je .L14 cmpb $0, 56(%rbx) jne .L15 movq %rbx, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%rbx), %rax movl $10, %esi movq %rbx, %rdi call *48(%rax) movl %eax, %esi jmp .L9 .L3: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3682: .size _Z5printPA21_b, .-_Z5printPA21_b .globl _Z12someoneAlivePA21_b .type _Z12someoneAlivePA21_b, @function _Z12someoneAlivePA21_b: .LFB3683: .cfi_startproc endbr64 leaq 21(%rdi), %rcx addq $462, %rdi .L17: leaq -21(%rcx), %rax .L19: movzbl (%rax), %edx testb %dl, %dl jne .L16 addq $1, %rax cmpq %rcx, %rax jne .L19 addq $21, %rcx cmpq %rdi, %rcx jne .L17 .L16: movl %edx, %eax ret .cfi_endproc .LFE3683: .size _Z12someoneAlivePA21_b, .-_Z12someoneAlivePA21_b .globl _Z27__device_stub__Z4jogoPA21_bPA21_b .type _Z27__device_stub__Z4jogoPA21_bPA21_b, @function _Z27__device_stub__Z4jogoPA21_bPA21_b: .LFB3709: .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 .L25 .L21: movq 88(%rsp), %rax subq %fs:40, %rax jne .L26 addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L25: .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 _Z4jogoPA21_b(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 112 jmp .L21 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE3709: .size _Z27__device_stub__Z4jogoPA21_bPA21_b, .-_Z27__device_stub__Z4jogoPA21_bPA21_b .globl _Z4jogoPA21_b .type _Z4jogoPA21_b, @function _Z4jogoPA21_b: .LFB3710: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z4jogoPA21_bPA21_b addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3710: .size _Z4jogoPA21_b, .-_Z4jogoPA21_b .globl main .type main, @function main: .LFB3684: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $944, %rsp .cfi_def_cfa_offset 960 movq %fs:40, %rax movq %rax, 936(%rsp) xorl %eax, %eax leaq 32(%rsp), %rdi movl $55, %ecx rep stosq movb $0, (%rdi) movb $1, 144(%rsp) movb $1, 166(%rsp) movb $1, 208(%rsp) movb $1, 167(%rsp) movb $1, 210(%rsp) movb $1, 231(%rsp) movb $1, 211(%rsp) movb $1, 253(%rsp) movb $1, 254(%rsp) leaq 480(%rsp), %rdi movl $441, %esi call cudaMalloc@PLT movl $4, 8(%rsp) movl $4, 12(%rsp) movl $1, 16(%rsp) movl $7, 20(%rsp) movl $7, 24(%rsp) movl $1, 28(%rsp) leaq 32(%rsp), %rdi call _Z12someoneAlivePA21_b testb %al, %al je .L30 leaq 32(%rsp), %rsi leaq 480(%rsp), %rdi movl $1, %ecx movl $441, %edx call cudaMemcpy@PLT movl 28(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 20(%rsp), %rdx movq 8(%rsp), %rdi movl 16(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L35 .L31: call cudaDeviceSynchronize@PLT leaq 480(%rsp), %rsi leaq 32(%rsp), %rbx movl $2, %ecx movl $441, %edx movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rdi call _Z5printPA21_b movl $100000, %edi call usleep@PLT .L32: movq 936(%rsp), %rax subq %fs:40, %rax jne .L36 movl $0, %eax addq $944, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L35: .cfi_restore_state leaq 480(%rsp), %rdi call _Z27__device_stub__Z4jogoPA21_bPA21_b jmp .L31 .L30: leaq 480(%rsp), %rdi call cudaFree@PLT jmp .L32 .L36: call __stack_chk_fail@PLT .cfi_endproc .LFE3684: .size main, .-main .section .rodata.str1.1 .LC3: .string "_Z4jogoPA21_b" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3712: .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 _Z4jogoPA21_b(%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 .LFE3712: .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 "q1try.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z5printPA21_b # -- Begin function _Z5printPA21_b .p2align 4, 0x90 .type _Z5printPA21_b,@function _Z5printPA21_b: # @_Z5printPA21_b .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %rbx .cfi_def_cfa_offset 32 .cfi_offset %rbx, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movq %rdi, %rbx movl $_ZSt4cout, %edi movl $.L.str, %esi movl $5, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l addq $22, %rbx movl $1, %r15d jmp .LBB0_1 .p2align 4, 0x90 .LBB0_8: # in Loop: Header=BB0_1 Depth=1 movq %r14, %rdi callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r14), %rax movq %r14, %rdi movl $10, %esi callq *48(%rax) .LBB0_9: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit # in Loop: Header=BB0_1 Depth=1 movsbl %al, %esi movl $_ZSt4cout, %edi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r15 addq $21, %rbx cmpq $20, %r15 je .LBB0_10 .LBB0_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB0_2 Depth 2 xorl %r14d, %r14d jmp .LBB0_2 .p2align 4, 0x90 .LBB0_4: # in Loop: Header=BB0_2 Depth=2 movl $_ZSt4cout, %edi movl $1, %edx callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l incq %r14 cmpq $19, %r14 je .LBB0_5 .LBB0_2: # Parent Loop BB0_1 Depth=1 # => This Inner Loop Header: Depth=2 cmpb $0, (%rbx,%r14) movl $.L.str.2, %esi je .LBB0_4 # %bb.3: # in Loop: Header=BB0_2 Depth=2 movl $.L.str.1, %esi jmp .LBB0_4 .p2align 4, 0x90 .LBB0_5: # in Loop: Header=BB0_1 Depth=1 movq _ZSt4cout(%rip), %rax movq -24(%rax), %rax movq _ZSt4cout+240(%rax), %r14 testq %r14, %r14 je .LBB0_11 # %bb.6: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i # in Loop: Header=BB0_1 Depth=1 cmpb $0, 56(%r14) je .LBB0_8 # %bb.7: # in Loop: Header=BB0_1 Depth=1 movzbl 67(%r14), %eax jmp .LBB0_9 .LBB0_10: popq %rbx .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 retq .LBB0_11: .cfi_def_cfa_offset 32 callq _ZSt16__throw_bad_castv .Lfunc_end0: .size _Z5printPA21_b, .Lfunc_end0-_Z5printPA21_b .cfi_endproc # -- End function .globl _Z12someoneAlivePA21_b # -- Begin function _Z12someoneAlivePA21_b .p2align 4, 0x90 .type _Z12someoneAlivePA21_b,@function _Z12someoneAlivePA21_b: # @_Z12someoneAlivePA21_b .cfi_startproc # %bb.0: leaq 1(%rdi), %rcx movb $1, %al xorl %edx, %edx jmp .LBB1_1 .p2align 4, 0x90 .LBB1_6: # %.critedge # in Loop: Header=BB1_1 Depth=1 cmpq $20, %rdx leaq 1(%rdx), %rsi setb %al addq $21, %rcx movq %rsi, %rdx cmpq $21, %rsi je .LBB1_7 .LBB1_1: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_3 Depth 2 leaq (%rdx,%rdx,4), %rsi leaq (%rdx,%rsi,4), %rsi cmpb $0, (%rdi,%rsi) jne .LBB1_7 # %bb.2: # %.lr.ph.preheader # in Loop: Header=BB1_1 Depth=1 movq $-1, %rsi .p2align 4, 0x90 .LBB1_3: # %.lr.ph # Parent Loop BB1_1 Depth=1 # => This Inner Loop Header: Depth=2 cmpq $19, %rsi je .LBB1_6 # %bb.4: # in Loop: Header=BB1_3 Depth=2 leaq 1(%rsi), %r8 cmpb $0, 1(%rcx,%rsi) movq %r8, %rsi je .LBB1_3 # %bb.5: # %._crit_edge # in Loop: Header=BB1_1 Depth=1 cmpq $20, %r8 jae .LBB1_6 .LBB1_7: # %.critedge26 andb $1, %al retq .Lfunc_end1: .size _Z12someoneAlivePA21_b, .Lfunc_end1-_Z12someoneAlivePA21_b .cfi_endproc # -- End function .globl _Z19__device_stub__jogoPA21_b # -- Begin function _Z19__device_stub__jogoPA21_b .p2align 4, 0x90 .type _Z19__device_stub__jogoPA21_b,@function _Z19__device_stub__jogoPA21_b: # @_Z19__device_stub__jogoPA21_b .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 $_Z4jogoPA21_b, %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_end2: .size _Z19__device_stub__jogoPA21_b, .Lfunc_end2-_Z19__device_stub__jogoPA21_b .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 $960, %rsp # imm = 0x3C0 .cfi_def_cfa_offset 976 .cfi_offset %rbx, -16 leaq 64(%rsp), %rdi movl $441, %edx # imm = 0x1B9 xorl %esi, %esi callq memset@PLT movb $1, 176(%rsp) movw $257, 198(%rsp) # imm = 0x101 movb $1, 240(%rsp) movw $257, 242(%rsp) # imm = 0x101 movb $1, 263(%rsp) movw $257, 285(%rsp) # imm = 0x101 leaq 512(%rsp), %rdi movl $441, %esi # imm = 0x1B9 callq hipMalloc cmpb $0, 64(%rsp) jne .LBB3_9 # %bb.1: # %.lr.ph.preheader.preheader movb $1, %al leaq 65(%rsp), %rcx xorl %edx, %edx .p2align 4, 0x90 .LBB3_3: # %.lr.ph.preheader # =>This Loop Header: Depth=1 # Child Loop BB3_4 Depth 2 movq %rdx, %rsi movq $-1, %rdx .p2align 4, 0x90 .LBB3_4: # %.lr.ph # Parent Loop BB3_3 Depth=1 # => This Inner Loop Header: Depth=2 cmpq $19, %rdx je .LBB3_7 # %bb.5: # in Loop: Header=BB3_4 Depth=2 leaq 1(%rdx), %rdi cmpb $0, 1(%rcx,%rdx) movq %rdi, %rdx je .LBB3_4 # %bb.6: # %._crit_edge # in Loop: Header=BB3_3 Depth=1 cmpq $20, %rdi jb .LBB3_8 .LBB3_7: # %.critedge.i # in Loop: Header=BB3_3 Depth=1 leaq 1(%rsi), %rdx cmpq $20, %rsi setb %al cmpq $21, %rdx je .LBB3_8 # %bb.2: # %.preheader.i # in Loop: Header=BB3_3 Depth=1 leaq (%rdx,%rdx,4), %rsi leaq (%rdx,%rsi,4), %rsi addq $21, %rcx cmpb $0, 64(%rsp,%rsi) je .LBB3_3 .LBB3_8: # %_Z12someoneAlivePA21_b.exit testb $1, %al je .LBB3_12 .LBB3_9: # %.critedge leaq 512(%rsp), %rbx leaq 64(%rsp), %rsi movl $441, %edx # imm = 0x1B9 movq %rbx, %rdi movl $1, %ecx callq hipMemcpy movabsq $17179869188, %rdi # imm = 0x400000004 movabsq $30064771079, %rdx # imm = 0x700000007 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_11 # %bb.10: movq %rbx, 56(%rsp) leaq 56(%rsp), %rax movq %rax, (%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 movq %rsp, %r9 movl $_Z4jogoPA21_b, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_11: callq hipDeviceSynchronize leaq 64(%rsp), %rbx leaq 512(%rsp), %rsi movl $441, %edx # imm = 0x1B9 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy movq %rbx, %rdi callq _Z5printPA21_b movl $100000, %edi # imm = 0x186A0 callq usleep .LBB3_13: xorl %eax, %eax addq $960, %rsp # imm = 0x3C0 .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 retq .LBB3_12: .cfi_def_cfa_offset 976 leaq 512(%rsp), %rdi callq hipFree jmp .LBB3_13 .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z4jogoPA21_b, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "\n\n\n\n\n" .size .L.str, 6 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "#" .size .L.str.1, 2 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "_" .size .L.str.2, 2 .type _Z4jogoPA21_b,@object # @_Z4jogoPA21_b .section .rodata,"a",@progbits .globl _Z4jogoPA21_b .p2align 3, 0x0 _Z4jogoPA21_b: .quad _Z19__device_stub__jogoPA21_b .size _Z4jogoPA21_b, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z4jogoPA21_b" .size .L__unnamed_1, 14 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z19__device_stub__jogoPA21_b .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _ZSt4cout .addrsig_sym _Z4jogoPA21_b .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void sum_grad(float* in, int inSize0, int inSize1, int inSize2, int inSize3, int nElement, float* out, int outStride0, int outStride1, int outStride2, int dim) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; for (int i = tid; i < nElement; i += stride) { int inOff2 = i / inSize3; int inDim3 = i - inOff2 * inSize3; int inOff1 = inOff2 / inSize2; int inDim2 = inOff2 - inOff1 * inSize2; int inDim0 = inOff1 / inSize1; int inDim1 = inOff1 - inDim0 * inSize1; int outOff = 0; if (dim == 0) outOff = inDim1 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 1) outOff = inDim0 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 2) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim3 * outStride2; if (dim == 3) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim2 * outStride2; in[i] += out[outOff]; } }
.file "tmpxft_0015b477_00000000-6_sum_grad.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 _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii .type _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii, @function _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii: .LFB2051: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movl %r9d, 20(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 28(%rsp), %rax movq %rax, 136(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) leaq 20(%rsp), %rax movq %rax, 152(%rsp) leaq 8(%rsp), %rax movq %rax, 160(%rsp) leaq 232(%rsp), %rax movq %rax, 168(%rsp) leaq 240(%rsp), %rax movq %rax, 176(%rsp) leaq 248(%rsp), %rax movq %rax, 184(%rsp) leaq 256(%rsp), %rax movq %rax, 192(%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 200(%rsp), %rax subq %fs:40, %rax jne .L8 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 232 pushq 56(%rsp) .cfi_def_cfa_offset 240 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z8sum_gradPfiiiiiS_iiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii, .-_Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii .globl _Z8sum_gradPfiiiiiS_iiii .type _Z8sum_gradPfiiiiiS_iiii, @function _Z8sum_gradPfiiiiiS_iiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 56 pushq 56(%rsp) .cfi_def_cfa_offset 64 call _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii addq $56, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z8sum_gradPfiiiiiS_iiii, .-_Z8sum_gradPfiiiiiS_iiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z8sum_gradPfiiiiiS_iiii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z8sum_gradPfiiiiiS_iiii(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include "includes.h" __global__ void sum_grad(float* in, int inSize0, int inSize1, int inSize2, int inSize3, int nElement, float* out, int outStride0, int outStride1, int outStride2, int dim) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; for (int i = tid; i < nElement; i += stride) { int inOff2 = i / inSize3; int inDim3 = i - inOff2 * inSize3; int inOff1 = inOff2 / inSize2; int inDim2 = inOff2 - inOff1 * inSize2; int inDim0 = inOff1 / inSize1; int inDim1 = inOff1 - inDim0 * inSize1; int outOff = 0; if (dim == 0) outOff = inDim1 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 1) outOff = inDim0 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 2) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim3 * outStride2; if (dim == 3) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim2 * outStride2; in[i] += out[outOff]; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void sum_grad(float* in, int inSize0, int inSize1, int inSize2, int inSize3, int nElement, float* out, int outStride0, int outStride1, int outStride2, int dim) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; for (int i = tid; i < nElement; i += stride) { int inOff2 = i / inSize3; int inDim3 = i - inOff2 * inSize3; int inOff1 = inOff2 / inSize2; int inDim2 = inOff2 - inOff1 * inSize2; int inDim0 = inOff1 / inSize1; int inDim1 = inOff1 - inDim0 * inSize1; int outOff = 0; if (dim == 0) outOff = inDim1 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 1) outOff = inDim0 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 2) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim3 * outStride2; if (dim == 3) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim2 * outStride2; in[i] += out[outOff]; } }
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 sum_grad(float* in, int inSize0, int inSize1, int inSize2, int inSize3, int nElement, float* out, int outStride0, int outStride1, int outStride2, int dim) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; for (int i = tid; i < nElement; i += stride) { int inOff2 = i / inSize3; int inDim3 = i - inOff2 * inSize3; int inOff1 = inOff2 / inSize2; int inDim2 = inOff2 - inOff1 * inSize2; int inDim0 = inOff1 / inSize1; int inDim1 = inOff1 - inDim0 * inSize1; int outOff = 0; if (dim == 0) outOff = inDim1 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 1) outOff = inDim0 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 2) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim3 * outStride2; if (dim == 3) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim2 * outStride2; in[i] += out[outOff]; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z8sum_gradPfiiiiiS_iiii .globl _Z8sum_gradPfiiiiiS_iiii .p2align 8 .type _Z8sum_gradPfiiiiiS_iiii,@function _Z8sum_gradPfiiiiiS_iiii: s_clause 0x1 s_load_b32 s2, s[0:1], 0x44 s_load_b32 s14, s[0:1], 0x18 s_add_u32 s4, s0, 56 s_addc_u32 s5, s1, 0 s_waitcnt lgkmcnt(0) s_and_b32 s10, s2, 0xffff s_mov_b32 s2, exec_lo v_mad_u64_u32 v[1:2], null, s15, s10, v[0:1] s_delay_alu instid0(VALU_DEP_1) v_cmpx_gt_i32_e64 s14, v1 s_cbranch_execz .LBB0_11 s_clause 0x2 s_load_b64 s[2:3], s[0:1], 0x30 s_load_b32 s24, s[0:1], 0x14 s_load_b64 s[8:9], s[0:1], 0xc s_load_b32 s11, s[4:5], 0x0 s_load_b64 s[12:13], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_delay_alu instid0(VALU_DEP_1) v_lshlrev_b64 v[2:3], 2, v[1:2] s_waitcnt lgkmcnt(0) s_cmp_eq_u32 s3, 0 s_cselect_b32 s15, -1, 0 s_cmp_eq_u32 s3, 1 s_mul_i32 s10, s11, s10 s_cselect_b32 s16, -1, 0 s_cmp_eq_u32 s3, 2 s_cselect_b32 s17, -1, 0 s_cmp_eq_u32 s3, 3 s_cselect_b32 s3, -1, 0 s_ashr_i32 s18, s24, 31 s_ashr_i32 s19, s9, 31 s_ashr_i32 s20, s8, 31 s_add_i32 s4, s24, s18 s_add_i32 s5, s9, s19 s_add_i32 s6, s8, s20 s_xor_b32 s21, s4, s18 s_xor_b32 s22, s5, s19 s_xor_b32 s23, s6, s20 s_load_b128 s[4:7], s[0:1], 0x20 v_cvt_f32_u32_e32 v0, s21 v_cvt_f32_u32_e32 v4, s22 v_cvt_f32_u32_e32 v5, s23 v_add_co_u32 v2, vcc_lo, s12, v2 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_3) v_rcp_iflag_f32_e32 v0, v0 v_rcp_iflag_f32_e32 v4, v4 s_delay_alu instid0(VALU_DEP_2) v_rcp_iflag_f32_e32 v5, v5 v_add_co_ci_u32_e32 v3, vcc_lo, s13, v3, vcc_lo s_ashr_i32 s11, s10, 31 s_mov_b32 s1, 0 s_lshl_b64 s[12:13], s[10:11], 2 s_sub_i32 s11, 0, s24 s_delay_alu instid0(TRANS32_DEP_3) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_mul_f32_e32 v0, 0x4f7ffffe, v0 s_waitcnt_depctr 0xfff v_dual_mul_f32 v4, 0x4f7ffffe, v4 :: v_dual_mul_f32 v5, 0x4f7ffffe, v5 v_cvt_u32_f32_e32 v0, v0 v_cvt_u32_f32_e32 v6, v4 s_delay_alu instid0(VALU_DEP_3) v_cvt_u32_f32_e32 v7, v5 s_branch .LBB0_3 .LBB0_2: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v5, 31, v4 v_lshlrev_b64 v[4:5], 2, v[4:5] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v4, vcc_lo, s4, v4 v_add_co_ci_u32_e32 v5, vcc_lo, s5, v5, vcc_lo global_load_b32 v8, v[2:3], off global_load_b32 v4, v[4:5], off s_waitcnt vmcnt(0) v_dual_add_f32 v4, v4, v8 :: v_dual_add_nc_u32 v1, s10, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_cmp_le_i32_e32 vcc_lo, s14, v1 global_store_b32 v[2:3], v4, off v_add_co_u32 v2, s0, v2, s12 v_add_co_ci_u32_e64 v3, s0, s13, v3, s0 s_or_b32 s1, vcc_lo, s1 s_delay_alu instid0(SALU_CYCLE_1) s_and_not1_b32 exec_lo, exec_lo, s1 s_cbranch_execz .LBB0_11 .LBB0_3: s_sub_i32 s0, 0, s21 v_ashrrev_i32_e32 v5, 31, v1 v_mul_lo_u32 v4, s0, v0 s_sub_i32 s0, 0, s22 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v8, v1, v5 v_mul_hi_u32 v4, v0, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_xor_b32_e32 v8, v8, v5 v_xor_b32_e32 v5, s18, v5 v_add_nc_u32_e32 v4, v0, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v4, v8, v4 v_mul_lo_u32 v9, v4, s21 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_sub_nc_u32_e32 v8, v8, v9 v_subrev_nc_u32_e32 v10, s21, v8 v_cmp_le_u32_e32 vcc_lo, s21, v8 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v8, v8, v10 :: v_dual_add_nc_u32 v9, 1, v4 v_cndmask_b32_e32 v4, v4, v9, vcc_lo s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_cmp_le_u32_e32 vcc_lo, s21, v8 v_add_nc_u32_e32 v9, 1, v4 v_mul_lo_u32 v8, s0, v6 s_sub_i32 s0, 0, s23 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e32 v4, v4, v9, vcc_lo v_xor_b32_e32 v4, v4, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_4) v_sub_nc_u32_e32 v10, v4, v5 v_mul_hi_u32 v5, v6, v8 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_ashrrev_i32_e32 v4, 31, v10 v_add_nc_u32_e32 v5, v6, v5 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v8, v10, v4 v_xor_b32_e32 v8, v8, v4 v_xor_b32_e32 v4, s19, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_hi_u32 v5, v8, v5 v_mul_lo_u32 v9, v5, s22 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v8, v8, v9 v_add_nc_u32_e32 v9, 1, v5 v_subrev_nc_u32_e32 v11, s22, v8 v_cmp_le_u32_e32 vcc_lo, s22, v8 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_dual_cndmask_b32 v5, v5, v9 :: v_dual_cndmask_b32 v8, v8, v11 v_add_nc_u32_e32 v9, 1, v5 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_3) v_cmp_le_u32_e32 vcc_lo, s22, v8 v_mul_lo_u32 v8, s0, v7 v_cndmask_b32_e32 v5, v5, v9, vcc_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_xor_b32_e32 v5, v5, v4 v_sub_nc_u32_e32 v11, v5, v4 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_2) v_mul_hi_u32 v5, v7, v8 v_ashrrev_i32_e32 v4, 31, v11 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v5, v7, v5 v_add_nc_u32_e32 v8, v11, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_xor_b32_e32 v8, v8, v4 v_xor_b32_e32 v4, s20, v4 v_mul_hi_u32 v5, v8, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_lo_u32 v9, v5, s23 v_sub_nc_u32_e32 v8, v8, v9 v_add_nc_u32_e32 v9, 1, v5 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_subrev_nc_u32_e32 v12, s23, v8 v_cmp_le_u32_e32 vcc_lo, s23, v8 v_dual_cndmask_b32 v5, v5, v9 :: v_dual_cndmask_b32 v8, v8, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v9, 1, v5 v_cmp_le_u32_e32 vcc_lo, s23, v8 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_cndmask_b32_e32 v5, v5, v9, vcc_lo s_and_not1_b32 vcc_lo, exec_lo, s15 v_xor_b32_e32 v5, v5, v4 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_sub_nc_u32_e32 v9, v5, v4 v_mad_u64_u32 v[4:5], null, s11, v10, v[1:2] v_mul_lo_u32 v5, v11, s9 v_mul_lo_u32 v12, v9, s8 s_delay_alu instid0(VALU_DEP_3) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_mul_lo_u32 v8, v4, s2 v_mov_b32_e32 v4, 0 v_sub_nc_u32_e32 v5, v10, v5 s_delay_alu instid0(VALU_DEP_4) v_sub_nc_u32_e32 v10, v11, v12 s_cbranch_vccnz .LBB0_5 s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mul_lo_u32 v4, v10, s6 v_mul_lo_u32 v11, v5, s7 v_add3_u32 v4, v11, v8, v4 .LBB0_5: s_waitcnt lgkmcnt(0) v_mul_lo_u32 v9, v9, s6 s_and_not1_b32 vcc_lo, exec_lo, s16 s_cbranch_vccnz .LBB0_7 v_mul_lo_u32 v4, v5, s7 s_delay_alu instid0(VALU_DEP_1) v_add3_u32 v4, v8, v9, v4 .LBB0_7: v_mul_lo_u32 v10, v10, s7 s_and_not1_b32 vcc_lo, exec_lo, s17 s_cbranch_vccnz .LBB0_9 s_delay_alu instid0(VALU_DEP_1) v_add3_u32 v4, v8, v9, v10 .LBB0_9: s_and_not1_b32 vcc_lo, exec_lo, s3 s_cbranch_vccnz .LBB0_2 v_mul_lo_u32 v4, v5, s2 s_delay_alu instid0(VALU_DEP_1) v_add3_u32 v4, v4, v9, v10 s_branch .LBB0_2 .LBB0_11: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z8sum_gradPfiiiiiS_iiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 312 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 13 .amdhsa_next_free_sgpr 25 .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 _Z8sum_gradPfiiiiiS_iiii, .Lfunc_end0-_Z8sum_gradPfiiiiiS_iiii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .offset: 16 .size: 4 .value_kind: by_value - .offset: 20 .size: 4 .value_kind: by_value - .offset: 24 .size: 4 .value_kind: by_value - .address_space: global .offset: 32 .size: 8 .value_kind: global_buffer - .offset: 40 .size: 4 .value_kind: by_value - .offset: 44 .size: 4 .value_kind: by_value - .offset: 48 .size: 4 .value_kind: by_value - .offset: 52 .size: 4 .value_kind: by_value - .offset: 56 .size: 4 .value_kind: hidden_block_count_x - .offset: 60 .size: 4 .value_kind: hidden_block_count_y - .offset: 64 .size: 4 .value_kind: hidden_block_count_z - .offset: 68 .size: 2 .value_kind: hidden_group_size_x - .offset: 70 .size: 2 .value_kind: hidden_group_size_y - .offset: 72 .size: 2 .value_kind: hidden_group_size_z - .offset: 74 .size: 2 .value_kind: hidden_remainder_x - .offset: 76 .size: 2 .value_kind: hidden_remainder_y - .offset: 78 .size: 2 .value_kind: hidden_remainder_z - .offset: 96 .size: 8 .value_kind: hidden_global_offset_x - .offset: 104 .size: 8 .value_kind: hidden_global_offset_y - .offset: 112 .size: 8 .value_kind: hidden_global_offset_z - .offset: 120 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 312 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z8sum_gradPfiiiiiS_iiii .private_segment_fixed_size: 0 .sgpr_count: 27 .sgpr_spill_count: 0 .symbol: _Z8sum_gradPfiiiiiS_iiii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 13 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void sum_grad(float* in, int inSize0, int inSize1, int inSize2, int inSize3, int nElement, float* out, int outStride0, int outStride1, int outStride2, int dim) { int tid = blockIdx.x * blockDim.x + threadIdx.x; int stride = gridDim.x * blockDim.x; for (int i = tid; i < nElement; i += stride) { int inOff2 = i / inSize3; int inDim3 = i - inOff2 * inSize3; int inOff1 = inOff2 / inSize2; int inDim2 = inOff2 - inOff1 * inSize2; int inDim0 = inOff1 / inSize1; int inDim1 = inOff1 - inDim0 * inSize1; int outOff = 0; if (dim == 0) outOff = inDim1 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 1) outOff = inDim0 * outStride0 + inDim2 * outStride1 + inDim3 * outStride2; if (dim == 2) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim3 * outStride2; if (dim == 3) outOff = inDim0 * outStride0 + inDim1 * outStride1 + inDim2 * outStride2; in[i] += out[outOff]; } }
.text .file "sum_grad.hip" .globl _Z23__device_stub__sum_gradPfiiiiiS_iiii # -- Begin function _Z23__device_stub__sum_gradPfiiiiiS_iiii .p2align 4, 0x90 .type _Z23__device_stub__sum_gradPfiiiiiS_iiii,@function _Z23__device_stub__sum_gradPfiiiiiS_iiii: # @_Z23__device_stub__sum_gradPfiiiiiS_iiii .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 72(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movl %ecx, 12(%rsp) movl %r8d, 8(%rsp) movl %r9d, 4(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 20(%rsp), %rax movq %rax, 88(%rsp) leaq 16(%rsp), %rax movq %rax, 96(%rsp) leaq 12(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq 176(%rsp), %rax movq %rax, 128(%rsp) leaq 184(%rsp), %rax movq %rax, 136(%rsp) leaq 192(%rsp), %rax movq %rax, 144(%rsp) leaq 200(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8sum_gradPfiiiiiS_iiii, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z23__device_stub__sum_gradPfiiiiiS_iiii, .Lfunc_end0-_Z23__device_stub__sum_gradPfiiiiiS_iiii .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8sum_gradPfiiiiiS_iiii, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z8sum_gradPfiiiiiS_iiii,@object # @_Z8sum_gradPfiiiiiS_iiii .section .rodata,"a",@progbits .globl _Z8sum_gradPfiiiiiS_iiii .p2align 3, 0x0 _Z8sum_gradPfiiiiiS_iiii: .quad _Z23__device_stub__sum_gradPfiiiiiS_iiii .size _Z8sum_gradPfiiiiiS_iiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8sum_gradPfiiiiiS_iiii" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__sum_gradPfiiiiiS_iiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8sum_gradPfiiiiiS_iiii .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_0015b477_00000000-6_sum_grad.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 _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii .type _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii, @function _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii: .LFB2051: .cfi_startproc endbr64 subq $216, %rsp .cfi_def_cfa_offset 224 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movl %ecx, 28(%rsp) movl %r8d, 24(%rsp) movl %r9d, 20(%rsp) movq 224(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 200(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 28(%rsp), %rax movq %rax, 136(%rsp) leaq 24(%rsp), %rax movq %rax, 144(%rsp) leaq 20(%rsp), %rax movq %rax, 152(%rsp) leaq 8(%rsp), %rax movq %rax, 160(%rsp) leaq 232(%rsp), %rax movq %rax, 168(%rsp) leaq 240(%rsp), %rax movq %rax, 176(%rsp) leaq 248(%rsp), %rax movq %rax, 184(%rsp) leaq 256(%rsp), %rax movq %rax, 192(%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 200(%rsp), %rax subq %fs:40, %rax jne .L8 addq $216, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 232 pushq 56(%rsp) .cfi_def_cfa_offset 240 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z8sum_gradPfiiiiiS_iiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 224 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii, .-_Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii .globl _Z8sum_gradPfiiiiiS_iiii .type _Z8sum_gradPfiiiiiS_iiii, @function _Z8sum_gradPfiiiiiS_iiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 56 pushq 56(%rsp) .cfi_def_cfa_offset 64 call _Z38__device_stub__Z8sum_gradPfiiiiiS_iiiiPfiiiiiS_iiii addq $56, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z8sum_gradPfiiiiiS_iiii, .-_Z8sum_gradPfiiiiiS_iiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z8sum_gradPfiiiiiS_iiii" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2054: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z8sum_gradPfiiiiiS_iiii(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "sum_grad.hip" .globl _Z23__device_stub__sum_gradPfiiiiiS_iiii # -- Begin function _Z23__device_stub__sum_gradPfiiiiiS_iiii .p2align 4, 0x90 .type _Z23__device_stub__sum_gradPfiiiiiS_iiii,@function _Z23__device_stub__sum_gradPfiiiiiS_iiii: # @_Z23__device_stub__sum_gradPfiiiiiS_iiii .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 72(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movl %ecx, 12(%rsp) movl %r8d, 8(%rsp) movl %r9d, 4(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 20(%rsp), %rax movq %rax, 88(%rsp) leaq 16(%rsp), %rax movq %rax, 96(%rsp) leaq 12(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) leaq 176(%rsp), %rax movq %rax, 128(%rsp) leaq 184(%rsp), %rax movq %rax, 136(%rsp) leaq 192(%rsp), %rax movq %rax, 144(%rsp) leaq 200(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z8sum_gradPfiiiiiS_iiii, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z23__device_stub__sum_gradPfiiiiiS_iiii, .Lfunc_end0-_Z23__device_stub__sum_gradPfiiiiiS_iiii .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB1_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB1_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z8sum_gradPfiiiiiS_iiii, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end1: .size __hip_module_ctor, .Lfunc_end1-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB2_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB2_2: retq .Lfunc_end2: .size __hip_module_dtor, .Lfunc_end2-__hip_module_dtor .cfi_endproc # -- End function .type _Z8sum_gradPfiiiiiS_iiii,@object # @_Z8sum_gradPfiiiiiS_iiii .section .rodata,"a",@progbits .globl _Z8sum_gradPfiiiiiS_iiii .p2align 3, 0x0 _Z8sum_gradPfiiiiiS_iiii: .quad _Z23__device_stub__sum_gradPfiiiiiS_iiii .size _Z8sum_gradPfiiiiiS_iiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z8sum_gradPfiiiiiS_iiii" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z23__device_stub__sum_gradPfiiiiiS_iiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z8sum_gradPfiiiiiS_iiii .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cuda_runtime.h> #include <iostream> #include <vector> #include <utility> #include <stdio.h> #include <math.h> using namespace std; #define K 3 #define BLCH 8 #define BLCW 32 #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); } inline void gpuAssert(cudaError_t code, const char *file, int line, bool abort=true) { if (code != cudaSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line); if (abort) exit(code); } } // declaration of constant memory where the fiter values are stored __constant__ float cm[K*K]; __device__ void conv(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k, float* sm, int gID, int tID, int nT, int rel_row, int rel_col, int nRows, int stopPrefetchRowID, int lastActiveThreadID) { for(int i=k; i<=nRows; i++) { /* ----prefetch a pixel value from GM and store it in register---- all threads fetch the cell value immediately below to the current cell iteratively last thread in the block would fetch k cells immediately below the current cell boundary check would be needed for the blocks that act on the bottom most partition of the input image to prevent it from prefetching out of image values. */ float reg; float regArr[K]; if(i <= stopPrefetchRowID){ reg = gm[i * iw + gID]; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ regArr[j] = gm[(i * iw) + gID + j]; } } } // load k * k pixels above the current cell float imgPixels[K*K]; for(int r=i-k; r<i; r++){ for(int c=0; c<k; c++){ /* translate the indices to [0,k] using r - (i-k) as imgPixels is of size k*k */ imgPixels[(r-i+k)*k + c] = sm[r * smW + tID + c]; } } /*multiply image pixel values with filter values (direct convolution) */ float convolvedCell = 0.0; for(int c=0; c<k*k; c++){ convolvedCell += cm[c]*imgPixels[c]; } //place the convolvedCell value into convolvedMatrix int cID = ( ( (rel_row * bh) + (i-k) ) * cw )+( rel_col * nT )+tID; if(cID < 0 || cID >= ch*cw ) { printf("cID : %d, tID : %d, gID : %d\n", cID, tID, gID ); } convolved[cID] = convolvedCell; __syncthreads(); if(i <= stopPrefetchRowID){ sm[i * smW + tID] = reg; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ int sID = i *smW + tID + j; sm[sID] = regArr[j]; } } } __syncthreads(); } } __global__ void conv_kernel(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k) { int tID = threadIdx.x; int bID = blockIdx.x; int nT = blockDim.x; int nB = gridDim.x; int nBx = iw / nT; //printf("num of blocks is %d\n", nB); //printf("nB in a row is %d\n", nBx); //check for right border or bottom border thread block bool isBottomBorder = false; bool isRightBorder = false; // bottom border thread block if(bID >= nB - nBx) { //printf("bID : %d is bottom border\n", bID); isBottomBorder = true; } // right border thread block if((bID+1) % nBx == 0){ //printf("bID : %d is right border\n", bID); isRightBorder = true; } // ---------------- Load k rows from GM into SM ---------------------- __shared__ float sm[ (BLCH + K - 1) * (BLCW + K - 1) ]; // rel_row and rel_col maps the Thread Block to appropriate position int rel_row = bID / nBx; int rel_col = bID % nBx; // (rel_row * bh * iw) covers all the cells before row_ids bh, 2bh, 3bh .. // gID finally maps threads to cells at rows 0, bh, 2bh, 3bh, ... int gID = (rel_row * bh * iw) + (rel_col * nT) + tID; for(int i=0; i<k; i++){ int sID = i * smW + tID; sm[sID] = gm[i * iw + gID]; /* if last thread in the block, it should fetch additional k-1 pixels in each row which are needed for computation of the convolution */ if(!isRightBorder && tID == nT-1){ for(int j=1; j<=k-1; j++){ sID = (i * smW) + tID + j; sm[sID] = gm[i * iw + gID + j]; } } } __syncthreads(); if( !isBottomBorder && !isRightBorder ){ int lastActiveThreadID = nT - 1; int nRows = bh + k - 1; int stopPrefetchRowID = nRows; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isBottomBorder && isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh; int stopPrefetchRowID = nRows - 1; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } else if( isBottomBorder ){ int nRows = bh; int stopPrefetchRowID = nRows-1; int lastActiveThreadID = nT - 1; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh + k - 1; int stopPrefetchRowID = nRows; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } } int main(int argc, char **argv){ /* set values for image dimensions, block dimensions, filter size, stride .. some of the constraints to keep in mind are 1. the value of k(filter size) should be less than blcH and blcW 2. stride value(s) should be 1 */ int imgH = 2048; int imgW = 2048; int blcH = BLCH; int blcW = BLCW; int k = K; int s = 1; int nB = (imgH * imgW) / (blcH * blcW); int nT = blcW; int imgDims = imgH * imgW; int imgSize = imgDims * sizeof(float); // create host array that can hold pixel intensity values float *h_img = new float[imgDims]; for(int i=0; i<imgDims; i++){ h_img[i] = 1.0; } // create device array that can hold pixel intensity values in GPU GM float *d_img; gpuErrchk(cudaMalloc((void **) &d_img, imgSize )); gpuErrchk(cudaMemcpy(d_img, h_img, imgSize, cudaMemcpyHostToDevice)); // create filter and copy to constant memory int filterDims = k * k; int filterSize = filterDims * sizeof(float); float *filter = new float[filterDims]; for(int i=0; i<filterDims; i++){ filter[i] = 0.5; } gpuErrchk(cudaMemcpyToSymbol(cm, filter, filterSize)); // create host and device array that holds the convoluted matrix int convH = ( (imgH - k) / s ) + 1; int convW = convH; int convDims = convH * convW; int convSize = convDims * sizeof(float); float *h_convolved = new float[convDims]; for(int i=0; i<convDims; i++){ h_convolved[i] = 0; } float *d_convolved; gpuErrchk(cudaMalloc((void **) &d_convolved, convSize)); gpuErrchk(cudaMemcpy(d_convolved, h_convolved, convSize, cudaMemcpyHostToDevice)); // calculate shared memory dimensions int smH = blcH + k - 1; int smW = blcW + k - 1; // call the kernel conv_kernel<<<nB, nT>>>(d_img, d_convolved, blcH, blcW, imgH, imgW, convH, convW, smH, smW, k); gpuErrchk(cudaMemcpy(h_convolved, d_convolved, convSize, cudaMemcpyDeviceToHost)); vector<pair<int,int> > miss; for(int i=0; i<convH; i++){ for(int j=0; j<convW; j++){ //cout<<h_convolved[i*convW +j]<<" "; if(h_convolved[i*convW +j] != 4.5){ miss.push_back(make_pair(i,j)); } } //cout<<"\n"; } cout<<miss.size()<<"\n"; for(int i=0;i<miss.size();i++){ cout<<miss[i].first<<","<<miss[i].second<<"\n"; } cudaDeviceReset(); delete h_img; delete h_convolved; return 0; }
.file "tmpxft_000691b4_00000000-6_convolution.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB4057: .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 .LFE4057: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata._Z9gpuAssert9cudaErrorPKcib.str1.1,"aMS",@progbits,1 .LC0: .string "GPUassert: %s %s %d\n" .section .text._Z9gpuAssert9cudaErrorPKcib,"axG",@progbits,_Z9gpuAssert9cudaErrorPKcib,comdat .weak _Z9gpuAssert9cudaErrorPKcib .type _Z9gpuAssert9cudaErrorPKcib, @function _Z9gpuAssert9cudaErrorPKcib: .LFB4032: .cfi_startproc endbr64 testl %edi, %edi jne .L9 ret .L9: 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 movl %edi, %ebx movq %rsi, %r13 movl %edx, %r12d movl %ecx, %ebp call cudaGetErrorString@PLT movq %rax, %rcx movl %r12d, %r9d movq %r13, %r8 leaq .LC0(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT testb %bpl, %bpl jne .L10 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 .L10: .cfi_restore_state movl %ebx, %edi call exit@PLT .cfi_endproc .LFE4032: .size _Z9gpuAssert9cudaErrorPKcib, .-_Z9gpuAssert9cudaErrorPKcib .text .globl _Z4convPKfPfiiiiiiiiiS1_iiiiiiii .type _Z4convPKfPfiiiiiiiiiS1_iiiiiiii, @function _Z4convPKfPfiiiiiiiiiS1_iiiiiiii: .LFB4033: .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 .LFE4033: .size _Z4convPKfPfiiiiiiiiiS1_iiiiiiii, .-_Z4convPKfPfiiiiiiiiiS1_iiiiiiii .globl _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .type _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii, @function _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii: .LFB4079: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 208(%rsp), %rax movq %rax, 144(%rsp) leaq 216(%rsp), %rax movq %rax, 152(%rsp) leaq 224(%rsp), %rax movq %rax, 160(%rsp) leaq 232(%rsp), %rax movq %rax, 168(%rsp) leaq 240(%rsp), %rax movq %rax, 176(%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 .L17 .L13: movq 184(%rsp), %rax subq %fs:40, %rax jne .L18 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 216 pushq 40(%rsp) .cfi_def_cfa_offset 224 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11conv_kernelPKfPfiiiiiiiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L13 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE4079: .size _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii, .-_Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .globl _Z11conv_kernelPKfPfiiiiiiiii .type _Z11conv_kernelPKfPfiiiiiiiii, @function _Z11conv_kernelPKfPfiiiiiiiii: .LFB4080: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 56 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 64 call _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii addq $56, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE4080: .size _Z11conv_kernelPKfPfiiiiiiiii, .-_Z11conv_kernelPKfPfiiiiiiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "_Z11conv_kernelPKfPfiiiiiiiii" .LC2: .string "cm" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB4082: .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 .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z11conv_kernelPKfPfiiiiiiiii(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $36, %r9d movl $0, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _ZL2cm(%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 .LFE4082: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .text._ZNSt6vectorISt4pairIiiESaIS1_EED2Ev,"axG",@progbits,_ZNSt6vectorISt4pairIiiESaIS1_EED5Ev,comdat .align 2 .weak _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .type _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev, @function _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev: .LFB4393: .cfi_startproc endbr64 movq (%rdi), %rax testq %rax, %rax je .L26 subq $8, %rsp .cfi_def_cfa_offset 16 movq 16(%rdi), %rsi subq %rax, %rsi movq %rax, %rdi call _ZdlPvm@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .L26: ret .cfi_endproc .LFE4393: .size _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev, .-_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .weak _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev .set _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev,_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .section .rodata._ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_.str1.1,"aMS",@progbits,1 .LC3: .string "vector::_M_realloc_insert" .section .text._ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_,"axG",@progbits,_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_,comdat .align 2 .weak _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .type _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_, @function _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_: .LFB4652: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $40, %rsp .cfi_def_cfa_offset 96 movq %rdx, 8(%rsp) movq 8(%rdi), %rbp movq (%rdi), %r12 movq %rbp, %rax subq %r12, %rax sarq $3, %rax movabsq $1152921504606846975, %rdx cmpq %rdx, %rax je .L47 movq %rdi, %r14 movq %rsi, %r13 movq %rsi, %rbx cmpq %r12, %rbp movl $1, %edx cmovne %rax, %rdx addq %rdx, %rax jc .L32 movabsq $1152921504606846975, %rdx cmpq %rdx, %rax cmovbe %rax, %rdx movq %rdx, 16(%rsp) movq %rsi, %rdi subq %r12, %rdi movq %rdi, 24(%rsp) movl $0, %r15d testq %rax, %rax je .L33 jmp .L39 .L47: leaq .LC3(%rip), %rdi call _ZSt20__throw_length_errorPKc@PLT .L42: movq %r15, %rsi jmp .L34 .L32: movq %rsi, %rax subq %r12, %rax movq %rax, 24(%rsp) movabsq $1152921504606846975, %rax movq %rax, 16(%rsp) .L39: movq 16(%rsp), %rax leaq 0(,%rax,8), %rdi call _Znwm@PLT movq %rax, %r15 .L33: movq 8(%rsp), %rax movq (%rax), %rax movq 24(%rsp), %rdi movq %rax, (%r15,%rdi) cmpq %r12, %r13 je .L42 movq %r13, %rsi subq %r12, %rsi addq %r15, %rsi movq %r15, %rax movq %r12, %rdx .L35: movq (%rdx), %rcx movq %rcx, (%rax) addq $8, %rdx addq $8, %rax cmpq %rax, %rsi jne .L35 .L34: leaq 8(%rsi), %rax movq %rax, 8(%rsp) cmpq %rbp, %r13 je .L36 movq %rbp, %rcx subq %r13, %rcx .L37: movq (%rbx), %rdx movq %rdx, (%rax) addq $8, %rbx addq $8, %rax cmpq %rbp, %rbx jne .L37 addq %rcx, 8(%rsp) .L36: testq %r12, %r12 je .L38 movq 16(%r14), %rsi subq %r12, %rsi movq %r12, %rdi call _ZdlPvm@PLT .L38: movq %r15, (%r14) movq 8(%rsp), %rax movq %rax, 8(%r14) movq 16(%rsp), %rax leaq (%r15,%rax,8), %rax movq %rax, 16(%r14) addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE4652: .size _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_, .-_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC5: .string "/home/ubuntu/Datasets/stackv2/train-structured/yottabytt/convolution_kernel/master/convolution.cu" .section .rodata.str1.1 .LC9: .string "\n" .LC10: .string "," .text .globl main .type main, @function main: .LFB4034: .cfi_startproc .cfi_personality 0x9b,DW.ref.__gxx_personality_v0 .cfi_lsda 0x1b,.LLSDA4034 endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $88, %rsp .cfi_def_cfa_offset 144 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $16777216, %edi .LEHB0: call _Znam@PLT movq %rax, %rbp leaq 16777216(%rax), %rdx movss .LC4(%rip), %xmm0 .L49: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L49 leaq 16(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT movl %eax, %edi movl $1, %ecx movl $233, %edx leaq .LC5(%rip), %rbx movq %rbx, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $1, %ecx movl $16777216, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $234, %edx movq %rbx, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $36, %edi call _Znam@PLT movq %rax, %rsi leaq 36(%rax), %rdx movss .LC6(%rip), %xmm0 .L50: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L50 movl $1, %r8d movl $0, %ecx movl $36, %edx leaq _ZL2cm(%rip), %rdi call cudaMemcpyToSymbol@PLT movl %eax, %edi movl $1, %ecx movl $242, %edx leaq .LC5(%rip), %rsi call _Z9gpuAssert9cudaErrorPKcib movl $16744464, %edi call _Znam@PLT movq %rax, %rbx movq %rax, %r13 leaq 16744464(%rax), %rdx .L51: movl $0x00000000, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L51 leaq 24(%rsp), %rdi movl $16744464, %esi call cudaMalloc@PLT movl %eax, %edi movl $1, %ecx movl $253, %edx leaq .LC5(%rip), %r12 movq %r12, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $1, %ecx movl $16744464, %edx movq %rbx, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $254, %edx movq %r12, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $32, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $16384, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $0, %r9d movl $0, %r8d movq 48(%rsp), %rdx movl $1, %ecx movq 36(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L74 .L52: movl $2, %ecx movl $16744464, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $266, %edx leaq .LC5(%rip), %rsi call _Z9gpuAssert9cudaErrorPKcib movq $0, 48(%rsp) movq $0, 56(%rsp) movq $0, 64(%rsp) movl $0, %r14d leaq 36(%rsp), %r15 jmp .L53 .L74: subq $8, %rsp .cfi_def_cfa_offset 152 pushq $3 .cfi_def_cfa_offset 160 pushq $34 .cfi_def_cfa_offset 168 pushq $10 .cfi_def_cfa_offset 176 pushq $2046 .cfi_def_cfa_offset 184 pushq $2046 .cfi_def_cfa_offset 192 movl $2048, %r9d movl $2048, %r8d movl $32, %ecx movl $8, %edx movq 72(%rsp), %rsi movq 64(%rsp), %rdi call _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .LEHE0: addq $48, %rsp .cfi_def_cfa_offset 144 jmp .L52 .L56: leaq 48(%rsp), %rdi movq %r15, %rdx .LEHB1: call _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .L54: addq $1, %r12 cmpq $2046, %r12 je .L75 .L58: movss 0(%r13,%r12,4), %xmm0 ucomiss .LC8(%rip), %xmm0 jp .L66 je .L54 .L66: movl %r14d, 36(%rsp) movl %r12d, 40(%rsp) movq 56(%rsp), %rsi cmpq 64(%rsp), %rsi je .L56 movq 36(%rsp), %rax movq %rax, (%rsi) addq $8, %rsi movq %rsi, 56(%rsp) jmp .L54 .L75: addl $1, %r14d addq $8184, %r13 cmpl $2046, %r14d je .L59 .L53: movl $0, %r12d jmp .L58 .L59: movq 56(%rsp), %r12 movq 48(%rsp), %r14 movq %r12, %r13 subq %r14, %r13 sarq $3, %r13 movq %r13, %rsi leaq _ZSt4cout(%rip), %rdi call _ZNSo9_M_insertImEERSoT_@PLT movq %rax, %rdi leaq .LC9(%rip), %rsi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT cmpq %r14, %r12 je .L60 movl $0, %r12d leaq _ZSt4cout(%rip), %r15 jmp .L61 .L76: movl $1, %edx leaq .LC10(%rip), %rsi movq %rax, 8(%rsp) movq %rax, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT movl 4(%r14,%r12,8), %esi movq 8(%rsp), %rdi call _ZNSolsEi@PLT movq %rax, %rdi movl $1, %edx leaq .LC9(%rip), %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $1, %r12 cmpq %r13, %r12 jnb .L60 .L61: movl (%r14,%r12,8), %esi movq %r15, %rdi call _ZNSolsEi@PLT jmp .L76 .L60: call cudaDeviceReset@PLT .LEHE1: movl $4, %esi movq %rbp, %rdi call _ZdlPvm@PLT movl $4, %esi movq %rbx, %rdi call _ZdlPvm@PLT leaq 48(%rsp), %rdi call _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev movq 72(%rsp), %rax subq %fs:40, %rax jne .L77 movl $0, %eax addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L65: .cfi_restore_state endbr64 movq %rax, %rbx leaq 48(%rsp), %rdi call _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev movq 72(%rsp), %rax subq %fs:40, %rax je .L63 call __stack_chk_fail@PLT .L63: movq %rbx, %rdi .LEHB2: call _Unwind_Resume@PLT .LEHE2: .L77: call __stack_chk_fail@PLT .cfi_endproc .LFE4034: .globl __gxx_personality_v0 .section .gcc_except_table,"a",@progbits .LLSDA4034: .byte 0xff .byte 0xff .byte 0x1 .uleb128 .LLSDACSE4034-.LLSDACSB4034 .LLSDACSB4034: .uleb128 .LEHB0-.LFB4034 .uleb128 .LEHE0-.LEHB0 .uleb128 0 .uleb128 0 .uleb128 .LEHB1-.LFB4034 .uleb128 .LEHE1-.LEHB1 .uleb128 .L65-.LFB4034 .uleb128 0 .uleb128 .LEHB2-.LFB4034 .uleb128 .LEHE2-.LEHB2 .uleb128 0 .uleb128 0 .LLSDACSE4034: .text .size main, .-main .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 _ZL2cm .comm _ZL2cm,36,32 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC4: .long 1065353216 .align 4 .LC6: .long 1056964608 .align 4 .LC8: .long 1083179008 .hidden DW.ref.__gxx_personality_v0 .weak DW.ref.__gxx_personality_v0 .section .data.rel.local.DW.ref.__gxx_personality_v0,"awG",@progbits,DW.ref.__gxx_personality_v0,comdat .align 8 .type DW.ref.__gxx_personality_v0, @object .size DW.ref.__gxx_personality_v0, 8 DW.ref.__gxx_personality_v0: .quad __gxx_personality_v0 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <cuda_runtime.h> #include <iostream> #include <vector> #include <utility> #include <stdio.h> #include <math.h> using namespace std; #define K 3 #define BLCH 8 #define BLCW 32 #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); } inline void gpuAssert(cudaError_t code, const char *file, int line, bool abort=true) { if (code != cudaSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line); if (abort) exit(code); } } // declaration of constant memory where the fiter values are stored __constant__ float cm[K*K]; __device__ void conv(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k, float* sm, int gID, int tID, int nT, int rel_row, int rel_col, int nRows, int stopPrefetchRowID, int lastActiveThreadID) { for(int i=k; i<=nRows; i++) { /* ----prefetch a pixel value from GM and store it in register---- all threads fetch the cell value immediately below to the current cell iteratively last thread in the block would fetch k cells immediately below the current cell boundary check would be needed for the blocks that act on the bottom most partition of the input image to prevent it from prefetching out of image values. */ float reg; float regArr[K]; if(i <= stopPrefetchRowID){ reg = gm[i * iw + gID]; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ regArr[j] = gm[(i * iw) + gID + j]; } } } // load k * k pixels above the current cell float imgPixels[K*K]; for(int r=i-k; r<i; r++){ for(int c=0; c<k; c++){ /* translate the indices to [0,k] using r - (i-k) as imgPixels is of size k*k */ imgPixels[(r-i+k)*k + c] = sm[r * smW + tID + c]; } } /*multiply image pixel values with filter values (direct convolution) */ float convolvedCell = 0.0; for(int c=0; c<k*k; c++){ convolvedCell += cm[c]*imgPixels[c]; } //place the convolvedCell value into convolvedMatrix int cID = ( ( (rel_row * bh) + (i-k) ) * cw )+( rel_col * nT )+tID; if(cID < 0 || cID >= ch*cw ) { printf("cID : %d, tID : %d, gID : %d\n", cID, tID, gID ); } convolved[cID] = convolvedCell; __syncthreads(); if(i <= stopPrefetchRowID){ sm[i * smW + tID] = reg; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ int sID = i *smW + tID + j; sm[sID] = regArr[j]; } } } __syncthreads(); } } __global__ void conv_kernel(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k) { int tID = threadIdx.x; int bID = blockIdx.x; int nT = blockDim.x; int nB = gridDim.x; int nBx = iw / nT; //printf("num of blocks is %d\n", nB); //printf("nB in a row is %d\n", nBx); //check for right border or bottom border thread block bool isBottomBorder = false; bool isRightBorder = false; // bottom border thread block if(bID >= nB - nBx) { //printf("bID : %d is bottom border\n", bID); isBottomBorder = true; } // right border thread block if((bID+1) % nBx == 0){ //printf("bID : %d is right border\n", bID); isRightBorder = true; } // ---------------- Load k rows from GM into SM ---------------------- __shared__ float sm[ (BLCH + K - 1) * (BLCW + K - 1) ]; // rel_row and rel_col maps the Thread Block to appropriate position int rel_row = bID / nBx; int rel_col = bID % nBx; // (rel_row * bh * iw) covers all the cells before row_ids bh, 2bh, 3bh .. // gID finally maps threads to cells at rows 0, bh, 2bh, 3bh, ... int gID = (rel_row * bh * iw) + (rel_col * nT) + tID; for(int i=0; i<k; i++){ int sID = i * smW + tID; sm[sID] = gm[i * iw + gID]; /* if last thread in the block, it should fetch additional k-1 pixels in each row which are needed for computation of the convolution */ if(!isRightBorder && tID == nT-1){ for(int j=1; j<=k-1; j++){ sID = (i * smW) + tID + j; sm[sID] = gm[i * iw + gID + j]; } } } __syncthreads(); if( !isBottomBorder && !isRightBorder ){ int lastActiveThreadID = nT - 1; int nRows = bh + k - 1; int stopPrefetchRowID = nRows; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isBottomBorder && isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh; int stopPrefetchRowID = nRows - 1; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } else if( isBottomBorder ){ int nRows = bh; int stopPrefetchRowID = nRows-1; int lastActiveThreadID = nT - 1; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh + k - 1; int stopPrefetchRowID = nRows; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } } int main(int argc, char **argv){ /* set values for image dimensions, block dimensions, filter size, stride .. some of the constraints to keep in mind are 1. the value of k(filter size) should be less than blcH and blcW 2. stride value(s) should be 1 */ int imgH = 2048; int imgW = 2048; int blcH = BLCH; int blcW = BLCW; int k = K; int s = 1; int nB = (imgH * imgW) / (blcH * blcW); int nT = blcW; int imgDims = imgH * imgW; int imgSize = imgDims * sizeof(float); // create host array that can hold pixel intensity values float *h_img = new float[imgDims]; for(int i=0; i<imgDims; i++){ h_img[i] = 1.0; } // create device array that can hold pixel intensity values in GPU GM float *d_img; gpuErrchk(cudaMalloc((void **) &d_img, imgSize )); gpuErrchk(cudaMemcpy(d_img, h_img, imgSize, cudaMemcpyHostToDevice)); // create filter and copy to constant memory int filterDims = k * k; int filterSize = filterDims * sizeof(float); float *filter = new float[filterDims]; for(int i=0; i<filterDims; i++){ filter[i] = 0.5; } gpuErrchk(cudaMemcpyToSymbol(cm, filter, filterSize)); // create host and device array that holds the convoluted matrix int convH = ( (imgH - k) / s ) + 1; int convW = convH; int convDims = convH * convW; int convSize = convDims * sizeof(float); float *h_convolved = new float[convDims]; for(int i=0; i<convDims; i++){ h_convolved[i] = 0; } float *d_convolved; gpuErrchk(cudaMalloc((void **) &d_convolved, convSize)); gpuErrchk(cudaMemcpy(d_convolved, h_convolved, convSize, cudaMemcpyHostToDevice)); // calculate shared memory dimensions int smH = blcH + k - 1; int smW = blcW + k - 1; // call the kernel conv_kernel<<<nB, nT>>>(d_img, d_convolved, blcH, blcW, imgH, imgW, convH, convW, smH, smW, k); gpuErrchk(cudaMemcpy(h_convolved, d_convolved, convSize, cudaMemcpyDeviceToHost)); vector<pair<int,int> > miss; for(int i=0; i<convH; i++){ for(int j=0; j<convW; j++){ //cout<<h_convolved[i*convW +j]<<" "; if(h_convolved[i*convW +j] != 4.5){ miss.push_back(make_pair(i,j)); } } //cout<<"\n"; } cout<<miss.size()<<"\n"; for(int i=0;i<miss.size();i++){ cout<<miss[i].first<<","<<miss[i].second<<"\n"; } cudaDeviceReset(); delete h_img; delete h_convolved; return 0; }
#include <hip/hip_runtime.h> #include <iostream> #include <vector> #include <utility> #include <stdio.h> #include <math.h> using namespace std; #define K 3 #define BLCH 8 #define BLCW 32 #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); } inline void gpuAssert(hipError_t code, const char *file, int line, bool abort=true) { if (code != hipSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", hipGetErrorString(code), file, line); if (abort) exit(code); } } // declaration of constant memory where the fiter values are stored __constant__ float cm[K*K]; __device__ void conv(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k, float* sm, int gID, int tID, int nT, int rel_row, int rel_col, int nRows, int stopPrefetchRowID, int lastActiveThreadID) { for(int i=k; i<=nRows; i++) { /* ----prefetch a pixel value from GM and store it in register---- all threads fetch the cell value immediately below to the current cell iteratively last thread in the block would fetch k cells immediately below the current cell boundary check would be needed for the blocks that act on the bottom most partition of the input image to prevent it from prefetching out of image values. */ float reg; float regArr[K]; if(i <= stopPrefetchRowID){ reg = gm[i * iw + gID]; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ regArr[j] = gm[(i * iw) + gID + j]; } } } // load k * k pixels above the current cell float imgPixels[K*K]; for(int r=i-k; r<i; r++){ for(int c=0; c<k; c++){ /* translate the indices to [0,k] using r - (i-k) as imgPixels is of size k*k */ imgPixels[(r-i+k)*k + c] = sm[r * smW + tID + c]; } } /*multiply image pixel values with filter values (direct convolution) */ float convolvedCell = 0.0; for(int c=0; c<k*k; c++){ convolvedCell += cm[c]*imgPixels[c]; } //place the convolvedCell value into convolvedMatrix int cID = ( ( (rel_row * bh) + (i-k) ) * cw )+( rel_col * nT )+tID; if(cID < 0 || cID >= ch*cw ) { printf("cID : %d, tID : %d, gID : %d\n", cID, tID, gID ); } convolved[cID] = convolvedCell; __syncthreads(); if(i <= stopPrefetchRowID){ sm[i * smW + tID] = reg; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ int sID = i *smW + tID + j; sm[sID] = regArr[j]; } } } __syncthreads(); } } __global__ void conv_kernel(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k) { int tID = threadIdx.x; int bID = blockIdx.x; int nT = blockDim.x; int nB = gridDim.x; int nBx = iw / nT; //printf("num of blocks is %d\n", nB); //printf("nB in a row is %d\n", nBx); //check for right border or bottom border thread block bool isBottomBorder = false; bool isRightBorder = false; // bottom border thread block if(bID >= nB - nBx) { //printf("bID : %d is bottom border\n", bID); isBottomBorder = true; } // right border thread block if((bID+1) % nBx == 0){ //printf("bID : %d is right border\n", bID); isRightBorder = true; } // ---------------- Load k rows from GM into SM ---------------------- __shared__ float sm[ (BLCH + K - 1) * (BLCW + K - 1) ]; // rel_row and rel_col maps the Thread Block to appropriate position int rel_row = bID / nBx; int rel_col = bID % nBx; // (rel_row * bh * iw) covers all the cells before row_ids bh, 2bh, 3bh .. // gID finally maps threads to cells at rows 0, bh, 2bh, 3bh, ... int gID = (rel_row * bh * iw) + (rel_col * nT) + tID; for(int i=0; i<k; i++){ int sID = i * smW + tID; sm[sID] = gm[i * iw + gID]; /* if last thread in the block, it should fetch additional k-1 pixels in each row which are needed for computation of the convolution */ if(!isRightBorder && tID == nT-1){ for(int j=1; j<=k-1; j++){ sID = (i * smW) + tID + j; sm[sID] = gm[i * iw + gID + j]; } } } __syncthreads(); if( !isBottomBorder && !isRightBorder ){ int lastActiveThreadID = nT - 1; int nRows = bh + k - 1; int stopPrefetchRowID = nRows; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isBottomBorder && isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh; int stopPrefetchRowID = nRows - 1; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } else if( isBottomBorder ){ int nRows = bh; int stopPrefetchRowID = nRows-1; int lastActiveThreadID = nT - 1; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh + k - 1; int stopPrefetchRowID = nRows; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } } int main(int argc, char **argv){ /* set values for image dimensions, block dimensions, filter size, stride .. some of the constraints to keep in mind are 1. the value of k(filter size) should be less than blcH and blcW 2. stride value(s) should be 1 */ int imgH = 2048; int imgW = 2048; int blcH = BLCH; int blcW = BLCW; int k = K; int s = 1; int nB = (imgH * imgW) / (blcH * blcW); int nT = blcW; int imgDims = imgH * imgW; int imgSize = imgDims * sizeof(float); // create host array that can hold pixel intensity values float *h_img = new float[imgDims]; for(int i=0; i<imgDims; i++){ h_img[i] = 1.0; } // create device array that can hold pixel intensity values in GPU GM float *d_img; gpuErrchk(hipMalloc((void **) &d_img, imgSize )); gpuErrchk(hipMemcpy(d_img, h_img, imgSize, hipMemcpyHostToDevice)); // create filter and copy to constant memory int filterDims = k * k; int filterSize = filterDims * sizeof(float); float *filter = new float[filterDims]; for(int i=0; i<filterDims; i++){ filter[i] = 0.5; } gpuErrchk(hipMemcpyToSymbol(HIP_SYMBOL(cm), filter, filterSize)); // create host and device array that holds the convoluted matrix int convH = ( (imgH - k) / s ) + 1; int convW = convH; int convDims = convH * convW; int convSize = convDims * sizeof(float); float *h_convolved = new float[convDims]; for(int i=0; i<convDims; i++){ h_convolved[i] = 0; } float *d_convolved; gpuErrchk(hipMalloc((void **) &d_convolved, convSize)); gpuErrchk(hipMemcpy(d_convolved, h_convolved, convSize, hipMemcpyHostToDevice)); // calculate shared memory dimensions int smH = blcH + k - 1; int smW = blcW + k - 1; // call the kernel conv_kernel<<<nB, nT>>>(d_img, d_convolved, blcH, blcW, imgH, imgW, convH, convW, smH, smW, k); gpuErrchk(hipMemcpy(h_convolved, d_convolved, convSize, hipMemcpyDeviceToHost)); vector<pair<int,int> > miss; for(int i=0; i<convH; i++){ for(int j=0; j<convW; j++){ //cout<<h_convolved[i*convW +j]<<" "; if(h_convolved[i*convW +j] != 4.5){ miss.push_back(make_pair(i,j)); } } //cout<<"\n"; } cout<<miss.size()<<"\n"; for(int i=0;i<miss.size();i++){ cout<<miss[i].first<<","<<miss[i].second<<"\n"; } hipDeviceReset(); delete h_img; delete h_convolved; return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <iostream> #include <vector> #include <utility> #include <stdio.h> #include <math.h> using namespace std; #define K 3 #define BLCH 8 #define BLCW 32 #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); } inline void gpuAssert(hipError_t code, const char *file, int line, bool abort=true) { if (code != hipSuccess) { fprintf(stderr,"GPUassert: %s %s %d\n", hipGetErrorString(code), file, line); if (abort) exit(code); } } // declaration of constant memory where the fiter values are stored __constant__ float cm[K*K]; __device__ void conv(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k, float* sm, int gID, int tID, int nT, int rel_row, int rel_col, int nRows, int stopPrefetchRowID, int lastActiveThreadID) { for(int i=k; i<=nRows; i++) { /* ----prefetch a pixel value from GM and store it in register---- all threads fetch the cell value immediately below to the current cell iteratively last thread in the block would fetch k cells immediately below the current cell boundary check would be needed for the blocks that act on the bottom most partition of the input image to prevent it from prefetching out of image values. */ float reg; float regArr[K]; if(i <= stopPrefetchRowID){ reg = gm[i * iw + gID]; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ regArr[j] = gm[(i * iw) + gID + j]; } } } // load k * k pixels above the current cell float imgPixels[K*K]; for(int r=i-k; r<i; r++){ for(int c=0; c<k; c++){ /* translate the indices to [0,k] using r - (i-k) as imgPixels is of size k*k */ imgPixels[(r-i+k)*k + c] = sm[r * smW + tID + c]; } } /*multiply image pixel values with filter values (direct convolution) */ float convolvedCell = 0.0; for(int c=0; c<k*k; c++){ convolvedCell += cm[c]*imgPixels[c]; } //place the convolvedCell value into convolvedMatrix int cID = ( ( (rel_row * bh) + (i-k) ) * cw )+( rel_col * nT )+tID; if(cID < 0 || cID >= ch*cw ) { printf("cID : %d, tID : %d, gID : %d\n", cID, tID, gID ); } convolved[cID] = convolvedCell; __syncthreads(); if(i <= stopPrefetchRowID){ sm[i * smW + tID] = reg; if(tID == lastActiveThreadID){ for(int j=1; j<=k-1; j++){ int sID = i *smW + tID + j; sm[sID] = regArr[j]; } } } __syncthreads(); } } __global__ void conv_kernel(const float* gm, float* convolved, int bh, int bw, int ih, int iw, int ch, int cw, int smH, int smW, int k) { int tID = threadIdx.x; int bID = blockIdx.x; int nT = blockDim.x; int nB = gridDim.x; int nBx = iw / nT; //printf("num of blocks is %d\n", nB); //printf("nB in a row is %d\n", nBx); //check for right border or bottom border thread block bool isBottomBorder = false; bool isRightBorder = false; // bottom border thread block if(bID >= nB - nBx) { //printf("bID : %d is bottom border\n", bID); isBottomBorder = true; } // right border thread block if((bID+1) % nBx == 0){ //printf("bID : %d is right border\n", bID); isRightBorder = true; } // ---------------- Load k rows from GM into SM ---------------------- __shared__ float sm[ (BLCH + K - 1) * (BLCW + K - 1) ]; // rel_row and rel_col maps the Thread Block to appropriate position int rel_row = bID / nBx; int rel_col = bID % nBx; // (rel_row * bh * iw) covers all the cells before row_ids bh, 2bh, 3bh .. // gID finally maps threads to cells at rows 0, bh, 2bh, 3bh, ... int gID = (rel_row * bh * iw) + (rel_col * nT) + tID; for(int i=0; i<k; i++){ int sID = i * smW + tID; sm[sID] = gm[i * iw + gID]; /* if last thread in the block, it should fetch additional k-1 pixels in each row which are needed for computation of the convolution */ if(!isRightBorder && tID == nT-1){ for(int j=1; j<=k-1; j++){ sID = (i * smW) + tID + j; sm[sID] = gm[i * iw + gID + j]; } } } __syncthreads(); if( !isBottomBorder && !isRightBorder ){ int lastActiveThreadID = nT - 1; int nRows = bh + k - 1; int stopPrefetchRowID = nRows; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isBottomBorder && isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh; int stopPrefetchRowID = nRows - 1; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } else if( isBottomBorder ){ int nRows = bh; int stopPrefetchRowID = nRows-1; int lastActiveThreadID = nT - 1; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } else if( isRightBorder ){ /* make the last k-1 threads in the block to be idle. as there is no convolution needed for them */ if(tID < (nT - (k-1))){ int nRows = bh + k - 1; int stopPrefetchRowID = nRows; int lastActiveThreadID = nT - k; conv( gm, convolved, bh, bw, ih, iw, ch, cw, smH, smW, k, sm, gID, tID, nT, rel_row, rel_col, nRows, stopPrefetchRowID, lastActiveThreadID ); } } } int main(int argc, char **argv){ /* set values for image dimensions, block dimensions, filter size, stride .. some of the constraints to keep in mind are 1. the value of k(filter size) should be less than blcH and blcW 2. stride value(s) should be 1 */ int imgH = 2048; int imgW = 2048; int blcH = BLCH; int blcW = BLCW; int k = K; int s = 1; int nB = (imgH * imgW) / (blcH * blcW); int nT = blcW; int imgDims = imgH * imgW; int imgSize = imgDims * sizeof(float); // create host array that can hold pixel intensity values float *h_img = new float[imgDims]; for(int i=0; i<imgDims; i++){ h_img[i] = 1.0; } // create device array that can hold pixel intensity values in GPU GM float *d_img; gpuErrchk(hipMalloc((void **) &d_img, imgSize )); gpuErrchk(hipMemcpy(d_img, h_img, imgSize, hipMemcpyHostToDevice)); // create filter and copy to constant memory int filterDims = k * k; int filterSize = filterDims * sizeof(float); float *filter = new float[filterDims]; for(int i=0; i<filterDims; i++){ filter[i] = 0.5; } gpuErrchk(hipMemcpyToSymbol(HIP_SYMBOL(cm), filter, filterSize)); // create host and device array that holds the convoluted matrix int convH = ( (imgH - k) / s ) + 1; int convW = convH; int convDims = convH * convW; int convSize = convDims * sizeof(float); float *h_convolved = new float[convDims]; for(int i=0; i<convDims; i++){ h_convolved[i] = 0; } float *d_convolved; gpuErrchk(hipMalloc((void **) &d_convolved, convSize)); gpuErrchk(hipMemcpy(d_convolved, h_convolved, convSize, hipMemcpyHostToDevice)); // calculate shared memory dimensions int smH = blcH + k - 1; int smW = blcW + k - 1; // call the kernel conv_kernel<<<nB, nT>>>(d_img, d_convolved, blcH, blcW, imgH, imgW, convH, convW, smH, smW, k); gpuErrchk(hipMemcpy(h_convolved, d_convolved, convSize, hipMemcpyDeviceToHost)); vector<pair<int,int> > miss; for(int i=0; i<convH; i++){ for(int j=0; j<convW; j++){ //cout<<h_convolved[i*convW +j]<<" "; if(h_convolved[i*convW +j] != 4.5){ miss.push_back(make_pair(i,j)); } } //cout<<"\n"; } cout<<miss.size()<<"\n"; for(int i=0;i<miss.size();i++){ cout<<miss[i].first<<","<<miss[i].second<<"\n"; } hipDeviceReset(); delete h_img; delete h_convolved; return 0; }
.text .file "convolution.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z26__device_stub__conv_kernelPKfPfiiiiiiiii # -- Begin function _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .p2align 4, 0x90 .type _Z26__device_stub__conv_kernelPKfPfiiiiiiiii,@function _Z26__device_stub__conv_kernelPKfPfiiiiiiiii: # @_Z26__device_stub__conv_kernelPKfPfiiiiiiiii .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 176(%rsp), %rax movq %rax, 128(%rsp) leaq 184(%rsp), %rax movq %rax, 136(%rsp) leaq 192(%rsp), %rax movq %rax, 144(%rsp) leaq 200(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 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 $_Z11conv_kernelPKfPfiiiiiiiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z26__device_stub__conv_kernelPKfPfiiiiiiiii, .Lfunc_end0-_Z26__device_stub__conv_kernelPKfPfiiiiiiiii .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function main .LCPI1_0: .long 0x40900000 # float 4.5 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .Lfunc_begin0: .cfi_startproc .cfi_personality 3, __gxx_personality_v0 .cfi_lsda 3, .Lexception0 # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $248, %rsp .cfi_def_cfa_offset 304 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 .cfi_escape 0x2e, 0x00 movl $16777216, %edi # imm = 0x1000000 callq _Znam movq %rax, %r15 xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%r15,%rax,4) # imm = 0x3F800000 incq %rax cmpq $4194304, %rax # imm = 0x400000 jne .LBB1_1 # %bb.2: .cfi_escape 0x2e, 0x00 leaq 64(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc testl %eax, %eax jne .LBB1_3 # %bb.5: # %_Z9gpuAssert10hipError_tPKcib.exit movq 64(%rsp), %rdi .cfi_escape 0x2e, 0x00 movl $16777216, %edx # imm = 0x1000000 movq %r15, %rsi movl $1, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_6 # %bb.9: # %_Z9gpuAssert10hipError_tPKcib.exit85 .cfi_escape 0x2e, 0x00 movl $36, %edi callq _Znam xorl %ecx, %ecx .p2align 4, 0x90 .LBB1_10: # =>This Inner Loop Header: Depth=1 movl $1056964608, (%rax,%rcx,4) # imm = 0x3F000000 incq %rcx cmpq $9, %rcx jne .LBB1_10 # %bb.7: .cfi_escape 0x2e, 0x00 movl $cm, %edi movl $36, %edx movq %rax, %rsi xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol testl %eax, %eax jne .LBB1_8 # %bb.11: # %_Z9gpuAssert10hipError_tPKcib.exit87 .cfi_escape 0x2e, 0x00 movl $16744464, %edi # imm = 0xFF8010 callq _Znam movq %rax, %rbx .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rax, %rdi xorl %esi, %esi callq memset@PLT .cfi_escape 0x2e, 0x00 leaq 8(%rsp), %rdi movl $16744464, %esi # imm = 0xFF8010 callq hipMalloc testl %eax, %eax jne .LBB1_12 # %bb.13: # %_Z9gpuAssert10hipError_tPKcib.exit89 movq 8(%rsp), %rdi .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_14 # %bb.15: # %_Z9gpuAssert10hipError_tPKcib.exit91 movabsq $4294967328, %rdx # imm = 0x100000020 leaq 16352(%rdx), %rdi .cfi_escape 0x2e, 0x00 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_17 # %bb.16: movq 64(%rsp), %rax movq 8(%rsp), %rcx movq %rax, 152(%rsp) movq %rcx, 144(%rsp) movl $8, 52(%rsp) movl $32, 48(%rsp) movl $2048, 44(%rsp) # imm = 0x800 movl $2048, 40(%rsp) # imm = 0x800 movl $2046, 36(%rsp) # imm = 0x7FE movl $2046, 32(%rsp) # imm = 0x7FE movl $10, 28(%rsp) movl $34, 24(%rsp) movl $3, 20(%rsp) leaq 152(%rsp), %rax movq %rax, 160(%rsp) leaq 144(%rsp), %rax movq %rax, 168(%rsp) leaq 52(%rsp), %rax movq %rax, 176(%rsp) leaq 48(%rsp), %rax movq %rax, 184(%rsp) leaq 44(%rsp), %rax movq %rax, 192(%rsp) leaq 40(%rsp), %rax movq %rax, 200(%rsp) leaq 36(%rsp), %rax movq %rax, 208(%rsp) leaq 32(%rsp), %rax movq %rax, 216(%rsp) leaq 28(%rsp), %rax movq %rax, 224(%rsp) leaq 24(%rsp), %rax movq %rax, 232(%rsp) leaq 20(%rsp), %rax movq %rax, 240(%rsp) .cfi_escape 0x2e, 0x00 leaq 128(%rsp), %rdi leaq 112(%rsp), %rsi leaq 104(%rsp), %rdx leaq 96(%rsp), %rcx callq __hipPopCallConfiguration movq 128(%rsp), %rsi movl 136(%rsp), %edx movq 112(%rsp), %rcx movl 120(%rsp), %r8d .cfi_escape 0x2e, 0x10 leaq 160(%rsp), %r9 movl $_Z11conv_kernelPKfPfiiiiiiiii, %edi pushq 96(%rsp) .cfi_adjust_cfa_offset 8 pushq 112(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_17: movq 8(%rsp), %rsi .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_59 # %bb.18: # %.preheader.preheader xorl %ecx, %ecx movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero xorl %eax, %eax movq %rax, (%rsp) # 8-byte Spill xorl %r12d, %r12d xorl %eax, %eax movq %rbx, 56(%rsp) # 8-byte Spill movq %r15, 72(%rsp) # 8-byte Spill jmp .LBB1_19 .p2align 4, 0x90 .LBB1_24: # %_Z9gpuAssert10hipError_tPKcib.exit93 # in Loop: Header=BB1_19 Depth=1 incq %rcx cmpq $2046, %rcx # imm = 0x7FE je .LBB1_25 .LBB1_19: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_20 Depth 2 # Child Loop BB1_47 Depth 3 imulq $8184, %rcx, %rdx # imm = 0x1FF8 addq %rbx, %rdx xorl %r14d, %r14d movq %rcx, 88(%rsp) # 8-byte Spill movq %rdx, 80(%rsp) # 8-byte Spill jmp .LBB1_20 .p2align 4, 0x90 .LBB1_22: # in Loop: Header=BB1_20 Depth=2 movq %rbp, (%r12) addq $8, %r12 .LBB1_23: # %_ZNSt6vectorISt4pairIiiESaIS1_EE9push_backEOS1_.exit # in Loop: Header=BB1_20 Depth=2 incq %r14 cmpq $2046, %r14 # imm = 0x7FE je .LBB1_24 .LBB1_20: # Parent Loop BB1_19 Depth=1 # => This Loop Header: Depth=2 # Child Loop BB1_47 Depth 3 movss (%rdx,%r14,4), %xmm0 # xmm0 = mem[0],zero,zero,zero ucomiss %xmm1, %xmm0 jne .LBB1_21 jnp .LBB1_23 .LBB1_21: # in Loop: Header=BB1_20 Depth=2 movq %r14, %rbp shlq $32, %rbp orq %rcx, %rbp cmpq %rax, %r12 jne .LBB1_22 # %bb.38: # in Loop: Header=BB1_20 Depth=2 movq %r12, %r13 subq (%rsp), %r13 # 8-byte Folded Reload movabsq $9223372036854775800, %rax # imm = 0x7FFFFFFFFFFFFFF8 cmpq %rax, %r13 je .LBB1_39 # %bb.41: # %_ZNKSt6vectorISt4pairIiiESaIS1_EE12_M_check_lenEmPKc.exit.i # in Loop: Header=BB1_20 Depth=2 sarq $3, %r13 cmpq $1, %r13 movq %r13, %rax adcq $0, %rax leaq (%rax,%r13), %r15 movabsq $1152921504606846975, %rcx # imm = 0xFFFFFFFFFFFFFFF cmpq %rcx, %r15 cmovaeq %rcx, %r15 addq %r13, %rax cmovbq %rcx, %r15 testq %r15, %r15 je .LBB1_42 # %bb.43: # in Loop: Header=BB1_20 Depth=2 leaq (,%r15,8), %rdi .Ltmp0: .cfi_escape 0x2e, 0x00 callq _Znwm .Ltmp1: # %bb.44: # in Loop: Header=BB1_20 Depth=2 movq %rax, %rbx movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero jmp .LBB1_45 .LBB1_42: # in Loop: Header=BB1_20 Depth=2 xorl %ebx, %ebx .LBB1_45: # %_ZNSt12_Vector_baseISt4pairIiiESaIS1_EE11_M_allocateEm.exit.i # in Loop: Header=BB1_20 Depth=2 movq %rbp, (%rbx,%r13,8) movq %rbx, %r13 cmpq %r12, (%rsp) # 8-byte Folded Reload je .LBB1_48 # %bb.46: # %.lr.ph.i.i.i.i.preheader # in Loop: Header=BB1_20 Depth=2 movq %rbx, %r13 movq (%rsp), %rax # 8-byte Reload .p2align 4, 0x90 .LBB1_47: # %.lr.ph.i.i.i.i # Parent Loop BB1_19 Depth=1 # Parent Loop BB1_20 Depth=2 # => This Inner Loop Header: Depth=3 movq (%rax), %rcx movq %rcx, (%r13) addq $8, %rax addq $8, %r13 cmpq %r12, %rax jne .LBB1_47 .LBB1_48: # %_ZNSt6vectorISt4pairIiiESaIS1_EE11_S_relocateEPS1_S4_S4_RS2_.exit.i # in Loop: Header=BB1_20 Depth=2 movq (%rsp), %rdi # 8-byte Reload testq %rdi, %rdi je .LBB1_50 # %bb.49: # in Loop: Header=BB1_20 Depth=2 .cfi_escape 0x2e, 0x00 callq _ZdlPv movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero .LBB1_50: # %_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_.exit # in Loop: Header=BB1_20 Depth=2 addq $8, %r13 leaq (%rbx,%r15,8), %rax movq %r13, %r12 movq %rbx, (%rsp) # 8-byte Spill movq 72(%rsp), %r15 # 8-byte Reload movq 56(%rsp), %rbx # 8-byte Reload movq 88(%rsp), %rcx # 8-byte Reload movq 80(%rsp), %rdx # 8-byte Reload jmp .LBB1_23 .LBB1_25: movq %r12, %r14 subq (%rsp), %r14 # 8-byte Folded Reload sarq $3, %r14 .Ltmp3: .cfi_escape 0x2e, 0x00 movl $_ZSt4cout, %edi movq %r14, %rsi callq _ZNSo9_M_insertImEERSoT_ .Ltmp4: # %bb.26: # %_ZNSolsEm.exit .Ltmp5: .cfi_escape 0x2e, 0x00 movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp6: # %bb.27: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit.preheader movq (%rsp), %rax # 8-byte Reload cmpq %rax, %r12 je .LBB1_34 # %bb.28: # %.lr.ph.preheader movq %rax, %r12 cmpq $1, %r14 adcq $0, %r14 xorl %ebx, %ebx .p2align 4, 0x90 .LBB1_29: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl (%r12,%rbx,8), %esi .Ltmp7: .cfi_escape 0x2e, 0x00 movl $_ZSt4cout, %edi callq _ZNSolsEi .Ltmp8: # %bb.30: # in Loop: Header=BB1_29 Depth=1 .Ltmp9: movq %rax, %r13 .cfi_escape 0x2e, 0x00 movl $.L.str.2, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp10: # %bb.31: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit96 # in Loop: Header=BB1_29 Depth=1 movl 4(%r12,%rbx,8), %esi .Ltmp11: .cfi_escape 0x2e, 0x00 movq %r13, %rdi callq _ZNSolsEi .Ltmp12: # %bb.32: # in Loop: Header=BB1_29 Depth=1 .Ltmp13: .cfi_escape 0x2e, 0x00 movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp14: # %bb.33: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit98 # in Loop: Header=BB1_29 Depth=1 incq %rbx cmpq %rbx, %r14 jne .LBB1_29 .LBB1_34: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit._crit_edge .Ltmp16: .cfi_escape 0x2e, 0x00 callq hipDeviceReset .Ltmp17: # %bb.35: .cfi_escape 0x2e, 0x00 movq %r15, %rdi callq _ZdlPv .cfi_escape 0x2e, 0x00 movq 56(%rsp), %rdi # 8-byte Reload callq _ZdlPv movq (%rsp), %rdi # 8-byte Reload testq %rdi, %rdi je .LBB1_37 # %bb.36: .cfi_escape 0x2e, 0x00 callq _ZdlPv .LBB1_37: # %_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev.exit xorl %eax, %eax addq $248, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB1_39: .cfi_def_cfa_offset 304 .Ltmp19: .cfi_escape 0x2e, 0x00 movl $.L.str.4, %edi callq _ZSt20__throw_length_errorPKc .Ltmp20: # %bb.40: # %.noexc102 .LBB1_3: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $233, %r8d jmp .LBB1_4 .LBB1_6: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $234, %r8d jmp .LBB1_4 .LBB1_8: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $242, %r8d jmp .LBB1_4 .LBB1_12: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $253, %r8d jmp .LBB1_4 .LBB1_14: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $255, %r8d jmp .LBB1_4 .LBB1_59: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $267, %r8d # imm = 0x10B .LBB1_4: xorl %eax, %eax callq fprintf .cfi_escape 0x2e, 0x00 movl %ebp, %edi callq exit .LBB1_54: .Ltmp18: jmp .LBB1_52 .LBB1_55: .Ltmp15: movq %rax, %rbx jmp .LBB1_56 .LBB1_51: # %.loopexit .Ltmp2: jmp .LBB1_52 .LBB1_53: # %.loopexit.split-lp .Ltmp21: .LBB1_52: movq %rax, %rbx movq (%rsp), %r12 # 8-byte Reload .LBB1_56: testq %r12, %r12 je .LBB1_58 # %bb.57: .cfi_escape 0x2e, 0x00 movq %r12, %rdi callq _ZdlPv .LBB1_58: # %_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev.exit100 .cfi_escape 0x2e, 0x00 movq %rbx, %rdi callq _Unwind_Resume@PLT .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc .section .gcc_except_table,"a",@progbits .p2align 2, 0x0 GCC_except_table1: .Lexception0: .byte 255 # @LPStart Encoding = omit .byte 255 # @TType Encoding = omit .byte 1 # Call site Encoding = uleb128 .uleb128 .Lcst_end0-.Lcst_begin0 .Lcst_begin0: .uleb128 .Lfunc_begin0-.Lfunc_begin0 # >> Call Site 1 << .uleb128 .Ltmp0-.Lfunc_begin0 # Call between .Lfunc_begin0 and .Ltmp0 .byte 0 # has no landing pad .byte 0 # On action: cleanup .uleb128 .Ltmp0-.Lfunc_begin0 # >> Call Site 2 << .uleb128 .Ltmp1-.Ltmp0 # Call between .Ltmp0 and .Ltmp1 .uleb128 .Ltmp2-.Lfunc_begin0 # jumps to .Ltmp2 .byte 0 # On action: cleanup .uleb128 .Ltmp3-.Lfunc_begin0 # >> Call Site 3 << .uleb128 .Ltmp6-.Ltmp3 # Call between .Ltmp3 and .Ltmp6 .uleb128 .Ltmp18-.Lfunc_begin0 # jumps to .Ltmp18 .byte 0 # On action: cleanup .uleb128 .Ltmp7-.Lfunc_begin0 # >> Call Site 4 << .uleb128 .Ltmp14-.Ltmp7 # Call between .Ltmp7 and .Ltmp14 .uleb128 .Ltmp15-.Lfunc_begin0 # jumps to .Ltmp15 .byte 0 # On action: cleanup .uleb128 .Ltmp16-.Lfunc_begin0 # >> Call Site 5 << .uleb128 .Ltmp17-.Ltmp16 # Call between .Ltmp16 and .Ltmp17 .uleb128 .Ltmp18-.Lfunc_begin0 # jumps to .Ltmp18 .byte 0 # On action: cleanup .uleb128 .Ltmp19-.Lfunc_begin0 # >> Call Site 6 << .uleb128 .Ltmp20-.Ltmp19 # Call between .Ltmp19 and .Ltmp20 .uleb128 .Ltmp21-.Lfunc_begin0 # jumps to .Ltmp21 .byte 0 # On action: cleanup .uleb128 .Ltmp20-.Lfunc_begin0 # >> Call Site 7 << .uleb128 .Lfunc_end1-.Ltmp20 # Call between .Ltmp20 and .Lfunc_end1 .byte 0 # has no landing pad .byte 0 # On action: cleanup .Lcst_end0: .p2align 2, 0x0 # -- End function .text .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: 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 $_Z11conv_kernelPKfPfiiiiiiiii, %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 $1, (%rsp) movl $cm, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $36, %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 cm,@object # @cm .local cm .comm cm,36,16 .type _Z11conv_kernelPKfPfiiiiiiiii,@object # @_Z11conv_kernelPKfPfiiiiiiiii .section .rodata,"a",@progbits .globl _Z11conv_kernelPKfPfiiiiiiiii .p2align 3, 0x0 _Z11conv_kernelPKfPfiiiiiiiii: .quad _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .size _Z11conv_kernelPKfPfiiiiiiiii, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/yottabytt/convolution_kernel/master/convolution.hip" .size .L.str, 109 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "\n" .size .L.str.1, 2 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "," .size .L.str.2, 2 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "GPUassert: %s %s %d\n" .size .L.str.3, 21 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "vector::_M_realloc_insert" .size .L.str.4, 26 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z11conv_kernelPKfPfiiiiiiiii" .size .L__unnamed_1, 30 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "cm" .size .L__unnamed_2, 3 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .addrsig_sym __gxx_personality_v0 .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Unwind_Resume .addrsig_sym cm .addrsig_sym _Z11conv_kernelPKfPfiiiiiiiii .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_000691b4_00000000-6_convolution.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB4057: .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 .LFE4057: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata._Z9gpuAssert9cudaErrorPKcib.str1.1,"aMS",@progbits,1 .LC0: .string "GPUassert: %s %s %d\n" .section .text._Z9gpuAssert9cudaErrorPKcib,"axG",@progbits,_Z9gpuAssert9cudaErrorPKcib,comdat .weak _Z9gpuAssert9cudaErrorPKcib .type _Z9gpuAssert9cudaErrorPKcib, @function _Z9gpuAssert9cudaErrorPKcib: .LFB4032: .cfi_startproc endbr64 testl %edi, %edi jne .L9 ret .L9: 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 movl %edi, %ebx movq %rsi, %r13 movl %edx, %r12d movl %ecx, %ebp call cudaGetErrorString@PLT movq %rax, %rcx movl %r12d, %r9d movq %r13, %r8 leaq .LC0(%rip), %rdx movl $2, %esi movq stderr(%rip), %rdi movl $0, %eax call __fprintf_chk@PLT testb %bpl, %bpl jne .L10 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 .L10: .cfi_restore_state movl %ebx, %edi call exit@PLT .cfi_endproc .LFE4032: .size _Z9gpuAssert9cudaErrorPKcib, .-_Z9gpuAssert9cudaErrorPKcib .text .globl _Z4convPKfPfiiiiiiiiiS1_iiiiiiii .type _Z4convPKfPfiiiiiiiiiS1_iiiiiiii, @function _Z4convPKfPfiiiiiiiiiS1_iiiiiiii: .LFB4033: .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 .LFE4033: .size _Z4convPKfPfiiiiiiiiiS1_iiiiiiii, .-_Z4convPKfPfiiiiiiiiiS1_iiiiiiii .globl _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .type _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii, @function _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii: .LFB4079: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 12(%rsp), %rax movq %rax, 112(%rsp) leaq 8(%rsp), %rax movq %rax, 120(%rsp) leaq 4(%rsp), %rax movq %rax, 128(%rsp) movq %rsp, %rax movq %rax, 136(%rsp) leaq 208(%rsp), %rax movq %rax, 144(%rsp) leaq 216(%rsp), %rax movq %rax, 152(%rsp) leaq 224(%rsp), %rax movq %rax, 160(%rsp) leaq 232(%rsp), %rax movq %rax, 168(%rsp) leaq 240(%rsp), %rax movq %rax, 176(%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 .L17 .L13: movq 184(%rsp), %rax subq %fs:40, %rax jne .L18 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L17: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 216 pushq 40(%rsp) .cfi_def_cfa_offset 224 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z11conv_kernelPKfPfiiiiiiiii(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L13 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE4079: .size _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii, .-_Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .globl _Z11conv_kernelPKfPfiiiiiiiii .type _Z11conv_kernelPKfPfiiiiiiiii, @function _Z11conv_kernelPKfPfiiiiiiiii: .LFB4080: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 40 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 48 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 56 movl 56(%rsp), %eax pushq %rax .cfi_def_cfa_offset 64 call _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii addq $56, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE4080: .size _Z11conv_kernelPKfPfiiiiiiiii, .-_Z11conv_kernelPKfPfiiiiiiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC1: .string "_Z11conv_kernelPKfPfiiiiiiiii" .LC2: .string "cm" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB4082: .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 .LC1(%rip), %rdx movq %rdx, %rcx leaq _Z11conv_kernelPKfPfiiiiiiiii(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $36, %r9d movl $0, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _ZL2cm(%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 .LFE4082: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .text._ZNSt6vectorISt4pairIiiESaIS1_EED2Ev,"axG",@progbits,_ZNSt6vectorISt4pairIiiESaIS1_EED5Ev,comdat .align 2 .weak _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .type _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev, @function _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev: .LFB4393: .cfi_startproc endbr64 movq (%rdi), %rax testq %rax, %rax je .L26 subq $8, %rsp .cfi_def_cfa_offset 16 movq 16(%rdi), %rsi subq %rax, %rsi movq %rax, %rdi call _ZdlPvm@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .L26: ret .cfi_endproc .LFE4393: .size _ZNSt6vectorISt4pairIiiESaIS1_EED2Ev, .-_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .weak _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev .set _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev,_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev .section .rodata._ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_.str1.1,"aMS",@progbits,1 .LC3: .string "vector::_M_realloc_insert" .section .text._ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_,"axG",@progbits,_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_,comdat .align 2 .weak _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .type _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_, @function _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_: .LFB4652: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $40, %rsp .cfi_def_cfa_offset 96 movq %rdx, 8(%rsp) movq 8(%rdi), %rbp movq (%rdi), %r12 movq %rbp, %rax subq %r12, %rax sarq $3, %rax movabsq $1152921504606846975, %rdx cmpq %rdx, %rax je .L47 movq %rdi, %r14 movq %rsi, %r13 movq %rsi, %rbx cmpq %r12, %rbp movl $1, %edx cmovne %rax, %rdx addq %rdx, %rax jc .L32 movabsq $1152921504606846975, %rdx cmpq %rdx, %rax cmovbe %rax, %rdx movq %rdx, 16(%rsp) movq %rsi, %rdi subq %r12, %rdi movq %rdi, 24(%rsp) movl $0, %r15d testq %rax, %rax je .L33 jmp .L39 .L47: leaq .LC3(%rip), %rdi call _ZSt20__throw_length_errorPKc@PLT .L42: movq %r15, %rsi jmp .L34 .L32: movq %rsi, %rax subq %r12, %rax movq %rax, 24(%rsp) movabsq $1152921504606846975, %rax movq %rax, 16(%rsp) .L39: movq 16(%rsp), %rax leaq 0(,%rax,8), %rdi call _Znwm@PLT movq %rax, %r15 .L33: movq 8(%rsp), %rax movq (%rax), %rax movq 24(%rsp), %rdi movq %rax, (%r15,%rdi) cmpq %r12, %r13 je .L42 movq %r13, %rsi subq %r12, %rsi addq %r15, %rsi movq %r15, %rax movq %r12, %rdx .L35: movq (%rdx), %rcx movq %rcx, (%rax) addq $8, %rdx addq $8, %rax cmpq %rax, %rsi jne .L35 .L34: leaq 8(%rsi), %rax movq %rax, 8(%rsp) cmpq %rbp, %r13 je .L36 movq %rbp, %rcx subq %r13, %rcx .L37: movq (%rbx), %rdx movq %rdx, (%rax) addq $8, %rbx addq $8, %rax cmpq %rbp, %rbx jne .L37 addq %rcx, 8(%rsp) .L36: testq %r12, %r12 je .L38 movq 16(%r14), %rsi subq %r12, %rsi movq %r12, %rdi call _ZdlPvm@PLT .L38: movq %r15, (%r14) movq 8(%rsp), %rax movq %rax, 8(%r14) movq 16(%rsp), %rax leaq (%r15,%rax,8), %rax movq %rax, 16(%r14) addq $40, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE4652: .size _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_, .-_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC5: .string "/home/ubuntu/Datasets/stackv2/train-structured/yottabytt/convolution_kernel/master/convolution.cu" .section .rodata.str1.1 .LC9: .string "\n" .LC10: .string "," .text .globl main .type main, @function main: .LFB4034: .cfi_startproc .cfi_personality 0x9b,DW.ref.__gxx_personality_v0 .cfi_lsda 0x1b,.LLSDA4034 endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $88, %rsp .cfi_def_cfa_offset 144 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax movl $16777216, %edi .LEHB0: call _Znam@PLT movq %rax, %rbp leaq 16777216(%rax), %rdx movss .LC4(%rip), %xmm0 .L49: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L49 leaq 16(%rsp), %rdi movl $16777216, %esi call cudaMalloc@PLT movl %eax, %edi movl $1, %ecx movl $233, %edx leaq .LC5(%rip), %rbx movq %rbx, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $1, %ecx movl $16777216, %edx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $234, %edx movq %rbx, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $36, %edi call _Znam@PLT movq %rax, %rsi leaq 36(%rax), %rdx movss .LC6(%rip), %xmm0 .L50: movss %xmm0, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L50 movl $1, %r8d movl $0, %ecx movl $36, %edx leaq _ZL2cm(%rip), %rdi call cudaMemcpyToSymbol@PLT movl %eax, %edi movl $1, %ecx movl $242, %edx leaq .LC5(%rip), %rsi call _Z9gpuAssert9cudaErrorPKcib movl $16744464, %edi call _Znam@PLT movq %rax, %rbx movq %rax, %r13 leaq 16744464(%rax), %rdx .L51: movl $0x00000000, (%rax) addq $4, %rax cmpq %rdx, %rax jne .L51 leaq 24(%rsp), %rdi movl $16744464, %esi call cudaMalloc@PLT movl %eax, %edi movl $1, %ecx movl $253, %edx leaq .LC5(%rip), %r12 movq %r12, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $1, %ecx movl $16744464, %edx movq %rbx, %rsi movq 24(%rsp), %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $254, %edx movq %r12, %rsi call _Z9gpuAssert9cudaErrorPKcib movl $32, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $16384, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $0, %r9d movl $0, %r8d movq 48(%rsp), %rdx movl $1, %ecx movq 36(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L74 .L52: movl $2, %ecx movl $16744464, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movl %eax, %edi movl $1, %ecx movl $266, %edx leaq .LC5(%rip), %rsi call _Z9gpuAssert9cudaErrorPKcib movq $0, 48(%rsp) movq $0, 56(%rsp) movq $0, 64(%rsp) movl $0, %r14d leaq 36(%rsp), %r15 jmp .L53 .L74: subq $8, %rsp .cfi_def_cfa_offset 152 pushq $3 .cfi_def_cfa_offset 160 pushq $34 .cfi_def_cfa_offset 168 pushq $10 .cfi_def_cfa_offset 176 pushq $2046 .cfi_def_cfa_offset 184 pushq $2046 .cfi_def_cfa_offset 192 movl $2048, %r9d movl $2048, %r8d movl $32, %ecx movl $8, %edx movq 72(%rsp), %rsi movq 64(%rsp), %rdi call _Z43__device_stub__Z11conv_kernelPKfPfiiiiiiiiiPKfPfiiiiiiiii .LEHE0: addq $48, %rsp .cfi_def_cfa_offset 144 jmp .L52 .L56: leaq 48(%rsp), %rdi movq %r15, %rdx .LEHB1: call _ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_ .L54: addq $1, %r12 cmpq $2046, %r12 je .L75 .L58: movss 0(%r13,%r12,4), %xmm0 ucomiss .LC8(%rip), %xmm0 jp .L66 je .L54 .L66: movl %r14d, 36(%rsp) movl %r12d, 40(%rsp) movq 56(%rsp), %rsi cmpq 64(%rsp), %rsi je .L56 movq 36(%rsp), %rax movq %rax, (%rsi) addq $8, %rsi movq %rsi, 56(%rsp) jmp .L54 .L75: addl $1, %r14d addq $8184, %r13 cmpl $2046, %r14d je .L59 .L53: movl $0, %r12d jmp .L58 .L59: movq 56(%rsp), %r12 movq 48(%rsp), %r14 movq %r12, %r13 subq %r14, %r13 sarq $3, %r13 movq %r13, %rsi leaq _ZSt4cout(%rip), %rdi call _ZNSo9_M_insertImEERSoT_@PLT movq %rax, %rdi leaq .LC9(%rip), %rsi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT cmpq %r14, %r12 je .L60 movl $0, %r12d leaq _ZSt4cout(%rip), %r15 jmp .L61 .L76: movl $1, %edx leaq .LC10(%rip), %rsi movq %rax, 8(%rsp) movq %rax, %rdi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT movl 4(%r14,%r12,8), %esi movq 8(%rsp), %rdi call _ZNSolsEi@PLT movq %rax, %rdi movl $1, %edx leaq .LC9(%rip), %rsi call _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l@PLT addq $1, %r12 cmpq %r13, %r12 jnb .L60 .L61: movl (%r14,%r12,8), %esi movq %r15, %rdi call _ZNSolsEi@PLT jmp .L76 .L60: call cudaDeviceReset@PLT .LEHE1: movl $4, %esi movq %rbp, %rdi call _ZdlPvm@PLT movl $4, %esi movq %rbx, %rdi call _ZdlPvm@PLT leaq 48(%rsp), %rdi call _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev movq 72(%rsp), %rax subq %fs:40, %rax jne .L77 movl $0, %eax addq $88, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L65: .cfi_restore_state endbr64 movq %rax, %rbx leaq 48(%rsp), %rdi call _ZNSt6vectorISt4pairIiiESaIS1_EED1Ev movq 72(%rsp), %rax subq %fs:40, %rax je .L63 call __stack_chk_fail@PLT .L63: movq %rbx, %rdi .LEHB2: call _Unwind_Resume@PLT .LEHE2: .L77: call __stack_chk_fail@PLT .cfi_endproc .LFE4034: .globl __gxx_personality_v0 .section .gcc_except_table,"a",@progbits .LLSDA4034: .byte 0xff .byte 0xff .byte 0x1 .uleb128 .LLSDACSE4034-.LLSDACSB4034 .LLSDACSB4034: .uleb128 .LEHB0-.LFB4034 .uleb128 .LEHE0-.LEHB0 .uleb128 0 .uleb128 0 .uleb128 .LEHB1-.LFB4034 .uleb128 .LEHE1-.LEHB1 .uleb128 .L65-.LFB4034 .uleb128 0 .uleb128 .LEHB2-.LFB4034 .uleb128 .LEHE2-.LEHB2 .uleb128 0 .uleb128 0 .LLSDACSE4034: .text .size main, .-main .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 _ZL2cm .comm _ZL2cm,36,32 .section .rodata.cst4,"aM",@progbits,4 .align 4 .LC4: .long 1065353216 .align 4 .LC6: .long 1056964608 .align 4 .LC8: .long 1083179008 .hidden DW.ref.__gxx_personality_v0 .weak DW.ref.__gxx_personality_v0 .section .data.rel.local.DW.ref.__gxx_personality_v0,"awG",@progbits,DW.ref.__gxx_personality_v0,comdat .align 8 .type DW.ref.__gxx_personality_v0, @object .size DW.ref.__gxx_personality_v0, 8 DW.ref.__gxx_personality_v0: .quad __gxx_personality_v0 .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 "convolution.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z26__device_stub__conv_kernelPKfPfiiiiiiiii # -- Begin function _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .p2align 4, 0x90 .type _Z26__device_stub__conv_kernelPKfPfiiiiiiiii,@function _Z26__device_stub__conv_kernelPKfPfiiiiiiiii: # @_Z26__device_stub__conv_kernelPKfPfiiiiiiiii .cfi_startproc # %bb.0: subq $168, %rsp .cfi_def_cfa_offset 176 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movl %edx, 12(%rsp) movl %ecx, 8(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 12(%rsp), %rax movq %rax, 96(%rsp) leaq 8(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 176(%rsp), %rax movq %rax, 128(%rsp) leaq 184(%rsp), %rax movq %rax, 136(%rsp) leaq 192(%rsp), %rax movq %rax, 144(%rsp) leaq 200(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 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 $_Z11conv_kernelPKfPfiiiiiiiii, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $184, %rsp .cfi_adjust_cfa_offset -184 retq .Lfunc_end0: .size _Z26__device_stub__conv_kernelPKfPfiiiiiiiii, .Lfunc_end0-_Z26__device_stub__conv_kernelPKfPfiiiiiiiii .cfi_endproc # -- End function .section .rodata.cst4,"aM",@progbits,4 .p2align 2, 0x0 # -- Begin function main .LCPI1_0: .long 0x40900000 # float 4.5 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .Lfunc_begin0: .cfi_startproc .cfi_personality 3, __gxx_personality_v0 .cfi_lsda 3, .Lexception0 # %bb.0: pushq %rbp .cfi_def_cfa_offset 16 pushq %r15 .cfi_def_cfa_offset 24 pushq %r14 .cfi_def_cfa_offset 32 pushq %r13 .cfi_def_cfa_offset 40 pushq %r12 .cfi_def_cfa_offset 48 pushq %rbx .cfi_def_cfa_offset 56 subq $248, %rsp .cfi_def_cfa_offset 304 .cfi_offset %rbx, -56 .cfi_offset %r12, -48 .cfi_offset %r13, -40 .cfi_offset %r14, -32 .cfi_offset %r15, -24 .cfi_offset %rbp, -16 .cfi_escape 0x2e, 0x00 movl $16777216, %edi # imm = 0x1000000 callq _Znam movq %rax, %r15 xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl $1065353216, (%r15,%rax,4) # imm = 0x3F800000 incq %rax cmpq $4194304, %rax # imm = 0x400000 jne .LBB1_1 # %bb.2: .cfi_escape 0x2e, 0x00 leaq 64(%rsp), %rdi movl $16777216, %esi # imm = 0x1000000 callq hipMalloc testl %eax, %eax jne .LBB1_3 # %bb.5: # %_Z9gpuAssert10hipError_tPKcib.exit movq 64(%rsp), %rdi .cfi_escape 0x2e, 0x00 movl $16777216, %edx # imm = 0x1000000 movq %r15, %rsi movl $1, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_6 # %bb.9: # %_Z9gpuAssert10hipError_tPKcib.exit85 .cfi_escape 0x2e, 0x00 movl $36, %edi callq _Znam xorl %ecx, %ecx .p2align 4, 0x90 .LBB1_10: # =>This Inner Loop Header: Depth=1 movl $1056964608, (%rax,%rcx,4) # imm = 0x3F000000 incq %rcx cmpq $9, %rcx jne .LBB1_10 # %bb.7: .cfi_escape 0x2e, 0x00 movl $cm, %edi movl $36, %edx movq %rax, %rsi xorl %ecx, %ecx movl $1, %r8d callq hipMemcpyToSymbol testl %eax, %eax jne .LBB1_8 # %bb.11: # %_Z9gpuAssert10hipError_tPKcib.exit87 .cfi_escape 0x2e, 0x00 movl $16744464, %edi # imm = 0xFF8010 callq _Znam movq %rax, %rbx .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rax, %rdi xorl %esi, %esi callq memset@PLT .cfi_escape 0x2e, 0x00 leaq 8(%rsp), %rdi movl $16744464, %esi # imm = 0xFF8010 callq hipMalloc testl %eax, %eax jne .LBB1_12 # %bb.13: # %_Z9gpuAssert10hipError_tPKcib.exit89 movq 8(%rsp), %rdi .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rbx, %rsi movl $1, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_14 # %bb.15: # %_Z9gpuAssert10hipError_tPKcib.exit91 movabsq $4294967328, %rdx # imm = 0x100000020 leaq 16352(%rdx), %rdi .cfi_escape 0x2e, 0x00 movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_17 # %bb.16: movq 64(%rsp), %rax movq 8(%rsp), %rcx movq %rax, 152(%rsp) movq %rcx, 144(%rsp) movl $8, 52(%rsp) movl $32, 48(%rsp) movl $2048, 44(%rsp) # imm = 0x800 movl $2048, 40(%rsp) # imm = 0x800 movl $2046, 36(%rsp) # imm = 0x7FE movl $2046, 32(%rsp) # imm = 0x7FE movl $10, 28(%rsp) movl $34, 24(%rsp) movl $3, 20(%rsp) leaq 152(%rsp), %rax movq %rax, 160(%rsp) leaq 144(%rsp), %rax movq %rax, 168(%rsp) leaq 52(%rsp), %rax movq %rax, 176(%rsp) leaq 48(%rsp), %rax movq %rax, 184(%rsp) leaq 44(%rsp), %rax movq %rax, 192(%rsp) leaq 40(%rsp), %rax movq %rax, 200(%rsp) leaq 36(%rsp), %rax movq %rax, 208(%rsp) leaq 32(%rsp), %rax movq %rax, 216(%rsp) leaq 28(%rsp), %rax movq %rax, 224(%rsp) leaq 24(%rsp), %rax movq %rax, 232(%rsp) leaq 20(%rsp), %rax movq %rax, 240(%rsp) .cfi_escape 0x2e, 0x00 leaq 128(%rsp), %rdi leaq 112(%rsp), %rsi leaq 104(%rsp), %rdx leaq 96(%rsp), %rcx callq __hipPopCallConfiguration movq 128(%rsp), %rsi movl 136(%rsp), %edx movq 112(%rsp), %rcx movl 120(%rsp), %r8d .cfi_escape 0x2e, 0x10 leaq 160(%rsp), %r9 movl $_Z11conv_kernelPKfPfiiiiiiiii, %edi pushq 96(%rsp) .cfi_adjust_cfa_offset 8 pushq 112(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_17: movq 8(%rsp), %rsi .cfi_escape 0x2e, 0x00 movl $16744464, %edx # imm = 0xFF8010 movq %rbx, %rdi movl $2, %ecx callq hipMemcpy testl %eax, %eax jne .LBB1_59 # %bb.18: # %.preheader.preheader xorl %ecx, %ecx movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero xorl %eax, %eax movq %rax, (%rsp) # 8-byte Spill xorl %r12d, %r12d xorl %eax, %eax movq %rbx, 56(%rsp) # 8-byte Spill movq %r15, 72(%rsp) # 8-byte Spill jmp .LBB1_19 .p2align 4, 0x90 .LBB1_24: # %_Z9gpuAssert10hipError_tPKcib.exit93 # in Loop: Header=BB1_19 Depth=1 incq %rcx cmpq $2046, %rcx # imm = 0x7FE je .LBB1_25 .LBB1_19: # %.preheader # =>This Loop Header: Depth=1 # Child Loop BB1_20 Depth 2 # Child Loop BB1_47 Depth 3 imulq $8184, %rcx, %rdx # imm = 0x1FF8 addq %rbx, %rdx xorl %r14d, %r14d movq %rcx, 88(%rsp) # 8-byte Spill movq %rdx, 80(%rsp) # 8-byte Spill jmp .LBB1_20 .p2align 4, 0x90 .LBB1_22: # in Loop: Header=BB1_20 Depth=2 movq %rbp, (%r12) addq $8, %r12 .LBB1_23: # %_ZNSt6vectorISt4pairIiiESaIS1_EE9push_backEOS1_.exit # in Loop: Header=BB1_20 Depth=2 incq %r14 cmpq $2046, %r14 # imm = 0x7FE je .LBB1_24 .LBB1_20: # Parent Loop BB1_19 Depth=1 # => This Loop Header: Depth=2 # Child Loop BB1_47 Depth 3 movss (%rdx,%r14,4), %xmm0 # xmm0 = mem[0],zero,zero,zero ucomiss %xmm1, %xmm0 jne .LBB1_21 jnp .LBB1_23 .LBB1_21: # in Loop: Header=BB1_20 Depth=2 movq %r14, %rbp shlq $32, %rbp orq %rcx, %rbp cmpq %rax, %r12 jne .LBB1_22 # %bb.38: # in Loop: Header=BB1_20 Depth=2 movq %r12, %r13 subq (%rsp), %r13 # 8-byte Folded Reload movabsq $9223372036854775800, %rax # imm = 0x7FFFFFFFFFFFFFF8 cmpq %rax, %r13 je .LBB1_39 # %bb.41: # %_ZNKSt6vectorISt4pairIiiESaIS1_EE12_M_check_lenEmPKc.exit.i # in Loop: Header=BB1_20 Depth=2 sarq $3, %r13 cmpq $1, %r13 movq %r13, %rax adcq $0, %rax leaq (%rax,%r13), %r15 movabsq $1152921504606846975, %rcx # imm = 0xFFFFFFFFFFFFFFF cmpq %rcx, %r15 cmovaeq %rcx, %r15 addq %r13, %rax cmovbq %rcx, %r15 testq %r15, %r15 je .LBB1_42 # %bb.43: # in Loop: Header=BB1_20 Depth=2 leaq (,%r15,8), %rdi .Ltmp0: .cfi_escape 0x2e, 0x00 callq _Znwm .Ltmp1: # %bb.44: # in Loop: Header=BB1_20 Depth=2 movq %rax, %rbx movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero jmp .LBB1_45 .LBB1_42: # in Loop: Header=BB1_20 Depth=2 xorl %ebx, %ebx .LBB1_45: # %_ZNSt12_Vector_baseISt4pairIiiESaIS1_EE11_M_allocateEm.exit.i # in Loop: Header=BB1_20 Depth=2 movq %rbp, (%rbx,%r13,8) movq %rbx, %r13 cmpq %r12, (%rsp) # 8-byte Folded Reload je .LBB1_48 # %bb.46: # %.lr.ph.i.i.i.i.preheader # in Loop: Header=BB1_20 Depth=2 movq %rbx, %r13 movq (%rsp), %rax # 8-byte Reload .p2align 4, 0x90 .LBB1_47: # %.lr.ph.i.i.i.i # Parent Loop BB1_19 Depth=1 # Parent Loop BB1_20 Depth=2 # => This Inner Loop Header: Depth=3 movq (%rax), %rcx movq %rcx, (%r13) addq $8, %rax addq $8, %r13 cmpq %r12, %rax jne .LBB1_47 .LBB1_48: # %_ZNSt6vectorISt4pairIiiESaIS1_EE11_S_relocateEPS1_S4_S4_RS2_.exit.i # in Loop: Header=BB1_20 Depth=2 movq (%rsp), %rdi # 8-byte Reload testq %rdi, %rdi je .LBB1_50 # %bb.49: # in Loop: Header=BB1_20 Depth=2 .cfi_escape 0x2e, 0x00 callq _ZdlPv movss .LCPI1_0(%rip), %xmm1 # xmm1 = mem[0],zero,zero,zero .LBB1_50: # %_ZNSt6vectorISt4pairIiiESaIS1_EE17_M_realloc_insertIJS1_EEEvN9__gnu_cxx17__normal_iteratorIPS1_S3_EEDpOT_.exit # in Loop: Header=BB1_20 Depth=2 addq $8, %r13 leaq (%rbx,%r15,8), %rax movq %r13, %r12 movq %rbx, (%rsp) # 8-byte Spill movq 72(%rsp), %r15 # 8-byte Reload movq 56(%rsp), %rbx # 8-byte Reload movq 88(%rsp), %rcx # 8-byte Reload movq 80(%rsp), %rdx # 8-byte Reload jmp .LBB1_23 .LBB1_25: movq %r12, %r14 subq (%rsp), %r14 # 8-byte Folded Reload sarq $3, %r14 .Ltmp3: .cfi_escape 0x2e, 0x00 movl $_ZSt4cout, %edi movq %r14, %rsi callq _ZNSo9_M_insertImEERSoT_ .Ltmp4: # %bb.26: # %_ZNSolsEm.exit .Ltmp5: .cfi_escape 0x2e, 0x00 movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp6: # %bb.27: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit.preheader movq (%rsp), %rax # 8-byte Reload cmpq %rax, %r12 je .LBB1_34 # %bb.28: # %.lr.ph.preheader movq %rax, %r12 cmpq $1, %r14 adcq $0, %r14 xorl %ebx, %ebx .p2align 4, 0x90 .LBB1_29: # %.lr.ph # =>This Inner Loop Header: Depth=1 movl (%r12,%rbx,8), %esi .Ltmp7: .cfi_escape 0x2e, 0x00 movl $_ZSt4cout, %edi callq _ZNSolsEi .Ltmp8: # %bb.30: # in Loop: Header=BB1_29 Depth=1 .Ltmp9: movq %rax, %r13 .cfi_escape 0x2e, 0x00 movl $.L.str.2, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp10: # %bb.31: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit96 # in Loop: Header=BB1_29 Depth=1 movl 4(%r12,%rbx,8), %esi .Ltmp11: .cfi_escape 0x2e, 0x00 movq %r13, %rdi callq _ZNSolsEi .Ltmp12: # %bb.32: # in Loop: Header=BB1_29 Depth=1 .Ltmp13: .cfi_escape 0x2e, 0x00 movl $.L.str.1, %esi movl $1, %edx movq %rax, %rdi callq _ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_PKS3_l .Ltmp14: # %bb.33: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit98 # in Loop: Header=BB1_29 Depth=1 incq %rbx cmpq %rbx, %r14 jne .LBB1_29 .LBB1_34: # %_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc.exit._crit_edge .Ltmp16: .cfi_escape 0x2e, 0x00 callq hipDeviceReset .Ltmp17: # %bb.35: .cfi_escape 0x2e, 0x00 movq %r15, %rdi callq _ZdlPv .cfi_escape 0x2e, 0x00 movq 56(%rsp), %rdi # 8-byte Reload callq _ZdlPv movq (%rsp), %rdi # 8-byte Reload testq %rdi, %rdi je .LBB1_37 # %bb.36: .cfi_escape 0x2e, 0x00 callq _ZdlPv .LBB1_37: # %_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev.exit xorl %eax, %eax addq $248, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %r12 .cfi_def_cfa_offset 40 popq %r13 .cfi_def_cfa_offset 32 popq %r14 .cfi_def_cfa_offset 24 popq %r15 .cfi_def_cfa_offset 16 popq %rbp .cfi_def_cfa_offset 8 retq .LBB1_39: .cfi_def_cfa_offset 304 .Ltmp19: .cfi_escape 0x2e, 0x00 movl $.L.str.4, %edi callq _ZSt20__throw_length_errorPKc .Ltmp20: # %bb.40: # %.noexc102 .LBB1_3: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $233, %r8d jmp .LBB1_4 .LBB1_6: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $234, %r8d jmp .LBB1_4 .LBB1_8: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $242, %r8d jmp .LBB1_4 .LBB1_12: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $253, %r8d jmp .LBB1_4 .LBB1_14: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $255, %r8d jmp .LBB1_4 .LBB1_59: movq stderr(%rip), %rbx .cfi_escape 0x2e, 0x00 movl %eax, %edi movl %eax, %ebp callq hipGetErrorString .cfi_escape 0x2e, 0x00 movl $.L.str.3, %esi movl $.L.str, %ecx movq %rbx, %rdi movq %rax, %rdx movl $267, %r8d # imm = 0x10B .LBB1_4: xorl %eax, %eax callq fprintf .cfi_escape 0x2e, 0x00 movl %ebp, %edi callq exit .LBB1_54: .Ltmp18: jmp .LBB1_52 .LBB1_55: .Ltmp15: movq %rax, %rbx jmp .LBB1_56 .LBB1_51: # %.loopexit .Ltmp2: jmp .LBB1_52 .LBB1_53: # %.loopexit.split-lp .Ltmp21: .LBB1_52: movq %rax, %rbx movq (%rsp), %r12 # 8-byte Reload .LBB1_56: testq %r12, %r12 je .LBB1_58 # %bb.57: .cfi_escape 0x2e, 0x00 movq %r12, %rdi callq _ZdlPv .LBB1_58: # %_ZNSt6vectorISt4pairIiiESaIS1_EED2Ev.exit100 .cfi_escape 0x2e, 0x00 movq %rbx, %rdi callq _Unwind_Resume@PLT .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc .section .gcc_except_table,"a",@progbits .p2align 2, 0x0 GCC_except_table1: .Lexception0: .byte 255 # @LPStart Encoding = omit .byte 255 # @TType Encoding = omit .byte 1 # Call site Encoding = uleb128 .uleb128 .Lcst_end0-.Lcst_begin0 .Lcst_begin0: .uleb128 .Lfunc_begin0-.Lfunc_begin0 # >> Call Site 1 << .uleb128 .Ltmp0-.Lfunc_begin0 # Call between .Lfunc_begin0 and .Ltmp0 .byte 0 # has no landing pad .byte 0 # On action: cleanup .uleb128 .Ltmp0-.Lfunc_begin0 # >> Call Site 2 << .uleb128 .Ltmp1-.Ltmp0 # Call between .Ltmp0 and .Ltmp1 .uleb128 .Ltmp2-.Lfunc_begin0 # jumps to .Ltmp2 .byte 0 # On action: cleanup .uleb128 .Ltmp3-.Lfunc_begin0 # >> Call Site 3 << .uleb128 .Ltmp6-.Ltmp3 # Call between .Ltmp3 and .Ltmp6 .uleb128 .Ltmp18-.Lfunc_begin0 # jumps to .Ltmp18 .byte 0 # On action: cleanup .uleb128 .Ltmp7-.Lfunc_begin0 # >> Call Site 4 << .uleb128 .Ltmp14-.Ltmp7 # Call between .Ltmp7 and .Ltmp14 .uleb128 .Ltmp15-.Lfunc_begin0 # jumps to .Ltmp15 .byte 0 # On action: cleanup .uleb128 .Ltmp16-.Lfunc_begin0 # >> Call Site 5 << .uleb128 .Ltmp17-.Ltmp16 # Call between .Ltmp16 and .Ltmp17 .uleb128 .Ltmp18-.Lfunc_begin0 # jumps to .Ltmp18 .byte 0 # On action: cleanup .uleb128 .Ltmp19-.Lfunc_begin0 # >> Call Site 6 << .uleb128 .Ltmp20-.Ltmp19 # Call between .Ltmp19 and .Ltmp20 .uleb128 .Ltmp21-.Lfunc_begin0 # jumps to .Ltmp21 .byte 0 # On action: cleanup .uleb128 .Ltmp20-.Lfunc_begin0 # >> Call Site 7 << .uleb128 .Lfunc_end1-.Ltmp20 # Call between .Ltmp20 and .Lfunc_end1 .byte 0 # has no landing pad .byte 0 # On action: cleanup .Lcst_end0: .p2align 2, 0x0 # -- End function .text .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: 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 $_Z11conv_kernelPKfPfiiiiiiiii, %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 $1, (%rsp) movl $cm, %esi movl $.L__unnamed_2, %edx movl $.L__unnamed_2, %ecx movl $36, %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 cm,@object # @cm .local cm .comm cm,36,16 .type _Z11conv_kernelPKfPfiiiiiiiii,@object # @_Z11conv_kernelPKfPfiiiiiiiii .section .rodata,"a",@progbits .globl _Z11conv_kernelPKfPfiiiiiiiii .p2align 3, 0x0 _Z11conv_kernelPKfPfiiiiiiiii: .quad _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .size _Z11conv_kernelPKfPfiiiiiiiii, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "/home/ubuntu/Datasets/stackv2/train-structured-repos-hip/yottabytt/convolution_kernel/master/convolution.hip" .size .L.str, 109 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "\n" .size .L.str.1, 2 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "," .size .L.str.2, 2 .type .L.str.3,@object # @.str.3 .L.str.3: .asciz "GPUassert: %s %s %d\n" .size .L.str.3, 21 .type .L.str.4,@object # @.str.4 .L.str.4: .asciz "vector::_M_realloc_insert" .size .L.str.4, 26 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z11conv_kernelPKfPfiiiiiiiii" .size .L__unnamed_1, 30 .type .L__unnamed_2,@object # @1 .L__unnamed_2: .asciz "cm" .size .L__unnamed_2, 3 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z26__device_stub__conv_kernelPKfPfiiiiiiiii .addrsig_sym __gxx_personality_v0 .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Unwind_Resume .addrsig_sym cm .addrsig_sym _Z11conv_kernelPKfPfiiiiiiiii .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #include <stdlib.h> #include <stdint.h> #include <string.h> #include <math.h> #include <time.h> #include <sys/time.h> // Necessary for random numbers in CUDA #include <curand_kernel.h> #include <curand.h> #define NUM_ITER 1000000000 #define TPB 128 // Threads PER block #define NUM_THREADS 10000 // Total number of threads to execute /** * Function which, for each instance of the kernel, generates a random point * and calculates whether or not it is within a circle. * * @param counts Array for each thread to store the total number of * randomly generate points that were within a circle * @param numIter Number of iterations / points each thread should make * @param numThreads Number of threads that should be doing work * @param curandState Array for each thread to store its own curandState * structure */ __global__ void estimatePiKernel(unsigned int *counts, unsigned int numIter, unsigned int numThreads, curandState *randState) { double x, y, distance; // Unique ID of the current thread to determine what work to compute int threadId = blockIdx.x * blockDim.x + threadIdx.x; // This thread has no work to do, exit if (threadId > numThreads) return; // Used threadId as a seed of randomness so that every thread is generating // different random values int seed = threadId; curand_init(threadId, seed, 0, &randState[threadId]); for (int iter = 0; iter < numIter; iter++) { // Generate random x, y coordinates from 0.0 (exclusive) to 1.0 // (inclusive) for a point x = (double) curand_uniform(&randState[threadId]); y = (double) curand_uniform(&randState[threadId]); // Distance from the origin of the circle distance = sqrt((x * x) + (y * y)); // If the distance from the origin of the circle is less than or equal // to 1, that means that the randomly generated point is inside the // circle because the circle has a radius of 1. Increment number of // points randomly generated within the circle if (distance <= 1.0) counts[threadId]++; } } /** * Tally up the counts in an array indicating the number of randomly generated * points that were inside a circle and estimate pi. * * @param counts Array of counts of points generate inside a circle */ void estimatePi(unsigned int *counts) { unsigned int totalCount = 0; // accumulate the counts of coins in the circle into totalCount for (int index = 0; index < NUM_THREADS; index++) { totalCount += counts[index]; } printf("total count: %d\n", totalCount); // Calculate pi according to the formula P(coin in circle) * 4 where // P(coin in circle) is equivalents to (coins in circle) / (total coins) double piEstimation = ((double) totalCount / (double) NUM_ITER) * 4.0; printf("The result is %f\n", piEstimation); } /** * Return a timestamp with double percision. */ double cpuSecond() { struct timeval tp; gettimeofday(&tp,NULL); return ((double)tp.tv_sec + (double)tp.tv_usec*1.e-6); } int main() { // Allocate space for curandState for each thread curandState *randState; cudaMalloc(&randState, NUM_THREADS * sizeof(curandState)); // Allocate space to keep track of counts of points generated in the circle unsigned int *deviceCounts; cudaMalloc(&deviceCounts, NUM_THREADS * sizeof(unsigned int)); // Allocate space to copy the GPU result back to the CPU unsigned int *hostCounts = (unsigned int*) malloc( NUM_THREADS * sizeof(unsigned int)); // Set all of the memory to 0 cudaMemset(deviceCounts, 0, NUM_THREADS * sizeof(unsigned int)); double startTime = cpuSecond(); // Launch the kernel estimatePiKernel <<<(NUM_THREADS + TPB - 1) / TPB, TPB>>> ( deviceCounts, NUM_ITER / NUM_THREADS, NUM_THREADS, randState); // Watch for the kernel to finish cudaDeviceSynchronize(); printf("Total time: %f\n", cpuSecond() - startTime); // Copy GPU counts to the CPU cudaMemcpy( hostCounts, deviceCounts, NUM_THREADS * sizeof(unsigned int), cudaMemcpyDeviceToHost ); // Print pi estimation estimatePi(hostCounts); return 0; }
.file "tmpxft_000f7565_00000000-6_exercise_bonus.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2276: .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 .LFE2276: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "total count: %d\n" .LC3: .string "The result is %f\n" .text .globl _Z10estimatePiPj .type _Z10estimatePiPj, @function _Z10estimatePiPj: .LFB2271: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq 40000(%rdi), %rdx movl $0, %eax .L4: addl (%rdi), %eax movl %eax, %ebx addq $4, %rdi cmpq %rdx, %rdi jne .L4 movl %eax, %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %ebx pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 mulsd .LC2(%rip), %xmm0 leaq .LC3(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2271: .size _Z10estimatePiPj, .-_Z10estimatePiPj .globl _Z9cpuSecondv .type _Z9cpuSecondv, @function _Z9cpuSecondv: .LFB2272: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $0, %esi call gettimeofday@PLT pxor %xmm0, %xmm0 cvtsi2sdq 8(%rsp), %xmm0 mulsd .LC4(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq (%rsp), %xmm1 addsd %xmm1, %xmm0 movq 24(%rsp), %rax subq %fs:40, %rax jne .L12 addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L12: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2272: .size _Z9cpuSecondv, .-_Z9cpuSecondv .globl _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW .type _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW, @function _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW: .LFB2298: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movq %rcx, 8(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 20(%rsp), %rax movq %rax, 104(%rsp) leaq 16(%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 .L17 .L13: movq 136(%rsp), %rax subq %fs:40, %rax jne .L18 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L17: .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 _Z16estimatePiKernelPjjjP17curandStateXORWOW(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L13 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE2298: .size _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW, .-_Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW .globl _Z16estimatePiKernelPjjjP17curandStateXORWOW .type _Z16estimatePiKernelPjjjP17curandStateXORWOW, @function _Z16estimatePiKernelPjjjP17curandStateXORWOW: .LFB2299: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2299: .size _Z16estimatePiKernelPjjjP17curandStateXORWOW, .-_Z16estimatePiKernelPjjjP17curandStateXORWOW .section .rodata.str1.1 .LC5: .string "Total time: %f\n" .text .globl main .type main, @function main: .LFB2273: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $64, %rsp .cfi_def_cfa_offset 80 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 16(%rsp), %rdi movl $480000, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $40000, %esi call cudaMalloc@PLT movl $40000, %edi call malloc@PLT movq %rax, %rbx movl $40000, %edx movl $0, %esi movq 24(%rsp), %rdi call cudaMemset@PLT call _Z9cpuSecondv movsd %xmm0, 8(%rsp) movl $128, 44(%rsp) movl $1, 48(%rsp) movl $79, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L22: call cudaDeviceSynchronize@PLT call _Z9cpuSecondv subsd 8(%rsp), %xmm0 leaq .LC5(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl $2, %ecx movl $40000, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rdi call _Z10estimatePiPj movq 56(%rsp), %rax subq %fs:40, %rax jne .L26 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L25: .cfi_restore_state movq 16(%rsp), %rcx movl $10000, %edx movl $100000, %esi movq 24(%rsp), %rdi call _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW jmp .L22 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE2273: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC6: .string "_Z16estimatePiKernelPjjjP17curandStateXORWOW" .section .rodata.str1.1 .LC7: .string "precalc_xorwow_matrix" .LC8: .string "precalc_xorwow_offset_matrix" .LC9: .string "mrg32k3aM1" .LC10: .string "mrg32k3aM2" .LC11: .string "mrg32k3aM1SubSeq" .LC12: .string "mrg32k3aM2SubSeq" .LC13: .string "mrg32k3aM1Seq" .LC14: .string "mrg32k3aM2Seq" .LC15: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2301: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z16estimatePiKernelPjjjP17curandStateXORWOW(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _ZL21precalc_xorwow_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _ZL28precalc_xorwow_offset_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM1(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM2(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM1SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM2SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC13(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM1Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC14(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM2Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $72, %r9d movl $0, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _ZL17__cr_lgamma_table(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2301: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1104006501 .align 8 .LC2: .long 0 .long 1074790400 .align 8 .LC4: .long -1598689907 .long 1051772663 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #include <stdlib.h> #include <stdint.h> #include <string.h> #include <math.h> #include <time.h> #include <sys/time.h> // Necessary for random numbers in CUDA #include <curand_kernel.h> #include <curand.h> #define NUM_ITER 1000000000 #define TPB 128 // Threads PER block #define NUM_THREADS 10000 // Total number of threads to execute /** * Function which, for each instance of the kernel, generates a random point * and calculates whether or not it is within a circle. * * @param counts Array for each thread to store the total number of * randomly generate points that were within a circle * @param numIter Number of iterations / points each thread should make * @param numThreads Number of threads that should be doing work * @param curandState Array for each thread to store its own curandState * structure */ __global__ void estimatePiKernel(unsigned int *counts, unsigned int numIter, unsigned int numThreads, curandState *randState) { double x, y, distance; // Unique ID of the current thread to determine what work to compute int threadId = blockIdx.x * blockDim.x + threadIdx.x; // This thread has no work to do, exit if (threadId > numThreads) return; // Used threadId as a seed of randomness so that every thread is generating // different random values int seed = threadId; curand_init(threadId, seed, 0, &randState[threadId]); for (int iter = 0; iter < numIter; iter++) { // Generate random x, y coordinates from 0.0 (exclusive) to 1.0 // (inclusive) for a point x = (double) curand_uniform(&randState[threadId]); y = (double) curand_uniform(&randState[threadId]); // Distance from the origin of the circle distance = sqrt((x * x) + (y * y)); // If the distance from the origin of the circle is less than or equal // to 1, that means that the randomly generated point is inside the // circle because the circle has a radius of 1. Increment number of // points randomly generated within the circle if (distance <= 1.0) counts[threadId]++; } } /** * Tally up the counts in an array indicating the number of randomly generated * points that were inside a circle and estimate pi. * * @param counts Array of counts of points generate inside a circle */ void estimatePi(unsigned int *counts) { unsigned int totalCount = 0; // accumulate the counts of coins in the circle into totalCount for (int index = 0; index < NUM_THREADS; index++) { totalCount += counts[index]; } printf("total count: %d\n", totalCount); // Calculate pi according to the formula P(coin in circle) * 4 where // P(coin in circle) is equivalents to (coins in circle) / (total coins) double piEstimation = ((double) totalCount / (double) NUM_ITER) * 4.0; printf("The result is %f\n", piEstimation); } /** * Return a timestamp with double percision. */ double cpuSecond() { struct timeval tp; gettimeofday(&tp,NULL); return ((double)tp.tv_sec + (double)tp.tv_usec*1.e-6); } int main() { // Allocate space for curandState for each thread curandState *randState; cudaMalloc(&randState, NUM_THREADS * sizeof(curandState)); // Allocate space to keep track of counts of points generated in the circle unsigned int *deviceCounts; cudaMalloc(&deviceCounts, NUM_THREADS * sizeof(unsigned int)); // Allocate space to copy the GPU result back to the CPU unsigned int *hostCounts = (unsigned int*) malloc( NUM_THREADS * sizeof(unsigned int)); // Set all of the memory to 0 cudaMemset(deviceCounts, 0, NUM_THREADS * sizeof(unsigned int)); double startTime = cpuSecond(); // Launch the kernel estimatePiKernel <<<(NUM_THREADS + TPB - 1) / TPB, TPB>>> ( deviceCounts, NUM_ITER / NUM_THREADS, NUM_THREADS, randState); // Watch for the kernel to finish cudaDeviceSynchronize(); printf("Total time: %f\n", cpuSecond() - startTime); // Copy GPU counts to the CPU cudaMemcpy( hostCounts, deviceCounts, NUM_THREADS * sizeof(unsigned int), cudaMemcpyDeviceToHost ); // Print pi estimation estimatePi(hostCounts); return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <stdint.h> #include <string.h> #include <math.h> #include <time.h> #include <sys/time.h> // Necessary for random numbers in CUDA #include <hiprand/hiprand_kernel.h> #include <hiprand/hiprand.h> #define NUM_ITER 1000000000 #define TPB 128 // Threads PER block #define NUM_THREADS 10000 // Total number of threads to execute /** * Function which, for each instance of the kernel, generates a random point * and calculates whether or not it is within a circle. * * @param counts Array for each thread to store the total number of * randomly generate points that were within a circle * @param numIter Number of iterations / points each thread should make * @param numThreads Number of threads that should be doing work * @param curandState Array for each thread to store its own curandState * structure */ __global__ void estimatePiKernel(unsigned int *counts, unsigned int numIter, unsigned int numThreads, hiprandState *randState) { double x, y, distance; // Unique ID of the current thread to determine what work to compute int threadId = blockIdx.x * blockDim.x + threadIdx.x; // This thread has no work to do, exit if (threadId > numThreads) return; // Used threadId as a seed of randomness so that every thread is generating // different random values int seed = threadId; hiprand_init(threadId, seed, 0, &randState[threadId]); for (int iter = 0; iter < numIter; iter++) { // Generate random x, y coordinates from 0.0 (exclusive) to 1.0 // (inclusive) for a point x = (double) hiprand_uniform(&randState[threadId]); y = (double) hiprand_uniform(&randState[threadId]); // Distance from the origin of the circle distance = sqrt((x * x) + (y * y)); // If the distance from the origin of the circle is less than or equal // to 1, that means that the randomly generated point is inside the // circle because the circle has a radius of 1. Increment number of // points randomly generated within the circle if (distance <= 1.0) counts[threadId]++; } } /** * Tally up the counts in an array indicating the number of randomly generated * points that were inside a circle and estimate pi. * * @param counts Array of counts of points generate inside a circle */ void estimatePi(unsigned int *counts) { unsigned int totalCount = 0; // accumulate the counts of coins in the circle into totalCount for (int index = 0; index < NUM_THREADS; index++) { totalCount += counts[index]; } printf("total count: %d\n", totalCount); // Calculate pi according to the formula P(coin in circle) * 4 where // P(coin in circle) is equivalents to (coins in circle) / (total coins) double piEstimation = ((double) totalCount / (double) NUM_ITER) * 4.0; printf("The result is %f\n", piEstimation); } /** * Return a timestamp with double percision. */ double cpuSecond() { struct timeval tp; gettimeofday(&tp,NULL); return ((double)tp.tv_sec + (double)tp.tv_usec*1.e-6); } int main() { // Allocate space for curandState for each thread hiprandState *randState; hipMalloc(&randState, NUM_THREADS * sizeof(hiprandState)); // Allocate space to keep track of counts of points generated in the circle unsigned int *deviceCounts; hipMalloc(&deviceCounts, NUM_THREADS * sizeof(unsigned int)); // Allocate space to copy the GPU result back to the CPU unsigned int *hostCounts = (unsigned int*) malloc( NUM_THREADS * sizeof(unsigned int)); // Set all of the memory to 0 hipMemset(deviceCounts, 0, NUM_THREADS * sizeof(unsigned int)); double startTime = cpuSecond(); // Launch the kernel estimatePiKernel <<<(NUM_THREADS + TPB - 1) / TPB, TPB>>> ( deviceCounts, NUM_ITER / NUM_THREADS, NUM_THREADS, randState); // Watch for the kernel to finish hipDeviceSynchronize(); printf("Total time: %f\n", cpuSecond() - startTime); // Copy GPU counts to the CPU hipMemcpy( hostCounts, deviceCounts, NUM_THREADS * sizeof(unsigned int), hipMemcpyDeviceToHost ); // Print pi estimation estimatePi(hostCounts); return 0; }
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> #include <stdint.h> #include <string.h> #include <math.h> #include <time.h> #include <sys/time.h> // Necessary for random numbers in CUDA #include <hiprand/hiprand_kernel.h> #include <hiprand/hiprand.h> #define NUM_ITER 1000000000 #define TPB 128 // Threads PER block #define NUM_THREADS 10000 // Total number of threads to execute /** * Function which, for each instance of the kernel, generates a random point * and calculates whether or not it is within a circle. * * @param counts Array for each thread to store the total number of * randomly generate points that were within a circle * @param numIter Number of iterations / points each thread should make * @param numThreads Number of threads that should be doing work * @param curandState Array for each thread to store its own curandState * structure */ __global__ void estimatePiKernel(unsigned int *counts, unsigned int numIter, unsigned int numThreads, hiprandState *randState) { double x, y, distance; // Unique ID of the current thread to determine what work to compute int threadId = blockIdx.x * blockDim.x + threadIdx.x; // This thread has no work to do, exit if (threadId > numThreads) return; // Used threadId as a seed of randomness so that every thread is generating // different random values int seed = threadId; hiprand_init(threadId, seed, 0, &randState[threadId]); for (int iter = 0; iter < numIter; iter++) { // Generate random x, y coordinates from 0.0 (exclusive) to 1.0 // (inclusive) for a point x = (double) hiprand_uniform(&randState[threadId]); y = (double) hiprand_uniform(&randState[threadId]); // Distance from the origin of the circle distance = sqrt((x * x) + (y * y)); // If the distance from the origin of the circle is less than or equal // to 1, that means that the randomly generated point is inside the // circle because the circle has a radius of 1. Increment number of // points randomly generated within the circle if (distance <= 1.0) counts[threadId]++; } } /** * Tally up the counts in an array indicating the number of randomly generated * points that were inside a circle and estimate pi. * * @param counts Array of counts of points generate inside a circle */ void estimatePi(unsigned int *counts) { unsigned int totalCount = 0; // accumulate the counts of coins in the circle into totalCount for (int index = 0; index < NUM_THREADS; index++) { totalCount += counts[index]; } printf("total count: %d\n", totalCount); // Calculate pi according to the formula P(coin in circle) * 4 where // P(coin in circle) is equivalents to (coins in circle) / (total coins) double piEstimation = ((double) totalCount / (double) NUM_ITER) * 4.0; printf("The result is %f\n", piEstimation); } /** * Return a timestamp with double percision. */ double cpuSecond() { struct timeval tp; gettimeofday(&tp,NULL); return ((double)tp.tv_sec + (double)tp.tv_usec*1.e-6); } int main() { // Allocate space for curandState for each thread hiprandState *randState; hipMalloc(&randState, NUM_THREADS * sizeof(hiprandState)); // Allocate space to keep track of counts of points generated in the circle unsigned int *deviceCounts; hipMalloc(&deviceCounts, NUM_THREADS * sizeof(unsigned int)); // Allocate space to copy the GPU result back to the CPU unsigned int *hostCounts = (unsigned int*) malloc( NUM_THREADS * sizeof(unsigned int)); // Set all of the memory to 0 hipMemset(deviceCounts, 0, NUM_THREADS * sizeof(unsigned int)); double startTime = cpuSecond(); // Launch the kernel estimatePiKernel <<<(NUM_THREADS + TPB - 1) / TPB, TPB>>> ( deviceCounts, NUM_ITER / NUM_THREADS, NUM_THREADS, randState); // Watch for the kernel to finish hipDeviceSynchronize(); printf("Total time: %f\n", cpuSecond() - startTime); // Copy GPU counts to the CPU hipMemcpy( hostCounts, deviceCounts, NUM_THREADS * sizeof(unsigned int), hipMemcpyDeviceToHost ); // Print pi estimation estimatePi(hostCounts); return 0; }
.text .file "exercise_bonus.hip" .globl _Z31__device_stub__estimatePiKernelPjjjP12hiprandState # -- Begin function _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .p2align 4, 0x90 .type _Z31__device_stub__estimatePiKernelPjjjP12hiprandState,@function _Z31__device_stub__estimatePiKernelPjjjP12hiprandState: # @_Z31__device_stub__estimatePiKernelPjjjP12hiprandState .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 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 $_Z16estimatePiKernelPjjjP12hiprandState, %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 _Z31__device_stub__estimatePiKernelPjjjP12hiprandState, .Lfunc_end0-_Z31__device_stub__estimatePiKernelPjjjP12hiprandState .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z10estimatePiPj .LCPI1_0: .quad 0x41cdcd6500000000 # double 1.0E+9 .LCPI1_1: .quad 0x4010000000000000 # double 4 .text .globl _Z10estimatePiPj .p2align 4, 0x90 .type _Z10estimatePiPj,@function _Z10estimatePiPj: # @_Z10estimatePiPj .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 xorl %eax, %eax xorl %ebx, %ebx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 addl (%rdi,%rax,4), %ebx incq %rax cmpq $10000, %rax # imm = 0x2710 jne .LBB1_1 # %bb.2: movl $.L.str, %edi movl %ebx, %esi xorl %eax, %eax callq printf movl %ebx, %eax cvtsi2sd %rax, %xmm0 divsd .LCPI1_0(%rip), %xmm0 mulsd .LCPI1_1(%rip), %xmm0 movl $.L.str.1, %edi movb $1, %al popq %rbx .cfi_def_cfa_offset 8 jmp printf # TAILCALL .Lfunc_end1: .size _Z10estimatePiPj, .Lfunc_end1-_Z10estimatePiPj .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z9cpuSecondv .LCPI2_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .text .globl _Z9cpuSecondv .p2align 4, 0x90 .type _Z9cpuSecondv,@function _Z9cpuSecondv: # @_Z9cpuSecondv .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 leaq 8(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 8(%rsp), %xmm1 cvtsi2sdq 16(%rsp), %xmm0 mulsd .LCPI2_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9cpuSecondv, .Lfunc_end2-_Z9cpuSecondv .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI3_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .LCPI3_1: .quad 0x41cdcd6500000000 # double 1.0E+9 .LCPI3_2: .quad 0x4010000000000000 # double 4 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $136, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 64(%rsp), %rdi movl $480000, %esi # imm = 0x75300 callq hipMalloc movq %rsp, %rdi movl $40000, %esi # imm = 0x9C40 callq hipMalloc movl $40000, %edi # imm = 0x9C40 callq malloc movq %rax, %r14 movq (%rsp), %rdi xorl %ebx, %ebx movl $40000, %edx # imm = 0x9C40 xorl %esi, %esi callq hipMemset leaq 16(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 16(%rsp), %xmm0 cvtsi2sdq 24(%rsp), %xmm1 mulsd .LCPI3_0(%rip), %xmm1 addsd %xmm0, %xmm1 movsd %xmm1, 56(%rsp) # 8-byte Spill movabsq $4294967375, %rdi # imm = 0x10000004F leaq 49(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_2 # %bb.1: movq (%rsp), %rax movq 64(%rsp), %rcx movq %rax, 128(%rsp) movl $100000, 12(%rsp) # imm = 0x186A0 movl $10000, 8(%rsp) # imm = 0x2710 movq %rcx, 120(%rsp) leaq 128(%rsp), %rax movq %rax, 16(%rsp) leaq 12(%rsp), %rax movq %rax, 24(%rsp) leaq 8(%rsp), %rax movq %rax, 32(%rsp) leaq 120(%rsp), %rax movq %rax, 40(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z16estimatePiKernelPjjjP12hiprandState, %edi pushq 72(%rsp) .cfi_adjust_cfa_offset 8 pushq 88(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_2: callq hipDeviceSynchronize leaq 16(%rsp), %rdi xorl %esi, %esi callq gettimeofday xorps %xmm1, %xmm1 cvtsi2sdq 16(%rsp), %xmm1 xorps %xmm0, %xmm0 cvtsi2sdq 24(%rsp), %xmm0 mulsd .LCPI3_0(%rip), %xmm0 addsd %xmm1, %xmm0 subsd 56(%rsp), %xmm0 # 8-byte Folded Reload movl $.L.str.2, %edi movb $1, %al callq printf movq (%rsp), %rsi movl $40000, %edx # imm = 0x9C40 movq %r14, %rdi movl $2, %ecx callq hipMemcpy xorl %eax, %eax .p2align 4, 0x90 .LBB3_3: # =>This Inner Loop Header: Depth=1 addl (%r14,%rax,4), %ebx incq %rax cmpq $10000, %rax # imm = 0x2710 jne .LBB3_3 # %bb.4: # %_Z10estimatePiPj.exit movl $.L.str, %edi movl %ebx, %esi xorl %eax, %eax callq printf movl %ebx, %eax xorps %xmm0, %xmm0 cvtsi2sd %rax, %xmm0 divsd .LCPI3_1(%rip), %xmm0 mulsd .LCPI3_2(%rip), %xmm0 movl $.L.str.1, %edi movb $1, %al callq printf xorl %eax, %eax addq $136, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z16estimatePiKernelPjjjP12hiprandState, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type _Z16estimatePiKernelPjjjP12hiprandState,@object # @_Z16estimatePiKernelPjjjP12hiprandState .section .rodata,"a",@progbits .globl _Z16estimatePiKernelPjjjP12hiprandState .p2align 3, 0x0 _Z16estimatePiKernelPjjjP12hiprandState: .quad _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .size _Z16estimatePiKernelPjjjP12hiprandState, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "total count: %d\n" .size .L.str, 17 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "The result is %f\n" .size .L.str.1, 18 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Total time: %f\n" .size .L.str.2, 16 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z16estimatePiKernelPjjjP12hiprandState" .size .L__unnamed_1, 40 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16estimatePiKernelPjjjP12hiprandState .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_000f7565_00000000-6_exercise_bonus.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2276: .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 .LFE2276: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "total count: %d\n" .LC3: .string "The result is %f\n" .text .globl _Z10estimatePiPj .type _Z10estimatePiPj, @function _Z10estimatePiPj: .LFB2271: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq 40000(%rdi), %rdx movl $0, %eax .L4: addl (%rdi), %eax movl %eax, %ebx addq $4, %rdi cmpq %rdx, %rdi jne .L4 movl %eax, %edx leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT movl %ebx, %ebx pxor %xmm0, %xmm0 cvtsi2sdq %rbx, %xmm0 divsd .LC1(%rip), %xmm0 mulsd .LC2(%rip), %xmm0 leaq .LC3(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2271: .size _Z10estimatePiPj, .-_Z10estimatePiPj .globl _Z9cpuSecondv .type _Z9cpuSecondv, @function _Z9cpuSecondv: .LFB2272: .cfi_startproc endbr64 subq $40, %rsp .cfi_def_cfa_offset 48 movq %fs:40, %rax movq %rax, 24(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $0, %esi call gettimeofday@PLT pxor %xmm0, %xmm0 cvtsi2sdq 8(%rsp), %xmm0 mulsd .LC4(%rip), %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq (%rsp), %xmm1 addsd %xmm1, %xmm0 movq 24(%rsp), %rax subq %fs:40, %rax jne .L12 addq $40, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L12: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2272: .size _Z9cpuSecondv, .-_Z9cpuSecondv .globl _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW .type _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW, @function _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW: .LFB2298: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movl %esi, 20(%rsp) movl %edx, 16(%rsp) movq %rcx, 8(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 20(%rsp), %rax movq %rax, 104(%rsp) leaq 16(%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 .L17 .L13: movq 136(%rsp), %rax subq %fs:40, %rax jne .L18 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L17: .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 _Z16estimatePiKernelPjjjP17curandStateXORWOW(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L13 .L18: call __stack_chk_fail@PLT .cfi_endproc .LFE2298: .size _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW, .-_Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW .globl _Z16estimatePiKernelPjjjP17curandStateXORWOW .type _Z16estimatePiKernelPjjjP17curandStateXORWOW, @function _Z16estimatePiKernelPjjjP17curandStateXORWOW: .LFB2299: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2299: .size _Z16estimatePiKernelPjjjP17curandStateXORWOW, .-_Z16estimatePiKernelPjjjP17curandStateXORWOW .section .rodata.str1.1 .LC5: .string "Total time: %f\n" .text .globl main .type main, @function main: .LFB2273: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 subq $64, %rsp .cfi_def_cfa_offset 80 movq %fs:40, %rax movq %rax, 56(%rsp) xorl %eax, %eax leaq 16(%rsp), %rdi movl $480000, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $40000, %esi call cudaMalloc@PLT movl $40000, %edi call malloc@PLT movq %rax, %rbx movl $40000, %edx movl $0, %esi movq 24(%rsp), %rdi call cudaMemset@PLT call _Z9cpuSecondv movsd %xmm0, 8(%rsp) movl $128, 44(%rsp) movl $1, 48(%rsp) movl $79, 32(%rsp) movl $1, 36(%rsp) movl $0, %r9d movl $0, %r8d movq 44(%rsp), %rdx movl $1, %ecx movq 32(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L22: call cudaDeviceSynchronize@PLT call _Z9cpuSecondv subsd 8(%rsp), %xmm0 leaq .LC5(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movl $2, %ecx movl $40000, %edx movq 24(%rsp), %rsi movq %rbx, %rdi call cudaMemcpy@PLT movq %rbx, %rdi call _Z10estimatePiPj movq 56(%rsp), %rax subq %fs:40, %rax jne .L26 movl $0, %eax addq $64, %rsp .cfi_remember_state .cfi_def_cfa_offset 16 popq %rbx .cfi_def_cfa_offset 8 ret .L25: .cfi_restore_state movq 16(%rsp), %rcx movl $10000, %edx movl $100000, %esi movq 24(%rsp), %rdi call _Z58__device_stub__Z16estimatePiKernelPjjjP17curandStateXORWOWPjjjP17curandStateXORWOW jmp .L22 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE2273: .size main, .-main .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC6: .string "_Z16estimatePiKernelPjjjP17curandStateXORWOW" .section .rodata.str1.1 .LC7: .string "precalc_xorwow_matrix" .LC8: .string "precalc_xorwow_offset_matrix" .LC9: .string "mrg32k3aM1" .LC10: .string "mrg32k3aM2" .LC11: .string "mrg32k3aM1SubSeq" .LC12: .string "mrg32k3aM2SubSeq" .LC13: .string "mrg32k3aM1Seq" .LC14: .string "mrg32k3aM2Seq" .LC15: .string "__cr_lgamma_table" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2301: .cfi_startproc endbr64 pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset 3, -16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rbx movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC6(%rip), %rdx movq %rdx, %rcx leaq _Z16estimatePiKernelPjjjP17curandStateXORWOW(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC7(%rip), %rdx movq %rdx, %rcx leaq _ZL21precalc_xorwow_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $102400, %r9d movl $0, %r8d leaq .LC8(%rip), %rdx movq %rdx, %rcx leaq _ZL28precalc_xorwow_offset_matrix(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC9(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM1(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC10(%rip), %rdx movq %rdx, %rcx leaq _ZL10mrg32k3aM2(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC11(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM1SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2016, %r9d movl $0, %r8d leaq .LC12(%rip), %rdx movq %rdx, %rcx leaq _ZL16mrg32k3aM2SubSeq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC13(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM1Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 movl $2304, %r9d movl $0, %r8d leaq .LC14(%rip), %rdx movq %rdx, %rcx leaq _ZL13mrg32k3aM2Seq(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $1 .cfi_def_cfa_offset 32 movl $72, %r9d movl $0, %r8d leaq .LC15(%rip), %rdx movq %rdx, %rcx leaq _ZL17__cr_lgamma_table(%rip), %rsi movq %rbx, %rdi call __cudaRegisterVar@PLT addq $16, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2301: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .local _ZL17__cr_lgamma_table .comm _ZL17__cr_lgamma_table,72,32 .local _ZL13mrg32k3aM2Seq .comm _ZL13mrg32k3aM2Seq,2304,32 .local _ZL13mrg32k3aM1Seq .comm _ZL13mrg32k3aM1Seq,2304,32 .local _ZL16mrg32k3aM2SubSeq .comm _ZL16mrg32k3aM2SubSeq,2016,32 .local _ZL16mrg32k3aM1SubSeq .comm _ZL16mrg32k3aM1SubSeq,2016,32 .local _ZL10mrg32k3aM2 .comm _ZL10mrg32k3aM2,2304,32 .local _ZL10mrg32k3aM1 .comm _ZL10mrg32k3aM1,2304,32 .local _ZL28precalc_xorwow_offset_matrix .comm _ZL28precalc_xorwow_offset_matrix,102400,32 .local _ZL21precalc_xorwow_matrix .comm _ZL21precalc_xorwow_matrix,102400,32 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC1: .long 0 .long 1104006501 .align 8 .LC2: .long 0 .long 1074790400 .align 8 .LC4: .long -1598689907 .long 1051772663 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "exercise_bonus.hip" .globl _Z31__device_stub__estimatePiKernelPjjjP12hiprandState # -- Begin function _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .p2align 4, 0x90 .type _Z31__device_stub__estimatePiKernelPjjjP12hiprandState,@function _Z31__device_stub__estimatePiKernelPjjjP12hiprandState: # @_Z31__device_stub__estimatePiKernelPjjjP12hiprandState .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 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 $_Z16estimatePiKernelPjjjP12hiprandState, %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 _Z31__device_stub__estimatePiKernelPjjjP12hiprandState, .Lfunc_end0-_Z31__device_stub__estimatePiKernelPjjjP12hiprandState .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z10estimatePiPj .LCPI1_0: .quad 0x41cdcd6500000000 # double 1.0E+9 .LCPI1_1: .quad 0x4010000000000000 # double 4 .text .globl _Z10estimatePiPj .p2align 4, 0x90 .type _Z10estimatePiPj,@function _Z10estimatePiPj: # @_Z10estimatePiPj .cfi_startproc # %bb.0: pushq %rbx .cfi_def_cfa_offset 16 .cfi_offset %rbx, -16 xorl %eax, %eax xorl %ebx, %ebx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 addl (%rdi,%rax,4), %ebx incq %rax cmpq $10000, %rax # imm = 0x2710 jne .LBB1_1 # %bb.2: movl $.L.str, %edi movl %ebx, %esi xorl %eax, %eax callq printf movl %ebx, %eax cvtsi2sd %rax, %xmm0 divsd .LCPI1_0(%rip), %xmm0 mulsd .LCPI1_1(%rip), %xmm0 movl $.L.str.1, %edi movb $1, %al popq %rbx .cfi_def_cfa_offset 8 jmp printf # TAILCALL .Lfunc_end1: .size _Z10estimatePiPj, .Lfunc_end1-_Z10estimatePiPj .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function _Z9cpuSecondv .LCPI2_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .text .globl _Z9cpuSecondv .p2align 4, 0x90 .type _Z9cpuSecondv,@function _Z9cpuSecondv: # @_Z9cpuSecondv .cfi_startproc # %bb.0: subq $24, %rsp .cfi_def_cfa_offset 32 leaq 8(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 8(%rsp), %xmm1 cvtsi2sdq 16(%rsp), %xmm0 mulsd .LCPI2_0(%rip), %xmm0 addsd %xmm1, %xmm0 addq $24, %rsp .cfi_def_cfa_offset 8 retq .Lfunc_end2: .size _Z9cpuSecondv, .Lfunc_end2-_Z9cpuSecondv .cfi_endproc # -- End function .section .rodata.cst8,"aM",@progbits,8 .p2align 3, 0x0 # -- Begin function main .LCPI3_0: .quad 0x3eb0c6f7a0b5ed8d # double 9.9999999999999995E-7 .LCPI3_1: .quad 0x41cdcd6500000000 # double 1.0E+9 .LCPI3_2: .quad 0x4010000000000000 # double 4 .text .globl main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r14 .cfi_def_cfa_offset 16 pushq %rbx .cfi_def_cfa_offset 24 subq $136, %rsp .cfi_def_cfa_offset 160 .cfi_offset %rbx, -24 .cfi_offset %r14, -16 leaq 64(%rsp), %rdi movl $480000, %esi # imm = 0x75300 callq hipMalloc movq %rsp, %rdi movl $40000, %esi # imm = 0x9C40 callq hipMalloc movl $40000, %edi # imm = 0x9C40 callq malloc movq %rax, %r14 movq (%rsp), %rdi xorl %ebx, %ebx movl $40000, %edx # imm = 0x9C40 xorl %esi, %esi callq hipMemset leaq 16(%rsp), %rdi xorl %esi, %esi callq gettimeofday cvtsi2sdq 16(%rsp), %xmm0 cvtsi2sdq 24(%rsp), %xmm1 mulsd .LCPI3_0(%rip), %xmm1 addsd %xmm0, %xmm1 movsd %xmm1, 56(%rsp) # 8-byte Spill movabsq $4294967375, %rdi # imm = 0x10000004F leaq 49(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB3_2 # %bb.1: movq (%rsp), %rax movq 64(%rsp), %rcx movq %rax, 128(%rsp) movl $100000, 12(%rsp) # imm = 0x186A0 movl $10000, 8(%rsp) # imm = 0x2710 movq %rcx, 120(%rsp) leaq 128(%rsp), %rax movq %rax, 16(%rsp) leaq 12(%rsp), %rax movq %rax, 24(%rsp) leaq 8(%rsp), %rax movq %rax, 32(%rsp) leaq 120(%rsp), %rax movq %rax, 40(%rsp) leaq 104(%rsp), %rdi leaq 88(%rsp), %rsi leaq 80(%rsp), %rdx leaq 72(%rsp), %rcx callq __hipPopCallConfiguration movq 104(%rsp), %rsi movl 112(%rsp), %edx movq 88(%rsp), %rcx movl 96(%rsp), %r8d leaq 16(%rsp), %r9 movl $_Z16estimatePiKernelPjjjP12hiprandState, %edi pushq 72(%rsp) .cfi_adjust_cfa_offset 8 pushq 88(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB3_2: callq hipDeviceSynchronize leaq 16(%rsp), %rdi xorl %esi, %esi callq gettimeofday xorps %xmm1, %xmm1 cvtsi2sdq 16(%rsp), %xmm1 xorps %xmm0, %xmm0 cvtsi2sdq 24(%rsp), %xmm0 mulsd .LCPI3_0(%rip), %xmm0 addsd %xmm1, %xmm0 subsd 56(%rsp), %xmm0 # 8-byte Folded Reload movl $.L.str.2, %edi movb $1, %al callq printf movq (%rsp), %rsi movl $40000, %edx # imm = 0x9C40 movq %r14, %rdi movl $2, %ecx callq hipMemcpy xorl %eax, %eax .p2align 4, 0x90 .LBB3_3: # =>This Inner Loop Header: Depth=1 addl (%r14,%rax,4), %ebx incq %rax cmpq $10000, %rax # imm = 0x2710 jne .LBB3_3 # %bb.4: # %_Z10estimatePiPj.exit movl $.L.str, %edi movl %ebx, %esi xorl %eax, %eax callq printf movl %ebx, %eax xorps %xmm0, %xmm0 cvtsi2sd %rax, %xmm0 divsd .LCPI3_1(%rip), %xmm0 mulsd .LCPI3_2(%rip), %xmm0 movl $.L.str.1, %edi movb $1, %al callq printf xorl %eax, %eax addq $136, %rsp .cfi_def_cfa_offset 24 popq %rbx .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 retq .Lfunc_end3: .size main, .Lfunc_end3-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB4_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB4_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z16estimatePiKernelPjjjP12hiprandState, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end4: .size __hip_module_ctor, .Lfunc_end4-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB5_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB5_2: retq .Lfunc_end5: .size __hip_module_dtor, .Lfunc_end5-__hip_module_dtor .cfi_endproc # -- End function .type _Z16estimatePiKernelPjjjP12hiprandState,@object # @_Z16estimatePiKernelPjjjP12hiprandState .section .rodata,"a",@progbits .globl _Z16estimatePiKernelPjjjP12hiprandState .p2align 3, 0x0 _Z16estimatePiKernelPjjjP12hiprandState: .quad _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .size _Z16estimatePiKernelPjjjP12hiprandState, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "total count: %d\n" .size .L.str, 17 .type .L.str.1,@object # @.str.1 .L.str.1: .asciz "The result is %f\n" .size .L.str.1, 18 .type .L.str.2,@object # @.str.2 .L.str.2: .asciz "Total time: %f\n" .size .L.str.2, 16 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z16estimatePiKernelPjjjP12hiprandState" .size .L__unnamed_1, 40 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__estimatePiKernelPjjjP12hiprandState .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16estimatePiKernelPjjjP12hiprandState .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <cstdlib> #include <iostream> #include <stdio.h> #include <cmath> #include <cuda_runtime.h> #include <time.h> #include "device_launch_parameters.h" #include "cuda.h" using namespace std; #define Nn 4096 #define Ni 25088 #define BATCH_SIZE 16 #define BLOCK_SIZE 32 #define BlockSize2D 16 #define VTYPE float /* * synapse (w) is (Nn x Ni)^T * neuron_i (x) is (BATCH_SIZE x Ni) * neuron_n (y) is (BATCH_SIZE x Nn) * * y = Xw^T */ void init_layer(VTYPE* h_neuron_i, VTYPE* h_neuron_n, VTYPE* synapse) { for (int i = 0; i < Nn; i++) { h_neuron_n[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * BATCH_SIZE; i++) { h_neuron_i[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * Nn; i++) { synapse[i] = rand() / (VTYPE)RAND_MAX; } } __launch_bounds__(1024,2) __global__ void d_MatMul_simple1(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; VTYPE temp = 0.0f; if (col < Nn && row < BATCH_SIZE) { #pragma unroll for (int i = 0; i < Ni; i++) { temp += d_neuron_i[row * Ni + i] * synapse[col + Nn * i]; } d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple2(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE neuron_i[BlockSize2D][BlockSize2D]; VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < Ni; i += BlockSize2D) { if (row < BATCH_SIZE && i + threadIdx.x < Ni) { neuron_i[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i + threadIdx.x]; } else { neuron_i[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron_i[threadIdx.y][j] * d_synapse[(j + i)* Nn + col]; } __syncthreads(); } if (col < Nn && row < BATCH_SIZE) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple3(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE synapse[BlockSize2D][BlockSize2D]; __shared__ VTYPE neuron[BlockSize2D][BlockSize2D]; // MxK = MxN * NxK VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < (Ni - 1) / BlockSize2D + 1; i++) { if (row < BATCH_SIZE && i * BlockSize2D + threadIdx.x < Ni) { neuron[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i * BlockSize2D + threadIdx.x]; } else { neuron[threadIdx.y][threadIdx.x] = 0.0f; } if (i * BlockSize2D + threadIdx.y < Ni && col < Nn) { synapse[threadIdx.y][threadIdx.x] = d_synapse[(i * BlockSize2D + threadIdx.y) * Nn + col]; } else { synapse[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron[threadIdx.y][j] * synapse[j][threadIdx.x]; } __syncthreads(); } if (row < BATCH_SIZE && col < Nn) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_test(VTYPE* d_synapse, VTYPE* d_neuron_i) { int idx = blockIdx.x * blockDim.x + threadIdx.x; d_neuron_i[idx] *= 1.1f; } bool compare(VTYPE* neuron1, VTYPE* neuron2) { bool good = true; #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { if (fabs(neuron1[k * Nn + i] - neuron2[k * Nn + i]) > 1e-2) { good = false; printf("At index (%d, %d) \t Host result: %lf \t Device result: %lf \n", k, i, neuron1[k * Nn + i], neuron2[k * Nn + i]); } } } return good; } int main() { // Initialize arrays on host VTYPE* h_neuron_i = (VTYPE*)malloc(Ni * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_neuron_n1 = (VTYPE*)malloc(Nn * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_synapse = (VTYPE*)malloc(Nn * Ni * sizeof(VTYPE)); VTYPE* h_neuron_n2 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n3 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); init_layer(h_neuron_i, h_neuron_n, h_synapse); // Allocate memory on device VTYPE* d_neuron_i = NULL; VTYPE* d_neuron_n1 = NULL; VTYPE* d_neuron_n2 = NULL; VTYPE* d_neuron_n3 = NULL; VTYPE* d_synapse = NULL; VTYPE* test_var = NULL; cudaMalloc((void**)&d_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_synapse, Nn * Ni * sizeof(VTYPE)); cudaMalloc((void**)&test_var, sizeof(VTYPE)); // Copy arrays from host to device cudaMemcpy(d_neuron_i, h_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyHostToDevice); cout << "Copy from Host to Device: " << cudaGetErrorString(cudaGetLastError()) << endl; cudaMemcpy(d_synapse, h_synapse, Nn * Ni * sizeof(VTYPE), cudaMemcpyHostToDevice); cout << "Copy from Host to Device: " << cudaGetErrorString(cudaGetLastError()) << endl; //Define kernel launch parameters dim3 ThreadsPerBlock2D = dim3(BlockSize2D, BlockSize2D); dim3 BlocksPerGrid2D = dim3((Nn + BlockSize2D - 1) / BlockSize2D, (BATCH_SIZE + BlockSize2D - 1) / BlockSize2D); //Launch kernel #1# d_MatMul_simple1<<<BlocksPerGrid2D, ThreadsPerBlock2D>>>(d_neuron_i, d_neuron_n1, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n1, d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; //Launch kernel #2# d_MatMul_simple2<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n2, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n2, d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; //Launch kernel #3# d_MatMul_simple3<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n3, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n3, d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; // Run and time on host clock_t begin = clock(); #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { VTYPE temp = 0.0f; #pragma unroll for (int j = 0; j < Ni; j++) { temp += h_neuron_i[k * Ni + j] * h_synapse[i + Nn * j]; } h_neuron_n[k * Nn + i] = temp; } /* * h_neuron_i 16 x 25088 * h_synapse 4096 x 25088 * h_neuron_n 16 x 4096 */ } double elapsed = ((double)clock() - (double)begin) / (double)CLOCKS_PER_SEC; printf("Took CPU %lf seconds to run\n", elapsed); /* VTYPE temp = 0.0f; int k = 1; int phase = 0; for (int i = phase * BlockSize2D; i < (phase + 1) * BlockSize2D; i++) { temp += h_synapse[k * Ni + i]; } printf("temp in host : %lf\n", temp); */ //Compare host and device results if (compare(h_neuron_n, h_neuron_n1)) { printf("1 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n2)) { printf("2 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n3)) { printf("3 Passed!\n"); } cout << "Host output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n[i]); } cout << endl; cout << "Kernel1 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n1[i]); } cout << endl; cout << "Kernel2 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n2[i]); } cout << endl; cout << "Kernel3 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n3[i]); } cout << endl; // Free up memory cudaFree(d_neuron_i); cudaFree(d_neuron_n1); cudaFree(d_neuron_n2); cudaFree(d_neuron_n3); cudaFree(d_synapse); cudaFree(test_var); free(h_neuron_i); free(h_neuron_n); free(h_synapse); free(h_neuron_n1); free(h_neuron_n2); free(h_neuron_n3); cout << "done\n"; return 0; }
.file "tmpxft_00080c7c_00000000-6_kernel.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3674: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3674: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z10init_layerPfS_S_ .type _Z10init_layerPfS_S_, @function _Z10init_layerPfS_S_: .LFB3669: .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 movq %rdx, %rbp movq %rsi, %rbx leaq 16384(%rsi), %r13 .L4: call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC0(%rip), %xmm0 movss %xmm0, (%rbx) addq $4, %rbx cmpq %r13, %rbx jne .L4 movq %r12, %rbx addq $1605632, %r12 .L5: call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC0(%rip), %xmm0 movss %xmm0, (%rbx) addq $4, %rbx cmpq %r12, %rbx jne .L5 movq %rbp, %rbx addq $411041792, %rbp .L6: call rand@PLT pxor %xmm0, %xmm0 cvtsi2ssl %eax, %xmm0 mulss .LC0(%rip), %xmm0 movss %xmm0, (%rbx) addq $4, %rbx cmpq %rbp, %rbx jne .L6 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 .LFE3669: .size _Z10init_layerPfS_S_, .-_Z10init_layerPfS_S_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC3: .string "At index (%d, %d) \t Host result: %lf \t Device result: %lf \n" .text .globl _Z7comparePfS_ .type _Z7comparePfS_, @function _Z7comparePfS_: .LFB3670: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $8, %rsp .cfi_def_cfa_offset 64 movq %rdi, %r12 movq %rsi, %rbp movl $0, %r13d movl $1, %eax leaq .LC3(%rip), %r15 movl $0, %r14d jmp .L12 .L13: addq $1, %rbx cmpq $4096, %rbx je .L20 .L15: movss (%r12,%rbx,4), %xmm0 movss 0(%rbp,%rbx,4), %xmm1 movaps %xmm0, %xmm2 subss %xmm1, %xmm2 andps .LC1(%rip), %xmm2 cvtss2sd %xmm2, %xmm2 comisd .LC2(%rip), %xmm2 jbe .L13 cvtss2sd %xmm0, %xmm0 cvtss2sd %xmm1, %xmm1 movl %ebx, %ecx movl %r13d, %edx movq %r15, %rsi movl $2, %edi movl $2, %eax call __printf_chk@PLT movl %r14d, %eax jmp .L13 .L20: addl $1, %r13d addq $16384, %r12 addq $16384, %rbp cmpl $16, %r13d je .L11 .L12: movl $0, %ebx jmp .L15 .L11: addq $8, %rsp .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3670: .size _Z7comparePfS_, .-_Z7comparePfS_ .globl _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_ .type _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_, @function _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_: .LFB3696: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L25 .L21: movq 120(%rsp), %rax subq %fs:40, %rax jne .L26 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L25: .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 _Z16d_MatMul_simple1PKfPfS0_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L21 .L26: call __stack_chk_fail@PLT .cfi_endproc .LFE3696: .size _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_, .-_Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_ .globl _Z16d_MatMul_simple1PKfPfS0_ .type _Z16d_MatMul_simple1PKfPfS0_, @function _Z16d_MatMul_simple1PKfPfS0_: .LFB3697: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3697: .size _Z16d_MatMul_simple1PKfPfS0_, .-_Z16d_MatMul_simple1PKfPfS0_ .globl _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_ .type _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_, @function _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_: .LFB3698: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L33 .L29: movq 120(%rsp), %rax subq %fs:40, %rax jne .L34 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L33: .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 _Z16d_MatMul_simple2PKfPfS0_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L29 .L34: call __stack_chk_fail@PLT .cfi_endproc .LFE3698: .size _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_, .-_Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_ .globl _Z16d_MatMul_simple2PKfPfS0_ .type _Z16d_MatMul_simple2PKfPfS0_, @function _Z16d_MatMul_simple2PKfPfS0_: .LFB3699: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3699: .size _Z16d_MatMul_simple2PKfPfS0_, .-_Z16d_MatMul_simple2PKfPfS0_ .globl _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_ .type _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_, @function _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_: .LFB3700: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L41 .L37: movq 120(%rsp), %rax subq %fs:40, %rax jne .L42 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L41: .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 _Z16d_MatMul_simple3PKfPfS0_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L37 .L42: call __stack_chk_fail@PLT .cfi_endproc .LFE3700: .size _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_, .-_Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_ .globl _Z16d_MatMul_simple3PKfPfS0_ .type _Z16d_MatMul_simple3PKfPfS0_, @function _Z16d_MatMul_simple3PKfPfS0_: .LFB3701: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3701: .size _Z16d_MatMul_simple3PKfPfS0_, .-_Z16d_MatMul_simple3PKfPfS0_ .section .rodata.str1.1,"aMS",@progbits,1 .LC5: .string "Copy from Host to Device: " .LC6: .string "MatMul_simple1: " .LC7: .string "Copy from Device to Host: " .LC9: .string "Took CPU %lf seconds to run\n" .LC10: .string "1 Passed!\n" .LC11: .string "2 Passed!\n" .LC12: .string "3 Passed!\n" .LC13: .string "Host output[0][6:9]: " .LC14: .string "%lf, " .LC15: .string "Kernel1 output[0][6:9]: " .LC16: .string "Kernel2 output[0][6:9]: " .LC17: .string "Kernel3 output[0][6:9]: " .LC18: .string "done\n" .text .globl main .type main, @function main: .LFB3671: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $104, %rsp .cfi_def_cfa_offset 160 movq %fs:40, %rax movq %rax, 88(%rsp) xorl %eax, %eax movl $1605632, %edi call malloc@PLT movq %rax, %r14 movl $262144, %edi call malloc@PLT movq %rax, %r13 movl $411041792, %edi call malloc@PLT movq %rax, %r15 movl $262144, %edi call malloc@PLT movq %rax, %r12 movl $262144, %edi call malloc@PLT movq %rax, %rbp movl $262144, %edi call malloc@PLT movq %rax, %rbx movq %r15, (%rsp) movq %r15, %rdx movq %rax, %rsi movq %r14, %rdi call _Z10init_layerPfS_S_ movq $0, 16(%rsp) movq $0, 24(%rsp) movq $0, 32(%rsp) movq $0, 40(%rsp) movq $0, 48(%rsp) movq $0, 56(%rsp) leaq 16(%rsp), %rdi movl $1605632, %esi call cudaMalloc@PLT leaq 24(%rsp), %rdi movl $262144, %esi call cudaMalloc@PLT leaq 32(%rsp), %rdi movl $262144, %esi call cudaMalloc@PLT leaq 40(%rsp), %rdi movl $262144, %esi call cudaMalloc@PLT leaq 48(%rsp), %rdi movl $411041792, %esi call cudaMalloc@PLT leaq 56(%rsp), %rdi movl $4, %esi call cudaMalloc@PLT movl $1, %ecx movl $1605632, %edx movq %r14, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT leaq .LC5(%rip), %r15 movq %r15, %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, 8(%rsp) call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq 8(%rsp), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $1, %ecx movl $411041792, %edx movq (%rsp), %rsi movq 48(%rsp), %rdi call cudaMemcpy@PLT movq %r15, %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %r15 call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $16, 64(%rsp) movl $16, 68(%rsp) movl $1, 72(%rsp) movl $256, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movl $1, %ecx movq 76(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L68 .L46: leaq .LC6(%rip), %rsi leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, 8(%rsp) call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq 8(%rsp), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $2, %ecx movl $262144, %edx movq 24(%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT leaq .LC7(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %r15 call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl 72(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movq 76(%rsp), %rdi movl 84(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L69 .L47: leaq .LC6(%rip), %rsi leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, 8(%rsp) call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq 8(%rsp), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $2, %ecx movl $262144, %edx movq 32(%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT leaq .LC7(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %r15 call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl 72(%rsp), %ecx movl $0, %r9d movl $0, %r8d movq 64(%rsp), %rdx movq 76(%rsp), %rdi movl 84(%rsp), %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L70 .L48: leaq .LC6(%rip), %rsi leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, 8(%rsp) call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq 8(%rsp), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movl $2, %ecx movl $262144, %edx movq 40(%rsp), %rsi movq %rbp, %rdi call cudaMemcpy@PLT leaq .LC7(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %r15 call cudaGetLastError@PLT movl %eax, %edi call cudaGetErrorString@PLT movq %rax, %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq %rax, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT call clock@PLT movq %rax, %r15 movl $0, %r9d .L49: movq (%rsp), %rax leaq 411041792(%rax), %rcx imulq $100352, %r9, %r8 addq %r14, %r8 movq %r9, %rdi salq $14, %rdi addq %rbx, %rdi movl $0, %esi .L53: leaq -411041792(%rcx), %rax movq %r8, %rdx pxor %xmm1, %xmm1 .L50: movss (%rdx), %xmm0 mulss (%rax), %xmm0 addss %xmm0, %xmm1 addq $4, %rdx addq $16384, %rax cmpq %rcx, %rax jne .L50 movss %xmm1, (%rdi,%rsi,4) addq $1, %rsi addq $4, %rcx cmpq $4096, %rsi jne .L53 addq $1, %r9 cmpq $16, %r9 jne .L49 call clock@PLT pxor %xmm0, %xmm0 cvtsi2sdq %rax, %xmm0 pxor %xmm1, %xmm1 cvtsi2sdq %r15, %xmm1 subsd %xmm1, %xmm0 divsd .LC8(%rip), %xmm0 leaq .LC9(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq %r13, %rsi movq %rbx, %rdi call _Z7comparePfS_ testb %al, %al jne .L71 .L54: movq %r12, %rsi movq %rbx, %rdi call _Z7comparePfS_ testb %al, %al jne .L72 .L55: movq %rbp, %rsi movq %rbx, %rdi call _Z7comparePfS_ testb %al, %al jne .L73 .L56: leaq .LC13(%rip), %rsi leaq _ZSt4cout(%rip), %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT leaq 24(%rbx), %r15 leaq 36(%rbx), %rax movq %rax, 8(%rsp) .L57: pxor %xmm0, %xmm0 cvtss2sd (%r15), %xmm0 leaq .LC14(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %r15 movq 8(%rsp), %rax cmpq %rax, %r15 jne .L57 leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT leaq .LC15(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT leaq 24(%r13), %r15 leaq 36(%r13), %rax movq %rax, 8(%rsp) .L58: pxor %xmm0, %xmm0 cvtss2sd (%r15), %xmm0 leaq .LC14(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %r15 movq 8(%rsp), %rax cmpq %rax, %r15 jne .L58 leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT leaq .LC16(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT leaq 24(%r12), %r15 leaq 36(%r12), %rax movq %rax, 8(%rsp) .L59: pxor %xmm0, %xmm0 cvtss2sd (%r15), %xmm0 leaq .LC14(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %r15 cmpq %r15, 8(%rsp) jne .L59 leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT leaq .LC17(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT leaq 24(%rbp), %r15 leaq 36(%rbp), %rax movq %rax, 8(%rsp) .L60: pxor %xmm0, %xmm0 cvtss2sd (%r15), %xmm0 leaq .LC14(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT addq $4, %r15 movq 8(%rsp), %rax cmpq %rax, %r15 jne .L60 leaq _ZSt4cout(%rip), %r15 movq %r15, %rdi call _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_@PLT movq 16(%rsp), %rdi call cudaFree@PLT movq 24(%rsp), %rdi call cudaFree@PLT movq 32(%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rdi call cudaFree@PLT movq 48(%rsp), %rdi call cudaFree@PLT movq 56(%rsp), %rdi call cudaFree@PLT movq %r14, %rdi call free@PLT movq %rbx, %rdi call free@PLT movq (%rsp), %rdi call free@PLT movq %r13, %rdi call free@PLT movq %r12, %rdi call free@PLT movq %rbp, %rdi call free@PLT leaq .LC18(%rip), %rsi movq %r15, %rdi call _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@PLT movq 88(%rsp), %rax subq %fs:40, %rax jne .L74 movl $0, %eax addq $104, %rsp .cfi_remember_state .cfi_def_cfa_offset 56 popq %rbx .cfi_def_cfa_offset 48 popq %rbp .cfi_def_cfa_offset 40 popq %r12 .cfi_def_cfa_offset 32 popq %r13 .cfi_def_cfa_offset 24 popq %r14 .cfi_def_cfa_offset 16 popq %r15 .cfi_def_cfa_offset 8 ret .L68: .cfi_restore_state movq 48(%rsp), %rdx movq 24(%rsp), %rsi movq 16(%rsp), %rdi call _Z42__device_stub__Z16d_MatMul_simple1PKfPfS0_PKfPfS0_ jmp .L46 .L69: movq 48(%rsp), %rdx movq 32(%rsp), %rsi movq 16(%rsp), %rdi call _Z42__device_stub__Z16d_MatMul_simple2PKfPfS0_PKfPfS0_ jmp .L47 .L70: movq 48(%rsp), %rdx movq 40(%rsp), %rsi movq 16(%rsp), %rdi call _Z42__device_stub__Z16d_MatMul_simple3PKfPfS0_PKfPfS0_ jmp .L48 .L71: leaq .LC10(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L54 .L72: leaq .LC11(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L55 .L73: leaq .LC12(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT jmp .L56 .L74: call __stack_chk_fail@PLT .cfi_endproc .LFE3671: .size main, .-main .globl _Z27__device_stub__Z6d_testPfS_PfS_ .type _Z27__device_stub__Z6d_testPfS_PfS_, @function _Z27__device_stub__Z6d_testPfS_PfS_: .LFB3702: .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 .L79 .L75: movq 104(%rsp), %rax subq %fs:40, %rax jne .L80 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L79: .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 _Z6d_testPfS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L75 .L80: call __stack_chk_fail@PLT .cfi_endproc .LFE3702: .size _Z27__device_stub__Z6d_testPfS_PfS_, .-_Z27__device_stub__Z6d_testPfS_PfS_ .globl _Z6d_testPfS_ .type _Z6d_testPfS_, @function _Z6d_testPfS_: .LFB3703: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z6d_testPfS_PfS_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3703: .size _Z6d_testPfS_, .-_Z6d_testPfS_ .section .rodata.str1.1 .LC19: .string "_Z6d_testPfS_" .LC20: .string "_Z16d_MatMul_simple3PKfPfS0_" .LC21: .string "_Z16d_MatMul_simple2PKfPfS0_" .LC22: .string "_Z16d_MatMul_simple1PKfPfS0_" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3705: .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 .LC19(%rip), %rdx movq %rdx, %rcx leaq _Z6d_testPfS_(%rip), %rsi movq %rax, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC20(%rip), %rdx movq %rdx, %rcx leaq _Z16d_MatMul_simple3PKfPfS0_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC21(%rip), %rdx movq %rdx, %rcx leaq _Z16d_MatMul_simple2PKfPfS0_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC22(%rip), %rdx movq %rdx, %rcx leaq _Z16d_MatMul_simple1PKfPfS0_(%rip), %rsi movq %rbx, %rdi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT popq %rbx .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3705: .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 805306368 .section .rodata.cst16,"aM",@progbits,16 .align 16 .LC1: .long 2147483647 .long 0 .long 0 .long 0 .section .rodata.cst8,"aM",@progbits,8 .align 8 .LC2: .long 1202590843 .long 1065646817 .align 8 .LC8: .long 0 .long 1093567616 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <cstdlib> #include <iostream> #include <stdio.h> #include <cmath> #include <cuda_runtime.h> #include <time.h> #include "device_launch_parameters.h" #include "cuda.h" using namespace std; #define Nn 4096 #define Ni 25088 #define BATCH_SIZE 16 #define BLOCK_SIZE 32 #define BlockSize2D 16 #define VTYPE float /* * synapse (w) is (Nn x Ni)^T * neuron_i (x) is (BATCH_SIZE x Ni) * neuron_n (y) is (BATCH_SIZE x Nn) * * y = Xw^T */ void init_layer(VTYPE* h_neuron_i, VTYPE* h_neuron_n, VTYPE* synapse) { for (int i = 0; i < Nn; i++) { h_neuron_n[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * BATCH_SIZE; i++) { h_neuron_i[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * Nn; i++) { synapse[i] = rand() / (VTYPE)RAND_MAX; } } __launch_bounds__(1024,2) __global__ void d_MatMul_simple1(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; VTYPE temp = 0.0f; if (col < Nn && row < BATCH_SIZE) { #pragma unroll for (int i = 0; i < Ni; i++) { temp += d_neuron_i[row * Ni + i] * synapse[col + Nn * i]; } d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple2(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE neuron_i[BlockSize2D][BlockSize2D]; VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < Ni; i += BlockSize2D) { if (row < BATCH_SIZE && i + threadIdx.x < Ni) { neuron_i[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i + threadIdx.x]; } else { neuron_i[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron_i[threadIdx.y][j] * d_synapse[(j + i)* Nn + col]; } __syncthreads(); } if (col < Nn && row < BATCH_SIZE) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple3(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE synapse[BlockSize2D][BlockSize2D]; __shared__ VTYPE neuron[BlockSize2D][BlockSize2D]; // MxK = MxN * NxK VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < (Ni - 1) / BlockSize2D + 1; i++) { if (row < BATCH_SIZE && i * BlockSize2D + threadIdx.x < Ni) { neuron[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i * BlockSize2D + threadIdx.x]; } else { neuron[threadIdx.y][threadIdx.x] = 0.0f; } if (i * BlockSize2D + threadIdx.y < Ni && col < Nn) { synapse[threadIdx.y][threadIdx.x] = d_synapse[(i * BlockSize2D + threadIdx.y) * Nn + col]; } else { synapse[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron[threadIdx.y][j] * synapse[j][threadIdx.x]; } __syncthreads(); } if (row < BATCH_SIZE && col < Nn) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_test(VTYPE* d_synapse, VTYPE* d_neuron_i) { int idx = blockIdx.x * blockDim.x + threadIdx.x; d_neuron_i[idx] *= 1.1f; } bool compare(VTYPE* neuron1, VTYPE* neuron2) { bool good = true; #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { if (fabs(neuron1[k * Nn + i] - neuron2[k * Nn + i]) > 1e-2) { good = false; printf("At index (%d, %d) \t Host result: %lf \t Device result: %lf \n", k, i, neuron1[k * Nn + i], neuron2[k * Nn + i]); } } } return good; } int main() { // Initialize arrays on host VTYPE* h_neuron_i = (VTYPE*)malloc(Ni * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_neuron_n1 = (VTYPE*)malloc(Nn * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_synapse = (VTYPE*)malloc(Nn * Ni * sizeof(VTYPE)); VTYPE* h_neuron_n2 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n3 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); init_layer(h_neuron_i, h_neuron_n, h_synapse); // Allocate memory on device VTYPE* d_neuron_i = NULL; VTYPE* d_neuron_n1 = NULL; VTYPE* d_neuron_n2 = NULL; VTYPE* d_neuron_n3 = NULL; VTYPE* d_synapse = NULL; VTYPE* test_var = NULL; cudaMalloc((void**)&d_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE)); cudaMalloc((void**)&d_synapse, Nn * Ni * sizeof(VTYPE)); cudaMalloc((void**)&test_var, sizeof(VTYPE)); // Copy arrays from host to device cudaMemcpy(d_neuron_i, h_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyHostToDevice); cout << "Copy from Host to Device: " << cudaGetErrorString(cudaGetLastError()) << endl; cudaMemcpy(d_synapse, h_synapse, Nn * Ni * sizeof(VTYPE), cudaMemcpyHostToDevice); cout << "Copy from Host to Device: " << cudaGetErrorString(cudaGetLastError()) << endl; //Define kernel launch parameters dim3 ThreadsPerBlock2D = dim3(BlockSize2D, BlockSize2D); dim3 BlocksPerGrid2D = dim3((Nn + BlockSize2D - 1) / BlockSize2D, (BATCH_SIZE + BlockSize2D - 1) / BlockSize2D); //Launch kernel #1# d_MatMul_simple1<<<BlocksPerGrid2D, ThreadsPerBlock2D>>>(d_neuron_i, d_neuron_n1, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n1, d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; //Launch kernel #2# d_MatMul_simple2<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n2, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n2, d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; //Launch kernel #3# d_MatMul_simple3<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n3, d_synapse); cout << "MatMul_simple1: " << cudaGetErrorString(cudaGetLastError()) << endl; // Copy results from device back to host cudaMemcpy(h_neuron_n3, d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE), cudaMemcpyDeviceToHost); cout << "Copy from Device to Host: " << cudaGetErrorString(cudaGetLastError()) << endl; // Run and time on host clock_t begin = clock(); #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { VTYPE temp = 0.0f; #pragma unroll for (int j = 0; j < Ni; j++) { temp += h_neuron_i[k * Ni + j] * h_synapse[i + Nn * j]; } h_neuron_n[k * Nn + i] = temp; } /* * h_neuron_i 16 x 25088 * h_synapse 4096 x 25088 * h_neuron_n 16 x 4096 */ } double elapsed = ((double)clock() - (double)begin) / (double)CLOCKS_PER_SEC; printf("Took CPU %lf seconds to run\n", elapsed); /* VTYPE temp = 0.0f; int k = 1; int phase = 0; for (int i = phase * BlockSize2D; i < (phase + 1) * BlockSize2D; i++) { temp += h_synapse[k * Ni + i]; } printf("temp in host : %lf\n", temp); */ //Compare host and device results if (compare(h_neuron_n, h_neuron_n1)) { printf("1 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n2)) { printf("2 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n3)) { printf("3 Passed!\n"); } cout << "Host output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n[i]); } cout << endl; cout << "Kernel1 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n1[i]); } cout << endl; cout << "Kernel2 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n2[i]); } cout << endl; cout << "Kernel3 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n3[i]); } cout << endl; // Free up memory cudaFree(d_neuron_i); cudaFree(d_neuron_n1); cudaFree(d_neuron_n2); cudaFree(d_neuron_n3); cudaFree(d_synapse); cudaFree(test_var); free(h_neuron_i); free(h_neuron_n); free(h_synapse); free(h_neuron_n1); free(h_neuron_n2); free(h_neuron_n3); cout << "done\n"; return 0; }
#include <cstdlib> #include <iostream> #include <stdio.h> #include <cmath> #include <hip/hip_runtime.h> #include <time.h> using namespace std; #define Nn 4096 #define Ni 25088 #define BATCH_SIZE 16 #define BLOCK_SIZE 32 #define BlockSize2D 16 #define VTYPE float /* * synapse (w) is (Nn x Ni)^T * neuron_i (x) is (BATCH_SIZE x Ni) * neuron_n (y) is (BATCH_SIZE x Nn) * * y = Xw^T */ void init_layer(VTYPE* h_neuron_i, VTYPE* h_neuron_n, VTYPE* synapse) { for (int i = 0; i < Nn; i++) { h_neuron_n[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * BATCH_SIZE; i++) { h_neuron_i[i] = rand() / (VTYPE)RAND_MAX; } for (int i = 0; i < Ni * Nn; i++) { synapse[i] = rand() / (VTYPE)RAND_MAX; } } __launch_bounds__(1024,2) __global__ void d_MatMul_simple1(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; VTYPE temp = 0.0f; if (col < Nn && row < BATCH_SIZE) { #pragma unroll for (int i = 0; i < Ni; i++) { temp += d_neuron_i[row * Ni + i] * synapse[col + Nn * i]; } d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple2(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE neuron_i[BlockSize2D][BlockSize2D]; VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < Ni; i += BlockSize2D) { if (row < BATCH_SIZE && i + threadIdx.x < Ni) { neuron_i[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i + threadIdx.x]; } else { neuron_i[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron_i[threadIdx.y][j] * d_synapse[(j + i)* Nn + col]; } __syncthreads(); } if (col < Nn && row < BATCH_SIZE) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_MatMul_simple3(const VTYPE* d_neuron_i, VTYPE* d_neuron_n, const VTYPE* d_synapse) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + threadIdx.y; __shared__ VTYPE synapse[BlockSize2D][BlockSize2D]; __shared__ VTYPE neuron[BlockSize2D][BlockSize2D]; // MxK = MxN * NxK VTYPE temp = 0.0f; #pragma unroll for (int i = 0; i < (Ni - 1) / BlockSize2D + 1; i++) { if (row < BATCH_SIZE && i * BlockSize2D + threadIdx.x < Ni) { neuron[threadIdx.y][threadIdx.x] = d_neuron_i[row * Ni + i * BlockSize2D + threadIdx.x]; } else { neuron[threadIdx.y][threadIdx.x] = 0.0f; } if (i * BlockSize2D + threadIdx.y < Ni && col < Nn) { synapse[threadIdx.y][threadIdx.x] = d_synapse[(i * BlockSize2D + threadIdx.y) * Nn + col]; } else { synapse[threadIdx.y][threadIdx.x] = 0.0f; } __syncthreads(); #pragma unroll for (int j = 0; j < BlockSize2D; j++) { temp += neuron[threadIdx.y][j] * synapse[j][threadIdx.x]; } __syncthreads(); } if (row < BATCH_SIZE && col < Nn) { d_neuron_n[row * Nn + col] = temp; } } __global__ void d_test(VTYPE* d_synapse, VTYPE* d_neuron_i) { int idx = blockIdx.x * blockDim.x + threadIdx.x; d_neuron_i[idx] *= 1.1f; } bool compare(VTYPE* neuron1, VTYPE* neuron2) { bool good = true; #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { if (fabs(neuron1[k * Nn + i] - neuron2[k * Nn + i]) > 1e-2) { good = false; printf("At index (%d, %d) \t Host result: %lf \t Device result: %lf \n", k, i, neuron1[k * Nn + i], neuron2[k * Nn + i]); } } } return good; } int main() { // Initialize arrays on host VTYPE* h_neuron_i = (VTYPE*)malloc(Ni * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_neuron_n1 = (VTYPE*)malloc(Nn * BATCH_SIZE *sizeof(VTYPE)); VTYPE* h_synapse = (VTYPE*)malloc(Nn * Ni * sizeof(VTYPE)); VTYPE* h_neuron_n2 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n3 = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); VTYPE* h_neuron_n = (VTYPE*)malloc(Nn * BATCH_SIZE * sizeof(VTYPE)); init_layer(h_neuron_i, h_neuron_n, h_synapse); // Allocate memory on device VTYPE* d_neuron_i = NULL; VTYPE* d_neuron_n1 = NULL; VTYPE* d_neuron_n2 = NULL; VTYPE* d_neuron_n3 = NULL; VTYPE* d_synapse = NULL; VTYPE* test_var = NULL; hipMalloc((void**)&d_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE)); hipMalloc((void**)&d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE)); hipMalloc((void**)&d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE)); hipMalloc((void**)&d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE)); hipMalloc((void**)&d_synapse, Nn * Ni * sizeof(VTYPE)); hipMalloc((void**)&test_var, sizeof(VTYPE)); // Copy arrays from host to device hipMemcpy(d_neuron_i, h_neuron_i, Ni * BATCH_SIZE * sizeof(VTYPE), hipMemcpyHostToDevice); cout << "Copy from Host to Device: " << hipGetErrorString(hipGetLastError()) << endl; hipMemcpy(d_synapse, h_synapse, Nn * Ni * sizeof(VTYPE), hipMemcpyHostToDevice); cout << "Copy from Host to Device: " << hipGetErrorString(hipGetLastError()) << endl; //Define kernel launch parameters dim3 ThreadsPerBlock2D = dim3(BlockSize2D, BlockSize2D); dim3 BlocksPerGrid2D = dim3((Nn + BlockSize2D - 1) / BlockSize2D, (BATCH_SIZE + BlockSize2D - 1) / BlockSize2D); //Launch kernel #1# d_MatMul_simple1<<<BlocksPerGrid2D, ThreadsPerBlock2D>>>(d_neuron_i, d_neuron_n1, d_synapse); cout << "MatMul_simple1: " << hipGetErrorString(hipGetLastError()) << endl; // Copy results from device back to host hipMemcpy(h_neuron_n1, d_neuron_n1, Nn * BATCH_SIZE * sizeof(VTYPE), hipMemcpyDeviceToHost); cout << "Copy from Device to Host: " << hipGetErrorString(hipGetLastError()) << endl; //Launch kernel #2# d_MatMul_simple2<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n2, d_synapse); cout << "MatMul_simple1: " << hipGetErrorString(hipGetLastError()) << endl; // Copy results from device back to host hipMemcpy(h_neuron_n2, d_neuron_n2, Nn * BATCH_SIZE * sizeof(VTYPE), hipMemcpyDeviceToHost); cout << "Copy from Device to Host: " << hipGetErrorString(hipGetLastError()) << endl; //Launch kernel #3# d_MatMul_simple3<<<BlocksPerGrid2D, ThreadsPerBlock2D >>>(d_neuron_i, d_neuron_n3, d_synapse); cout << "MatMul_simple1: " << hipGetErrorString(hipGetLastError()) << endl; // Copy results from device back to host hipMemcpy(h_neuron_n3, d_neuron_n3, Nn * BATCH_SIZE * sizeof(VTYPE), hipMemcpyDeviceToHost); cout << "Copy from Device to Host: " << hipGetErrorString(hipGetLastError()) << endl; // Run and time on host clock_t begin = clock(); #pragma unroll for (int k = 0; k < BATCH_SIZE; k++) { #pragma unroll for (int i = 0; i < Nn; i++) { VTYPE temp = 0.0f; #pragma unroll for (int j = 0; j < Ni; j++) { temp += h_neuron_i[k * Ni + j] * h_synapse[i + Nn * j]; } h_neuron_n[k * Nn + i] = temp; } /* * h_neuron_i 16 x 25088 * h_synapse 4096 x 25088 * h_neuron_n 16 x 4096 */ } double elapsed = ((double)clock() - (double)begin) / (double)CLOCKS_PER_SEC; printf("Took CPU %lf seconds to run\n", elapsed); /* VTYPE temp = 0.0f; int k = 1; int phase = 0; for (int i = phase * BlockSize2D; i < (phase + 1) * BlockSize2D; i++) { temp += h_synapse[k * Ni + i]; } printf("temp in host : %lf\n", temp); */ //Compare host and device results if (compare(h_neuron_n, h_neuron_n1)) { printf("1 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n2)) { printf("2 Passed!\n"); } if (compare(h_neuron_n, h_neuron_n3)) { printf("3 Passed!\n"); } cout << "Host output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n[i]); } cout << endl; cout << "Kernel1 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n1[i]); } cout << endl; cout << "Kernel2 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n2[i]); } cout << endl; cout << "Kernel3 output[0][6:9]: "; for (int i = 6; i < 9; i++) { printf("%lf, ", h_neuron_n3[i]); } cout << endl; // Free up memory hipFree(d_neuron_i); hipFree(d_neuron_n1); hipFree(d_neuron_n2); hipFree(d_neuron_n3); hipFree(d_synapse); hipFree(test_var); free(h_neuron_i); free(h_neuron_n); free(h_synapse); free(h_neuron_n1); free(h_neuron_n2); free(h_neuron_n3); cout << "done\n"; return 0; }
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
code for sm_80 Function : _Z16gpu_seqwr_kernelPimm .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*/ ISETP.NE.U32.AND P0, PT, RZ, c[0x0][0x168], PT ; /* 0x00005a00ff007a0c */ /* 0x000fc80003f05070 */ /*0020*/ ISETP.NE.AND.EX P0, PT, RZ, c[0x0][0x16c], PT, P0 ; /* 0x00005b00ff007a0c */ /* 0x000fda0003f05300 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e220000002500 */ /*0050*/ IMAD.MOV.U32 R9, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff097624 */ /* 0x000fe200078e00ff */ /*0060*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0070*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fe200078e00ff */ /*0080*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0090*/ IMAD.MOV.U32 R8, RZ, RZ, RZ ; /* 0x000000ffff087224 */ /* 0x000fe400078e00ff */ /*00a0*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*00b0*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe20003f06070 */ /*00c0*/ BSSY B0, 0x1f0 ; /* 0x0000012000007945 */ /* 0x000fe20003800000 */ /*00d0*/ IADD3 R6, P1, R6, 0x1, RZ ; /* 0x0000000106067810 */ /* 0x000fe40007f3e0ff */ /*00e0*/ ISETP.GE.U32.AND.EX P0, PT, RZ, c[0x0][0x174], PT, P0 ; /* 0x00005d00ff007a0c */ /* 0x000fc60003f06100 */ /*00f0*/ IMAD.X R8, RZ, RZ, R8, P1 ; /* 0x000000ffff087224 */ /* 0x000fe200008e0608 */ /*0100*/ ISETP.GE.U32.AND P1, PT, R6, c[0x0][0x168], PT ; /* 0x00005a0006007a0c */ /* 0x000fc80003f26070 */ /*0110*/ ISETP.GE.U32.AND.EX P1, PT, R8, c[0x0][0x16c], PT, P1 ; /* 0x00005b0008007a0c */ /* 0x000fca0003f26110 */ /*0120*/ @P0 BRA 0x1e0 ; /* 0x000000b000000947 */ /* 0x000fea0003800000 */ /*0130*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0000 */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc600078e00ff */ /*0150*/ LEA R2, P0, R4, c[0x0][0x160], 0x2 ; /* 0x0000580004027a11 */ /* 0x000fe200078010ff */ /*0160*/ IMAD R7, R9, c[0x0][0xc], RZ ; /* 0x0000030009077a24 */ /* 0x000fc600078e02ff */ /*0170*/ LEA.HI.X R3, R4, c[0x0][0x164], R5, 0x2, P0 ; /* 0x0000590004037a11 */ /* 0x000fe400000f1405 */ /*0180*/ IADD3 R4, P0, R7, R4, RZ ; /* 0x0000000407047210 */ /* 0x000fc60007f1e0ff */ /*0190*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x0001e4000c101904 */ /*01a0*/ IMAD.X R5, RZ, RZ, R5, P0 ; /* 0x000000ffff057224 */ /* 0x000fe200000e0605 */ /*01b0*/ ISETP.GE.U32.AND P0, PT, R4, c[0x0][0x170], PT ; /* 0x00005c0004007a0c */ /* 0x000fc80003f06070 */ /*01c0*/ ISETP.GE.U32.AND.EX P0, PT, R5, c[0x0][0x174], PT, P0 ; /* 0x00005d0005007a0c */ /* 0x000fda0003f06100 */ /*01d0*/ @!P0 BRA 0x150 ; /* 0xffffff7000008947 */ /* 0x001fea000383ffff */ /*01e0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*01f0*/ @!P1 BRA 0xb0 ; /* 0xfffffeb000009947 */ /* 0x000fea000383ffff */ /*0200*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0210*/ BRA 0x210; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
.file "tmpxft_00017ee3_00000000-6_gpu_seqwr_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 _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm .type _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm, @function _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z16gpu_seqwr_kernelPimm(%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 _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm, .-_Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm .globl _Z16gpu_seqwr_kernelPimm .type _Z16gpu_seqwr_kernelPimm, @function _Z16gpu_seqwr_kernelPimm: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z16gpu_seqwr_kernelPimm, .-_Z16gpu_seqwr_kernelPimm .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z16gpu_seqwr_kernelPimm" .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 _Z16gpu_seqwr_kernelPimm(%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 gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
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 gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z16gpu_seqwr_kernelPimm .globl _Z16gpu_seqwr_kernelPimm .p2align 8 .type _Z16gpu_seqwr_kernelPimm,@function _Z16gpu_seqwr_kernelPimm: s_load_b64 s[2:3], s[0:1], 0x8 s_mov_b64 s[4:5], 0 s_waitcnt lgkmcnt(0) s_cmp_eq_u64 s[2:3], 0 s_cbranch_scc1 .LBB0_6 s_clause 0x3 s_load_b32 s8, s[0:1], 0x24 s_load_b64 s[6:7], s[0:1], 0x10 s_load_b64 s[10:11], s[0:1], 0x0 s_load_b32 s1, s[0:1], 0x18 s_mov_b32 s9, 0 s_waitcnt lgkmcnt(0) s_and_b32 s8, s8, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_mad_u64_u32 v[1:2], null, s15, s8, v[0:1] v_mov_b32_e32 v2, 0 v_mov_b32_e32 v0, 0 s_mul_i32 s8, s1, s8 v_lshlrev_b64 v[3:4], 2, v[1:2] v_cmp_gt_u64_e32 vcc_lo, s[6:7], v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v3, s0, s10, v3 v_add_co_ci_u32_e64 v4, s0, s11, v4, s0 s_lshl_b64 s[10:11], s[8:9], 2 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_or_b32 exec_lo, exec_lo, s12 s_add_u32 s4, s4, 1 s_addc_u32 s5, s5, 0 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u64 s[4:5], s[2:3] s_cbranch_scc1 .LBB0_6 .LBB0_3: s_and_saveexec_b32 s12, vcc_lo s_cbranch_execz .LBB0_2 v_dual_mov_b32 v6, v4 :: v_dual_mov_b32 v5, v3 v_dual_mov_b32 v8, v2 :: v_dual_mov_b32 v7, v1 s_mov_b32 s13, 0 .LBB0_5: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v7, s0, v7, s8 v_add_co_ci_u32_e64 v8, s0, s9, v8, s0 global_store_b32 v[5:6], v0, off v_add_co_u32 v5, s1, v5, s10 v_cmp_le_u64_e64 s0, s[6:7], v[7:8] v_add_co_ci_u32_e64 v6, s1, s11, v6, s1 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_or_b32 s13, s0, s13 s_and_not1_b32 exec_lo, exec_lo, s13 s_cbranch_execnz .LBB0_5 s_branch .LBB0_2 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16gpu_seqwr_kernelPimm .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 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 _Z16gpu_seqwr_kernelPimm, .Lfunc_end0-_Z16gpu_seqwr_kernelPimm .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: 8 .value_kind: by_value - .offset: 16 .size: 8 .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: _Z16gpu_seqwr_kernelPimm .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z16gpu_seqwr_kernelPimm.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" __global__ void gpu_seqwr_kernel(int *buffer, size_t reps, size_t elements) { for(size_t j = 0; j < reps; j++) { size_t ofs = blockIdx.x * blockDim.x + threadIdx.x; size_t step = blockDim.x * gridDim.x; while(ofs < elements) { buffer[ofs] = 0; ofs += step; } } }
.text .file "gpu_seqwr_kernel.hip" .globl _Z31__device_stub__gpu_seqwr_kernelPimm # -- Begin function _Z31__device_stub__gpu_seqwr_kernelPimm .p2align 4, 0x90 .type _Z31__device_stub__gpu_seqwr_kernelPimm,@function _Z31__device_stub__gpu_seqwr_kernelPimm: # @_Z31__device_stub__gpu_seqwr_kernelPimm .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z16gpu_seqwr_kernelPimm, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z31__device_stub__gpu_seqwr_kernelPimm, .Lfunc_end0-_Z31__device_stub__gpu_seqwr_kernelPimm .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 $_Z16gpu_seqwr_kernelPimm, %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 _Z16gpu_seqwr_kernelPimm,@object # @_Z16gpu_seqwr_kernelPimm .section .rodata,"a",@progbits .globl _Z16gpu_seqwr_kernelPimm .p2align 3, 0x0 _Z16gpu_seqwr_kernelPimm: .quad _Z31__device_stub__gpu_seqwr_kernelPimm .size _Z16gpu_seqwr_kernelPimm, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z16gpu_seqwr_kernelPimm" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__gpu_seqwr_kernelPimm .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16gpu_seqwr_kernelPimm .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 : _Z16gpu_seqwr_kernelPimm .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*/ ISETP.NE.U32.AND P0, PT, RZ, c[0x0][0x168], PT ; /* 0x00005a00ff007a0c */ /* 0x000fc80003f05070 */ /*0020*/ ISETP.NE.AND.EX P0, PT, RZ, c[0x0][0x16c], PT, P0 ; /* 0x00005b00ff007a0c */ /* 0x000fda0003f05300 */ /*0030*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*0040*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e220000002500 */ /*0050*/ IMAD.MOV.U32 R9, RZ, RZ, c[0x0][0x0] ; /* 0x00000000ff097624 */ /* 0x000fe200078e00ff */ /*0060*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0070*/ IMAD.MOV.U32 R6, RZ, RZ, RZ ; /* 0x000000ffff067224 */ /* 0x000fe200078e00ff */ /*0080*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0090*/ IMAD.MOV.U32 R8, RZ, RZ, RZ ; /* 0x000000ffff087224 */ /* 0x000fe400078e00ff */ /*00a0*/ IMAD R0, R0, c[0x0][0x0], R3 ; /* 0x0000000000007a24 */ /* 0x001fca00078e0203 */ /*00b0*/ ISETP.GE.U32.AND P0, PT, R0, c[0x0][0x170], PT ; /* 0x00005c0000007a0c */ /* 0x000fe20003f06070 */ /*00c0*/ BSSY B0, 0x1f0 ; /* 0x0000012000007945 */ /* 0x000fe20003800000 */ /*00d0*/ IADD3 R6, P1, R6, 0x1, RZ ; /* 0x0000000106067810 */ /* 0x000fe40007f3e0ff */ /*00e0*/ ISETP.GE.U32.AND.EX P0, PT, RZ, c[0x0][0x174], PT, P0 ; /* 0x00005d00ff007a0c */ /* 0x000fc60003f06100 */ /*00f0*/ IMAD.X R8, RZ, RZ, R8, P1 ; /* 0x000000ffff087224 */ /* 0x000fe200008e0608 */ /*0100*/ ISETP.GE.U32.AND P1, PT, R6, c[0x0][0x168], PT ; /* 0x00005a0006007a0c */ /* 0x000fc80003f26070 */ /*0110*/ ISETP.GE.U32.AND.EX P1, PT, R8, c[0x0][0x16c], PT, P1 ; /* 0x00005b0008007a0c */ /* 0x000fca0003f26110 */ /*0120*/ @P0 BRA 0x1e0 ; /* 0x000000b000000947 */ /* 0x000fea0003800000 */ /*0130*/ IMAD.MOV.U32 R4, RZ, RZ, R0 ; /* 0x000000ffff047224 */ /* 0x000fe400078e0000 */ /*0140*/ IMAD.MOV.U32 R5, RZ, RZ, RZ ; /* 0x000000ffff057224 */ /* 0x000fc600078e00ff */ /*0150*/ LEA R2, P0, R4, c[0x0][0x160], 0x2 ; /* 0x0000580004027a11 */ /* 0x000fe200078010ff */ /*0160*/ IMAD R7, R9, c[0x0][0xc], RZ ; /* 0x0000030009077a24 */ /* 0x000fc600078e02ff */ /*0170*/ LEA.HI.X R3, R4, c[0x0][0x164], R5, 0x2, P0 ; /* 0x0000590004037a11 */ /* 0x000fe400000f1405 */ /*0180*/ IADD3 R4, P0, R7, R4, RZ ; /* 0x0000000407047210 */ /* 0x000fc60007f1e0ff */ /*0190*/ STG.E [R2.64], RZ ; /* 0x000000ff02007986 */ /* 0x0001e4000c101904 */ /*01a0*/ IMAD.X R5, RZ, RZ, R5, P0 ; /* 0x000000ffff057224 */ /* 0x000fe200000e0605 */ /*01b0*/ ISETP.GE.U32.AND P0, PT, R4, c[0x0][0x170], PT ; /* 0x00005c0004007a0c */ /* 0x000fc80003f06070 */ /*01c0*/ ISETP.GE.U32.AND.EX P0, PT, R5, c[0x0][0x174], PT, P0 ; /* 0x00005d0005007a0c */ /* 0x000fda0003f06100 */ /*01d0*/ @!P0 BRA 0x150 ; /* 0xffffff7000008947 */ /* 0x001fea000383ffff */ /*01e0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*01f0*/ @!P1 BRA 0xb0 ; /* 0xfffffeb000009947 */ /* 0x000fea000383ffff */ /*0200*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0210*/ BRA 0x210; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*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 _Z16gpu_seqwr_kernelPimm .globl _Z16gpu_seqwr_kernelPimm .p2align 8 .type _Z16gpu_seqwr_kernelPimm,@function _Z16gpu_seqwr_kernelPimm: s_load_b64 s[2:3], s[0:1], 0x8 s_mov_b64 s[4:5], 0 s_waitcnt lgkmcnt(0) s_cmp_eq_u64 s[2:3], 0 s_cbranch_scc1 .LBB0_6 s_clause 0x3 s_load_b32 s8, s[0:1], 0x24 s_load_b64 s[6:7], s[0:1], 0x10 s_load_b64 s[10:11], s[0:1], 0x0 s_load_b32 s1, s[0:1], 0x18 s_mov_b32 s9, 0 s_waitcnt lgkmcnt(0) s_and_b32 s8, s8, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_2) v_mad_u64_u32 v[1:2], null, s15, s8, v[0:1] v_mov_b32_e32 v2, 0 v_mov_b32_e32 v0, 0 s_mul_i32 s8, s1, s8 v_lshlrev_b64 v[3:4], 2, v[1:2] v_cmp_gt_u64_e32 vcc_lo, s[6:7], v[1:2] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v3, s0, s10, v3 v_add_co_ci_u32_e64 v4, s0, s11, v4, s0 s_lshl_b64 s[10:11], s[8:9], 2 s_set_inst_prefetch_distance 0x1 s_branch .LBB0_3 .p2align 6 .LBB0_2: s_or_b32 exec_lo, exec_lo, s12 s_add_u32 s4, s4, 1 s_addc_u32 s5, s5, 0 s_delay_alu instid0(SALU_CYCLE_1) s_cmp_eq_u64 s[4:5], s[2:3] s_cbranch_scc1 .LBB0_6 .LBB0_3: s_and_saveexec_b32 s12, vcc_lo s_cbranch_execz .LBB0_2 v_dual_mov_b32 v6, v4 :: v_dual_mov_b32 v5, v3 v_dual_mov_b32 v8, v2 :: v_dual_mov_b32 v7, v1 s_mov_b32 s13, 0 .LBB0_5: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v7, s0, v7, s8 v_add_co_ci_u32_e64 v8, s0, s9, v8, s0 global_store_b32 v[5:6], v0, off v_add_co_u32 v5, s1, v5, s10 v_cmp_le_u64_e64 s0, s[6:7], v[7:8] v_add_co_ci_u32_e64 v6, s1, s11, v6, s1 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_or_b32 s13, s0, s13 s_and_not1_b32 exec_lo, exec_lo, s13 s_cbranch_execnz .LBB0_5 s_branch .LBB0_2 .LBB0_6: s_set_inst_prefetch_distance 0x2 s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z16gpu_seqwr_kernelPimm .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 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 _Z16gpu_seqwr_kernelPimm, .Lfunc_end0-_Z16gpu_seqwr_kernelPimm .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: 8 .value_kind: by_value - .offset: 16 .size: 8 .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: _Z16gpu_seqwr_kernelPimm .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z16gpu_seqwr_kernelPimm.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_00017ee3_00000000-6_gpu_seqwr_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 _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm .type _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm, @function _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm: .LFB2051: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z16gpu_seqwr_kernelPimm(%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 _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm, .-_Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm .globl _Z16gpu_seqwr_kernelPimm .type _Z16gpu_seqwr_kernelPimm, @function _Z16gpu_seqwr_kernelPimm: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z38__device_stub__Z16gpu_seqwr_kernelPimmPimm addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z16gpu_seqwr_kernelPimm, .-_Z16gpu_seqwr_kernelPimm .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z16gpu_seqwr_kernelPimm" .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 _Z16gpu_seqwr_kernelPimm(%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 "gpu_seqwr_kernel.hip" .globl _Z31__device_stub__gpu_seqwr_kernelPimm # -- Begin function _Z31__device_stub__gpu_seqwr_kernelPimm .p2align 4, 0x90 .type _Z31__device_stub__gpu_seqwr_kernelPimm,@function _Z31__device_stub__gpu_seqwr_kernelPimm: # @_Z31__device_stub__gpu_seqwr_kernelPimm .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z16gpu_seqwr_kernelPimm, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z31__device_stub__gpu_seqwr_kernelPimm, .Lfunc_end0-_Z31__device_stub__gpu_seqwr_kernelPimm .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 $_Z16gpu_seqwr_kernelPimm, %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 _Z16gpu_seqwr_kernelPimm,@object # @_Z16gpu_seqwr_kernelPimm .section .rodata,"a",@progbits .globl _Z16gpu_seqwr_kernelPimm .p2align 3, 0x0 _Z16gpu_seqwr_kernelPimm: .quad _Z31__device_stub__gpu_seqwr_kernelPimm .size _Z16gpu_seqwr_kernelPimm, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z16gpu_seqwr_kernelPimm" .size .L__unnamed_1, 25 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z31__device_stub__gpu_seqwr_kernelPimm .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z16gpu_seqwr_kernelPimm .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
code for sm_80 Function : _Z6matmulPfS_S_ii .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ ISETP.LT.AND P0, PT, RZ, c[0x0][0x178], PT ; /* 0x00005e00ff007a0c */ /* 0x000fe20003f01270 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd80000000a00 */ /*0030*/ @!P0 MOV R26, RZ ; /* 0x000000ff001a8202 */ /* 0x000fe20000000f00 */ /*0040*/ @!P0 BRA 0x5e0 ; /* 0x0000059000008947 */ /* 0x000fea0003800000 */ /*0050*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0060*/ MOV R26, RZ ; /* 0x000000ff001a7202 */ /* 0x000fc60000000f00 */ /*0070*/ S2R R0, SR_TID.Y ; /* 0x0000000000007919 */ /* 0x000e680000002200 */ /*0080*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000ea80000002500 */ /*0090*/ S2R R4, SR_CTAID.Y ; /* 0x0000000000047919 */ /* 0x000ee20000002600 */ /*00a0*/ ISETP.GT.U32.AND P0, PT, R3, 0xf, PT ; /* 0x0000000f0300780c */ /* 0x001fe40003f04070 */ /*00b0*/ LEA R17, R0, R3.reuse, 0xb ; /* 0x0000000300117211 */ /* 0x082fe200078e58ff */ /*00c0*/ HFMA2.MMA R0, -RZ, RZ, 0, 0 ; /* 0x00000000ff007435 */ /* 0x000fe200000001ff */ /*00d0*/ LEA R2, R2, R3, 0x8 ; /* 0x0000000302027211 */ /* 0x004fc400078e40ff */ /*00e0*/ LEA R17, R4, R17, 0xb ; /* 0x0000001104117211 */ /* 0x008fe400078e58ff */ /*00f0*/ IADD3 R16, R2, 0x7800, RZ ; /* 0x0000780002107810 */ /* 0x000fca0007ffe0ff */ /*0100*/ @!P0 IADD3 R6, R17, R0, RZ ; /* 0x0000000011068210 */ /* 0x000fe40007ffe0ff */ /*0110*/ MOV R27, 0x4 ; /* 0x00000004001b7802 */ /* 0x000fca0000000f00 */ /*0120*/ @!P0 IMAD.WIDE.U32 R6, R6, R27, c[0x0][0x160] ; /* 0x0000580006068625 */ /* 0x000fcc00078e001b */ /*0130*/ @!P0 LDG.E R6, [R6.64] ; /* 0x0000000406068981 */ /* 0x000ea2000c1e1900 */ /*0140*/ IMAD.WIDE.U32 R8, R2.reuse, R27, c[0x0][0x168] ; /* 0x00005a0002087625 */ /* 0x040fe200078e001b */ /*0150*/ IADD3 R12, R2.reuse, 0x800, RZ ; /* 0x00000800020c7810 */ /* 0x040fe40007ffe0ff */ /*0160*/ IADD3 R18, R2, 0x1000, RZ ; /* 0x0000100002127810 */ /* 0x000fc60007ffe0ff */ /*0170*/ IMAD.WIDE.U32 R12, R12, R27, c[0x0][0x168] ; /* 0x00005a000c0c7625 */ /* 0x000fc800078e001b */ /*0180*/ IMAD.WIDE.U32 R18, R18, R27, c[0x0][0x168] ; /* 0x00005a0012127625 */ /* 0x000fe200078e001b */ /*0190*/ IADD3 R10, R2.reuse, 0x1800, RZ ; /* 0x00001800020a7810 */ /* 0x040fe40007ffe0ff */ /*01a0*/ IADD3 R24, R2, 0x2000, RZ ; /* 0x0000200002187810 */ /* 0x000fca0007ffe0ff */ /*01b0*/ IMAD.WIDE.U32 R24, R24, R27.reuse, c[0x0][0x168] ; /* 0x00005a0018187625 */ /* 0x080fe200078e001b */ /*01c0*/ IADD3 R20, R2, 0x2800, RZ ; /* 0x0000280002147810 */ /* 0x000fe20007ffe0ff */ /*01d0*/ @!P0 STS [R3.X4], R6 ; /* 0x0000000603008388 */ /* 0x0041e80000004800 */ /*01e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*01f0*/ LDG.E R21, [R8.64] ; /* 0x0000000408157981 */ /* 0x0002a8000c1e1900 */ /*0200*/ LDG.E R4, [R12.64] ; /* 0x000000040c047981 */ /* 0x000728000c1e1900 */ /*0210*/ LDG.E R5, [R18.64] ; /* 0x0000000412057981 */ /* 0x000b22000c1e1900 */ /*0220*/ IMAD.WIDE.U32 R6, R10, R27, c[0x0][0x168] ; /* 0x00005a000a067625 */ /* 0x001fc600078e001b */ /*0230*/ LDG.E R15, [R24.64] ; /* 0x00000004180f7981 */ /* 0x000128000c1e1900 */ /*0240*/ LDG.E R14, [R6.64] ; /* 0x00000004060e7981 */ /* 0x000122000c1e1900 */ /*0250*/ IADD3 R22, R2, 0x3000, RZ ; /* 0x0000300002167810 */ /* 0x000fe20007ffe0ff */ /*0260*/ IMAD.WIDE.U32 R12, R20, R27.reuse, c[0x0][0x168] ; /* 0x00005a00140c7625 */ /* 0x088fe400078e001b */ /*0270*/ LDS.128 R8, [RZ] ; /* 0x00000000ff087984 */ /* 0x002ea40000000c00 */ /*0280*/ IMAD.WIDE.U32 R22, R22, R27, c[0x0][0x168] ; /* 0x00005a0016167625 */ /* 0x000fc400078e001b */ /*0290*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x0002e8000c1e1900 */ /*02a0*/ LDG.E R13, [R22.64] ; /* 0x00000004160d7981 */ /* 0x0022e2000c1e1900 */ /*02b0*/ IADD3 R28, R2, 0x3800, RZ ; /* 0x00003800021c7810 */ /* 0x000fe40007ffe0ff */ /*02c0*/ IADD3 R18, R16, -0x3800, RZ ; /* 0xffffc80010127810 */ /* 0x020fc60007ffe0ff */ /*02d0*/ IMAD.WIDE.U32 R28, R28, R27, c[0x0][0x168] ; /* 0x00005a001c1c7625 */ /* 0x000fe200078e001b */ /*02e0*/ IADD3 R24, R16, -0x3000, RZ ; /* 0xffffd00010187810 */ /* 0x001fc60007ffe0ff */ /*02f0*/ IMAD.WIDE.U32 R6, R18, R27.reuse, c[0x0][0x168] ; /* 0x00005a0012067625 */ /* 0x080fe400078e001b */ /*0300*/ LDG.E R18, [R28.64] ; /* 0x000000041c127981 */ /* 0x000162000c1e1900 */ /*0310*/ IADD3 R22, R16, -0x2800, RZ ; /* 0xffffd80010167810 */ /* 0x002fe20007ffe0ff */ /*0320*/ IMAD.WIDE.U32 R24, R24, R27.reuse, c[0x0][0x168] ; /* 0x00005a0018187625 */ /* 0x080fe400078e001b */ /*0330*/ LDG.E R19, [R6.64] ; /* 0x0000000406137981 */ /* 0x0002e4000c1e1900 */ /*0340*/ IMAD.WIDE.U32 R22, R22, R27, c[0x0][0x168] ; /* 0x00005a0016167625 */ /* 0x000fe400078e001b */ /*0350*/ LDG.E R20, [R24.64] ; /* 0x0000000418147981 */ /* 0x0004e2000c1e1900 */ /*0360*/ IADD3 R28, R16, -0x1800, RZ ; /* 0xffffe800101c7810 */ /* 0x001fc40007ffe0ff */ /*0370*/ IADD3 R6, R16, -0x2000, RZ ; /* 0xffffe00010067810 */ /* 0x002fc60007ffe0ff */ /*0380*/ IMAD.WIDE.U32 R28, R28, R27, c[0x0][0x168] ; /* 0x00005a001c1c7625 */ /* 0x000fc800078e001b */ /*0390*/ IMAD.WIDE.U32 R6, R6, R27, c[0x0][0x168] ; /* 0x00005a0006067625 */ /* 0x000fc800078e001b */ /*03a0*/ FFMA R8, R21, R8, R26 ; /* 0x0000000815087223 */ /* 0x004fe2000000001a */ /*03b0*/ IADD3 R26, R16, -0x1000, RZ ; /* 0xfffff000101a7810 */ /* 0x000fe20007ffe0ff */ /*03c0*/ LDG.E R21, [R22.64] ; /* 0x0000000416157981 */ /* 0x0000a8000c1e1900 */ /*03d0*/ IMAD.WIDE.U32 R24, R26, R27, c[0x0][0x168] ; /* 0x00005a001a187625 */ /* 0x000fe200078e001b */ /*03e0*/ LDG.E R22, [R6.64] ; /* 0x0000000406167981 */ /* 0x0010a8000c1e1900 */ /*03f0*/ LDG.E R23, [R28.64] ; /* 0x000000041c177981 */ /* 0x0002a8000c1e1900 */ /*0400*/ LDG.E R24, [R24.64] ; /* 0x0000000418187981 */ /* 0x0008a2000c1e1900 */ /*0410*/ IADD3 R6, R16, -0x800, RZ ; /* 0xfffff80010067810 */ /* 0x001fca0007ffe0ff */ /*0420*/ IMAD.WIDE.U32 R28, R6, R27, c[0x0][0x168] ; /* 0x00005a00061c7625 */ /* 0x002fc800078e001b */ /*0430*/ IMAD.WIDE.U32 R26, R16, R27, c[0x0][0x168] ; /* 0x00005a00101a7625 */ /* 0x000fe200078e001b */ /*0440*/ LDG.E R25, [R28.64] ; /* 0x000000041c197981 */ /* 0x010f2a000c1e1900 */ /*0450*/ LDG.E R26, [R26.64] ; /* 0x000000041a1a7981 */ /* 0x000f22000c1e1900 */ /*0460*/ FFMA R4, R4, R9, R8 ; /* 0x0000000904047223 */ /* 0x000fc80000000008 */ /*0470*/ FFMA R10, R5, R10, R4 ; /* 0x0000000a050a7223 */ /* 0x000fe40000000004 */ /*0480*/ LDS.128 R4, [0x10] ; /* 0x00001000ff047984 */ /* 0x000e240000000c00 */ /*0490*/ FFMA R10, R14, R11, R10 ; /* 0x0000000b0e0a7223 */ /* 0x000fc8000000000a */ /*04a0*/ FFMA R4, R15, R4, R10 ; /* 0x000000040f047223 */ /* 0x001fe4000000000a */ /*04b0*/ LDS.128 R8, [0x20] ; /* 0x00002000ff087984 */ /* 0x000e240000000c00 */ /*04c0*/ FFMA R4, R12, R5, R4 ; /* 0x000000050c047223 */ /* 0x008fc80000000004 */ /*04d0*/ FFMA R4, R13, R6, R4 ; /* 0x000000060d047223 */ /* 0x000fe40000000004 */ /*04e0*/ LDS.128 R12, [0x30] ; /* 0x00003000ff0c7984 */ /* 0x000e640000000c00 */ /*04f0*/ FFMA R4, R18, R7, R4 ; /* 0x0000000712047223 */ /* 0x020fe20000000004 */ /*0500*/ IADD3 R0, R0, 0x10, RZ ; /* 0x0000001000007810 */ /* 0x000fe20007ffe0ff */ /*0510*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe60000010000 */ /*0520*/ ISETP.GE.AND P1, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe20003f26270 */ /*0530*/ FFMA R4, R19, R8, R4 ; /* 0x0000000813047223 */ /* 0x001fc80000000004 */ /*0540*/ FFMA R4, R20, R9, R4 ; /* 0x0000000914047223 */ /* 0x000fe20000000004 */ /*0550*/ IADD3 R2, R2, 0x8000, RZ ; /* 0x0000800002027810 */ /* 0x000fe40007ffe0ff */ /*0560*/ IADD3 R16, R16, 0x8000, RZ ; /* 0x0000800010107810 */ /* 0x000fe20007ffe0ff */ /*0570*/ FFMA R4, R21, R10, R4 ; /* 0x0000000a15047223 */ /* 0x004fc80000000004 */ /*0580*/ FFMA R4, R22, R11, R4 ; /* 0x0000000b16047223 */ /* 0x000fc80000000004 */ /*0590*/ FFMA R4, R23, R12, R4 ; /* 0x0000000c17047223 */ /* 0x002fc80000000004 */ /*05a0*/ FFMA R4, R24, R13, R4 ; /* 0x0000000d18047223 */ /* 0x000fc80000000004 */ /*05b0*/ FFMA R4, R25, R14, R4 ; /* 0x0000000e19047223 */ /* 0x010fc80000000004 */ /*05c0*/ FFMA R26, R26, R15, R4 ; /* 0x0000000f1a1a7223 */ /* 0x000fe20000000004 */ /*05d0*/ @!P1 BRA 0x100 ; /* 0xfffffb2000009947 */ /* 0x000fea000383ffff */ /*05e0*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e220000002500 */ /*05f0*/ S2UR UR6, SR_CTAID.Y ; /* 0x00000000000679c3 */ /* 0x000e620000002600 */ /*0600*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe400000001ff */ /*0610*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0620*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0630*/ LEA R3, R0, R3, 0x8 ; /* 0x0000000300037211 */ /* 0x001fc400078e40ff */ /*0640*/ IADD3 R0, R2, UR6, RZ ; /* 0x0000000602007c10 */ /* 0x002fc8000fffe0ff */ /*0650*/ LEA R0, R0, R3, 0xb ; /* 0x0000000300007211 */ /* 0x000fca00078e58ff */ /*0660*/ IMAD.WIDE.U32 R2, R0, R5, c[0x0][0x170] ; /* 0x00005c0000027625 */ /* 0x000fca00078e0005 */ /*0670*/ STG.E [R2.64], R26 ; /* 0x0000001a02007986 */ /* 0x000fe2000c101904 */ /*0680*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0690*/ BRA 0x690; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0700*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0710*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0720*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0730*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0740*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0750*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0760*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0770*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
.file "tmpxft_0019ad65_00000000-6_gcompiler_matmul_THREADBLOCK_X_16__THREAD_Y_32_.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii .type _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii, @function _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movq %rsp, %rax movq %rax, 128(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z6matmulPfS_S_ii(%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 _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii, .-_Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii .globl _Z6matmulPfS_S_ii .type _Z6matmulPfS_S_ii, @function _Z6matmulPfS_S_ii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z6matmulPfS_S_ii, .-_Z6matmulPfS_S_ii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z6matmulPfS_S_ii" .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 _Z6matmulPfS_S_ii(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2054: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
#include <hip/hip_runtime.h> #define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z6matmulPfS_S_ii .globl _Z6matmulPfS_S_ii .p2align 8 .type _Z6matmulPfS_S_ii,@function _Z6matmulPfS_S_ii: s_load_b32 s3, s[0:1], 0x18 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s3, 1 s_cbranch_scc1 .LBB0_5 s_load_b128 s[4:7], s[0:1], 0x0 v_dual_mov_b32 v2, 0 :: v_dual_and_b32 v5, 0x3ff, v0 v_add_lshl_u32 v6, s15, v3, 11 s_lshl_b32 s8, s14, 8 v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_3) v_cmp_gt_u32_e32 vcc_lo, 16, v5 v_lshlrev_b32_e32 v7, 2, v5 s_mov_b32 s9, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s10 v_dual_mov_b32 v9, 0 :: v_dual_add_nc_u32 v8, s8, v5 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x800, v8 v_lshlrev_b64 v[10:11], 2, v[8:9] v_dual_mov_b32 v15, v9 :: v_dual_add_nc_u32 v14, 0x1000, v8 v_dual_mov_b32 v17, v9 :: v_dual_add_nc_u32 v16, 0x2000, v8 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_lshlrev_b64 v[14:15], 2, v[14:15] v_add_co_u32 v12, s2, s6, v12 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 global_load_b32 v1, v[10:11], off v_dual_mov_b32 v19, v9 :: v_dual_add_nc_u32 v18, 0x2800, v8 global_load_b32 v26, v[12:13], off v_dual_mov_b32 v11, v9 :: v_dual_add_nc_u32 v10, 0x1800, v8 v_add_co_u32 v14, s2, s6, v14 v_lshlrev_b64 v[16:17], 2, v[16:17] v_dual_mov_b32 v21, v9 :: v_dual_add_nc_u32 v20, 0x3000, v8 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_3) | instid1(VALU_DEP_4) v_lshlrev_b64 v[10:11], 2, v[10:11] v_add_co_ci_u32_e64 v15, s2, s7, v15, s2 v_lshlrev_b64 v[18:19], 2, v[18:19] v_dual_mov_b32 v23, v9 :: v_dual_add_nc_u32 v22, 0x3800, v8 v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_co_u32 v16, s2, s6, v16 v_lshlrev_b64 v[20:21], 2, v[20:21] v_add_co_ci_u32_e64 v17, s2, s7, v17, s2 v_add_co_u32 v18, s2, s6, v18 v_lshlrev_b64 v[22:23], 2, v[22:23] v_add_co_ci_u32_e64 v19, s2, s7, v19, s2 v_add_co_u32 v20, s2, s6, v20 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v21, s2, s7, v21, s2 v_add_co_u32 v22, s2, s6, v22 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v23, s2, s7, v23, s2 s_clause 0x5 global_load_b32 v27, v[14:15], off global_load_b32 v28, v[10:11], off global_load_b32 v29, v[16:17], off global_load_b32 v30, v[18:19], off global_load_b32 v20, v[20:21], off global_load_b32 v21, v[22:23], off v_dual_mov_b32 v25, v9 :: v_dual_add_nc_u32 v24, 0x4000, v8 v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x4800, v8 v_dual_mov_b32 v17, v9 :: v_dual_add_nc_u32 v16, 0x5800, v8 v_mov_b32_e32 v19, v9 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_lshlrev_b64 v[24:25], 2, v[24:25] v_lshlrev_b64 v[10:11], 2, v[12:13] v_add_nc_u32_e32 v12, 0x5000, v8 v_add_nc_u32_e32 v18, 0x6000, v8 v_lshlrev_b64 v[16:17], 2, v[16:17] v_add_nc_u32_e32 v6, 16, v6 v_add_co_u32 v14, s2, s6, v24 v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_ci_u32_e64 v15, s2, s7, v25, s2 v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_co_u32 v12, s2, s6, v12 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 s_clause 0x2 global_load_b32 v22, v[14:15], off global_load_b32 v23, v[10:11], off global_load_b32 v24, v[12:13], off v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x6800, v8 v_lshlrev_b64 v[10:11], 2, v[18:19] v_add_co_u32 v14, s2, s6, v16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v15, s2, s7, v17, s2 v_mov_b32_e32 v17, v9 v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_nc_u32_e32 v16, 0x7000, v8 v_add_co_u32 v10, s2, s6, v10 v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_nc_u32_e32 v8, 0x7800, v8 v_add_co_u32 v12, s2, s6, v12 global_load_b32 v18, v[14:15], off v_lshlrev_b64 v[14:15], 2, v[16:17] v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 v_lshlrev_b64 v[8:9], 2, v[8:9] s_clause 0x1 global_load_b32 v16, v[10:11], off global_load_b32 v17, v[12:13], off v_add_co_u32 v10, s2, s6, v14 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v11, s2, s7, v15, s2 v_add_co_u32 v8, s2, s6, v8 v_add_co_ci_u32_e64 v9, s2, s7, v9, s2 s_clause 0x1 global_load_b32 v19, v[10:11], off global_load_b32 v25, v[8:9], off ds_load_2addr_b32 v[8:9], v2 offset1:1 ds_load_2addr_b32 v[10:11], v2 offset0:2 offset1:3 ds_load_2addr_b32 v[12:13], v2 offset0:4 offset1:5 ds_load_2addr_b32 v[14:15], v2 offset0:6 offset1:7 s_add_i32 s9, s9, 16 s_add_i32 s8, s8, 0x8000 s_cmp_lt_i32 s9, s3 s_waitcnt vmcnt(15) lgkmcnt(3) v_fmac_f32_e32 v4, v8, v1 s_waitcnt vmcnt(14) s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v9, v26 ds_load_2addr_b32 v[8:9], v2 offset0:8 offset1:9 s_waitcnt vmcnt(13) lgkmcnt(3) v_fmac_f32_e32 v4, v10, v27 s_waitcnt vmcnt(12) v_fmac_f32_e32 v4, v11, v28 ds_load_2addr_b32 v[10:11], v2 offset0:10 offset1:11 s_waitcnt vmcnt(11) lgkmcnt(3) v_fmac_f32_e32 v4, v12, v29 s_waitcnt vmcnt(10) s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v13, v30 s_waitcnt vmcnt(9) lgkmcnt(2) v_fmac_f32_e32 v4, v14, v20 s_waitcnt vmcnt(8) s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v15, v21 ds_load_2addr_b32 v[12:13], v2 offset0:12 offset1:13 ds_load_2addr_b32 v[14:15], v2 offset0:14 offset1:15 s_waitcnt vmcnt(0) lgkmcnt(0) s_barrier buffer_gl0_inv v_fmac_f32_e32 v4, v8, v22 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v9, v23 v_fmac_f32_e32 v4, v10, v24 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v11, v18 v_fmac_f32_e32 v4, v12, v16 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v13, v17 v_fmac_f32_e32 v4, v14, v19 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v15, v25 s_cbranch_scc0 .LBB0_6 .LBB0_3: s_and_saveexec_b32 s10, vcc_lo s_cbranch_execz .LBB0_2 v_add_nc_u32_e32 v1, v5, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[8:9], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v8, s2, s4, v8 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v9, s2, s5, v9, s2 global_load_b32 v1, v[8:9], off s_waitcnt vmcnt(0) ds_store_b32 v7, v1 s_branch .LBB0_2 .LBB0_5: v_mov_b32_e32 v4, 0 .LBB0_6: s_load_b64 s[0:1], s[0:1], 0x10 v_add_lshl_u32 v2, s15, v3, 11 v_dual_mov_b32 v1, 0 :: v_dual_and_b32 v0, 0x3ff, v0 s_lshl_b32 s2, s14, 8 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_add3_u32 v0, s2, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v4, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z6matmulPfS_S_ii .amdhsa_group_segment_fixed_size 64 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .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 31 .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 _Z6matmulPfS_S_ii, .Lfunc_end0-_Z6matmulPfS_S_ii .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 .group_segment_fixed_size: 64 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z6matmulPfS_S_ii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z6matmulPfS_S_ii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 31 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadIdx.y) #define merger_y 1 #define coalesced_idy (bidy/(COALESCED_NUM/(merger_y*blockDimY))*COALESCED_NUM) #define A(y,x) A[(y)*WIDTH_A+(x)] #define B(y,x) B[(y)*WIDTH_B+(x)] #define C(y,x) C[(y)*WIDTH_C+(x)] #define WIDTH_C 2048 #define WIDTH_B 2048 #define WIDTH_A 2048 __global__ void matmul(float * A, float * B, float * C, int width, int height) { __shared__ float shared_0[16]; int i; float sum; sum=0; for (i=0; i<width; i=(i+16)) { int it_1; if ((tidx<16)) { shared_0[(tidx+0)]=A(idy, (i+tidx)); } __syncthreads(); #pragma unroll for (it_1=0; it_1<16; it_1=(it_1+1)) { float a; float b; a=shared_0[it_1]; b=B((it_1+i), idx); sum+=(a*b); } __syncthreads(); } { C(idy, idx)=sum; } }
.text .file "gcompiler_matmul_THREADBLOCK_X_16__THREAD_Y_32_.hip" .globl _Z21__device_stub__matmulPfS_S_ii # -- Begin function _Z21__device_stub__matmulPfS_S_ii .p2align 4, 0x90 .type _Z21__device_stub__matmulPfS_S_ii,@function _Z21__device_stub__matmulPfS_S_ii: # @_Z21__device_stub__matmulPfS_S_ii .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 4(%rsp), %rax movq %rax, 104(%rsp) movq %rsp, %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z6matmulPfS_S_ii, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z21__device_stub__matmulPfS_S_ii, .Lfunc_end0-_Z21__device_stub__matmulPfS_S_ii .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 $_Z6matmulPfS_S_ii, %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 _Z6matmulPfS_S_ii,@object # @_Z6matmulPfS_S_ii .section .rodata,"a",@progbits .globl _Z6matmulPfS_S_ii .p2align 3, 0x0 _Z6matmulPfS_S_ii: .quad _Z21__device_stub__matmulPfS_S_ii .size _Z6matmulPfS_S_ii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z6matmulPfS_S_ii" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z21__device_stub__matmulPfS_S_ii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6matmulPfS_S_ii .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 : _Z6matmulPfS_S_ii .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ ISETP.LT.AND P0, PT, RZ, c[0x0][0x178], PT ; /* 0x00005e00ff007a0c */ /* 0x000fe20003f01270 */ /*0020*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd80000000a00 */ /*0030*/ @!P0 MOV R26, RZ ; /* 0x000000ff001a8202 */ /* 0x000fe20000000f00 */ /*0040*/ @!P0 BRA 0x5e0 ; /* 0x0000059000008947 */ /* 0x000fea0003800000 */ /*0050*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e220000002100 */ /*0060*/ MOV R26, RZ ; /* 0x000000ff001a7202 */ /* 0x000fc60000000f00 */ /*0070*/ S2R R0, SR_TID.Y ; /* 0x0000000000007919 */ /* 0x000e680000002200 */ /*0080*/ S2R R2, SR_CTAID.X ; /* 0x0000000000027919 */ /* 0x000ea80000002500 */ /*0090*/ S2R R4, SR_CTAID.Y ; /* 0x0000000000047919 */ /* 0x000ee20000002600 */ /*00a0*/ ISETP.GT.U32.AND P0, PT, R3, 0xf, PT ; /* 0x0000000f0300780c */ /* 0x001fe40003f04070 */ /*00b0*/ LEA R17, R0, R3.reuse, 0xb ; /* 0x0000000300117211 */ /* 0x082fe200078e58ff */ /*00c0*/ HFMA2.MMA R0, -RZ, RZ, 0, 0 ; /* 0x00000000ff007435 */ /* 0x000fe200000001ff */ /*00d0*/ LEA R2, R2, R3, 0x8 ; /* 0x0000000302027211 */ /* 0x004fc400078e40ff */ /*00e0*/ LEA R17, R4, R17, 0xb ; /* 0x0000001104117211 */ /* 0x008fe400078e58ff */ /*00f0*/ IADD3 R16, R2, 0x7800, RZ ; /* 0x0000780002107810 */ /* 0x000fca0007ffe0ff */ /*0100*/ @!P0 IADD3 R6, R17, R0, RZ ; /* 0x0000000011068210 */ /* 0x000fe40007ffe0ff */ /*0110*/ MOV R27, 0x4 ; /* 0x00000004001b7802 */ /* 0x000fca0000000f00 */ /*0120*/ @!P0 IMAD.WIDE.U32 R6, R6, R27, c[0x0][0x160] ; /* 0x0000580006068625 */ /* 0x000fcc00078e001b */ /*0130*/ @!P0 LDG.E R6, [R6.64] ; /* 0x0000000406068981 */ /* 0x000ea2000c1e1900 */ /*0140*/ IMAD.WIDE.U32 R8, R2.reuse, R27, c[0x0][0x168] ; /* 0x00005a0002087625 */ /* 0x040fe200078e001b */ /*0150*/ IADD3 R12, R2.reuse, 0x800, RZ ; /* 0x00000800020c7810 */ /* 0x040fe40007ffe0ff */ /*0160*/ IADD3 R18, R2, 0x1000, RZ ; /* 0x0000100002127810 */ /* 0x000fc60007ffe0ff */ /*0170*/ IMAD.WIDE.U32 R12, R12, R27, c[0x0][0x168] ; /* 0x00005a000c0c7625 */ /* 0x000fc800078e001b */ /*0180*/ IMAD.WIDE.U32 R18, R18, R27, c[0x0][0x168] ; /* 0x00005a0012127625 */ /* 0x000fe200078e001b */ /*0190*/ IADD3 R10, R2.reuse, 0x1800, RZ ; /* 0x00001800020a7810 */ /* 0x040fe40007ffe0ff */ /*01a0*/ IADD3 R24, R2, 0x2000, RZ ; /* 0x0000200002187810 */ /* 0x000fca0007ffe0ff */ /*01b0*/ IMAD.WIDE.U32 R24, R24, R27.reuse, c[0x0][0x168] ; /* 0x00005a0018187625 */ /* 0x080fe200078e001b */ /*01c0*/ IADD3 R20, R2, 0x2800, RZ ; /* 0x0000280002147810 */ /* 0x000fe20007ffe0ff */ /*01d0*/ @!P0 STS [R3.X4], R6 ; /* 0x0000000603008388 */ /* 0x0041e80000004800 */ /*01e0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*01f0*/ LDG.E R21, [R8.64] ; /* 0x0000000408157981 */ /* 0x0002a8000c1e1900 */ /*0200*/ LDG.E R4, [R12.64] ; /* 0x000000040c047981 */ /* 0x000728000c1e1900 */ /*0210*/ LDG.E R5, [R18.64] ; /* 0x0000000412057981 */ /* 0x000b22000c1e1900 */ /*0220*/ IMAD.WIDE.U32 R6, R10, R27, c[0x0][0x168] ; /* 0x00005a000a067625 */ /* 0x001fc600078e001b */ /*0230*/ LDG.E R15, [R24.64] ; /* 0x00000004180f7981 */ /* 0x000128000c1e1900 */ /*0240*/ LDG.E R14, [R6.64] ; /* 0x00000004060e7981 */ /* 0x000122000c1e1900 */ /*0250*/ IADD3 R22, R2, 0x3000, RZ ; /* 0x0000300002167810 */ /* 0x000fe20007ffe0ff */ /*0260*/ IMAD.WIDE.U32 R12, R20, R27.reuse, c[0x0][0x168] ; /* 0x00005a00140c7625 */ /* 0x088fe400078e001b */ /*0270*/ LDS.128 R8, [RZ] ; /* 0x00000000ff087984 */ /* 0x002ea40000000c00 */ /*0280*/ IMAD.WIDE.U32 R22, R22, R27, c[0x0][0x168] ; /* 0x00005a0016167625 */ /* 0x000fc400078e001b */ /*0290*/ LDG.E R12, [R12.64] ; /* 0x000000040c0c7981 */ /* 0x0002e8000c1e1900 */ /*02a0*/ LDG.E R13, [R22.64] ; /* 0x00000004160d7981 */ /* 0x0022e2000c1e1900 */ /*02b0*/ IADD3 R28, R2, 0x3800, RZ ; /* 0x00003800021c7810 */ /* 0x000fe40007ffe0ff */ /*02c0*/ IADD3 R18, R16, -0x3800, RZ ; /* 0xffffc80010127810 */ /* 0x020fc60007ffe0ff */ /*02d0*/ IMAD.WIDE.U32 R28, R28, R27, c[0x0][0x168] ; /* 0x00005a001c1c7625 */ /* 0x000fe200078e001b */ /*02e0*/ IADD3 R24, R16, -0x3000, RZ ; /* 0xffffd00010187810 */ /* 0x001fc60007ffe0ff */ /*02f0*/ IMAD.WIDE.U32 R6, R18, R27.reuse, c[0x0][0x168] ; /* 0x00005a0012067625 */ /* 0x080fe400078e001b */ /*0300*/ LDG.E R18, [R28.64] ; /* 0x000000041c127981 */ /* 0x000162000c1e1900 */ /*0310*/ IADD3 R22, R16, -0x2800, RZ ; /* 0xffffd80010167810 */ /* 0x002fe20007ffe0ff */ /*0320*/ IMAD.WIDE.U32 R24, R24, R27.reuse, c[0x0][0x168] ; /* 0x00005a0018187625 */ /* 0x080fe400078e001b */ /*0330*/ LDG.E R19, [R6.64] ; /* 0x0000000406137981 */ /* 0x0002e4000c1e1900 */ /*0340*/ IMAD.WIDE.U32 R22, R22, R27, c[0x0][0x168] ; /* 0x00005a0016167625 */ /* 0x000fe400078e001b */ /*0350*/ LDG.E R20, [R24.64] ; /* 0x0000000418147981 */ /* 0x0004e2000c1e1900 */ /*0360*/ IADD3 R28, R16, -0x1800, RZ ; /* 0xffffe800101c7810 */ /* 0x001fc40007ffe0ff */ /*0370*/ IADD3 R6, R16, -0x2000, RZ ; /* 0xffffe00010067810 */ /* 0x002fc60007ffe0ff */ /*0380*/ IMAD.WIDE.U32 R28, R28, R27, c[0x0][0x168] ; /* 0x00005a001c1c7625 */ /* 0x000fc800078e001b */ /*0390*/ IMAD.WIDE.U32 R6, R6, R27, c[0x0][0x168] ; /* 0x00005a0006067625 */ /* 0x000fc800078e001b */ /*03a0*/ FFMA R8, R21, R8, R26 ; /* 0x0000000815087223 */ /* 0x004fe2000000001a */ /*03b0*/ IADD3 R26, R16, -0x1000, RZ ; /* 0xfffff000101a7810 */ /* 0x000fe20007ffe0ff */ /*03c0*/ LDG.E R21, [R22.64] ; /* 0x0000000416157981 */ /* 0x0000a8000c1e1900 */ /*03d0*/ IMAD.WIDE.U32 R24, R26, R27, c[0x0][0x168] ; /* 0x00005a001a187625 */ /* 0x000fe200078e001b */ /*03e0*/ LDG.E R22, [R6.64] ; /* 0x0000000406167981 */ /* 0x0010a8000c1e1900 */ /*03f0*/ LDG.E R23, [R28.64] ; /* 0x000000041c177981 */ /* 0x0002a8000c1e1900 */ /*0400*/ LDG.E R24, [R24.64] ; /* 0x0000000418187981 */ /* 0x0008a2000c1e1900 */ /*0410*/ IADD3 R6, R16, -0x800, RZ ; /* 0xfffff80010067810 */ /* 0x001fca0007ffe0ff */ /*0420*/ IMAD.WIDE.U32 R28, R6, R27, c[0x0][0x168] ; /* 0x00005a00061c7625 */ /* 0x002fc800078e001b */ /*0430*/ IMAD.WIDE.U32 R26, R16, R27, c[0x0][0x168] ; /* 0x00005a00101a7625 */ /* 0x000fe200078e001b */ /*0440*/ LDG.E R25, [R28.64] ; /* 0x000000041c197981 */ /* 0x010f2a000c1e1900 */ /*0450*/ LDG.E R26, [R26.64] ; /* 0x000000041a1a7981 */ /* 0x000f22000c1e1900 */ /*0460*/ FFMA R4, R4, R9, R8 ; /* 0x0000000904047223 */ /* 0x000fc80000000008 */ /*0470*/ FFMA R10, R5, R10, R4 ; /* 0x0000000a050a7223 */ /* 0x000fe40000000004 */ /*0480*/ LDS.128 R4, [0x10] ; /* 0x00001000ff047984 */ /* 0x000e240000000c00 */ /*0490*/ FFMA R10, R14, R11, R10 ; /* 0x0000000b0e0a7223 */ /* 0x000fc8000000000a */ /*04a0*/ FFMA R4, R15, R4, R10 ; /* 0x000000040f047223 */ /* 0x001fe4000000000a */ /*04b0*/ LDS.128 R8, [0x20] ; /* 0x00002000ff087984 */ /* 0x000e240000000c00 */ /*04c0*/ FFMA R4, R12, R5, R4 ; /* 0x000000050c047223 */ /* 0x008fc80000000004 */ /*04d0*/ FFMA R4, R13, R6, R4 ; /* 0x000000060d047223 */ /* 0x000fe40000000004 */ /*04e0*/ LDS.128 R12, [0x30] ; /* 0x00003000ff0c7984 */ /* 0x000e640000000c00 */ /*04f0*/ FFMA R4, R18, R7, R4 ; /* 0x0000000712047223 */ /* 0x020fe20000000004 */ /*0500*/ IADD3 R0, R0, 0x10, RZ ; /* 0x0000001000007810 */ /* 0x000fe20007ffe0ff */ /*0510*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fe60000010000 */ /*0520*/ ISETP.GE.AND P1, PT, R0, c[0x0][0x178], PT ; /* 0x00005e0000007a0c */ /* 0x000fe20003f26270 */ /*0530*/ FFMA R4, R19, R8, R4 ; /* 0x0000000813047223 */ /* 0x001fc80000000004 */ /*0540*/ FFMA R4, R20, R9, R4 ; /* 0x0000000914047223 */ /* 0x000fe20000000004 */ /*0550*/ IADD3 R2, R2, 0x8000, RZ ; /* 0x0000800002027810 */ /* 0x000fe40007ffe0ff */ /*0560*/ IADD3 R16, R16, 0x8000, RZ ; /* 0x0000800010107810 */ /* 0x000fe20007ffe0ff */ /*0570*/ FFMA R4, R21, R10, R4 ; /* 0x0000000a15047223 */ /* 0x004fc80000000004 */ /*0580*/ FFMA R4, R22, R11, R4 ; /* 0x0000000b16047223 */ /* 0x000fc80000000004 */ /*0590*/ FFMA R4, R23, R12, R4 ; /* 0x0000000c17047223 */ /* 0x002fc80000000004 */ /*05a0*/ FFMA R4, R24, R13, R4 ; /* 0x0000000d18047223 */ /* 0x000fc80000000004 */ /*05b0*/ FFMA R4, R25, R14, R4 ; /* 0x0000000e19047223 */ /* 0x010fc80000000004 */ /*05c0*/ FFMA R26, R26, R15, R4 ; /* 0x0000000f1a1a7223 */ /* 0x000fe20000000004 */ /*05d0*/ @!P1 BRA 0x100 ; /* 0xfffffb2000009947 */ /* 0x000fea000383ffff */ /*05e0*/ S2R R0, SR_CTAID.X ; /* 0x0000000000007919 */ /* 0x000e220000002500 */ /*05f0*/ S2UR UR6, SR_CTAID.Y ; /* 0x00000000000679c3 */ /* 0x000e620000002600 */ /*0600*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe400000001ff */ /*0610*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e280000002100 */ /*0620*/ S2R R2, SR_TID.Y ; /* 0x0000000000027919 */ /* 0x000e620000002200 */ /*0630*/ LEA R3, R0, R3, 0x8 ; /* 0x0000000300037211 */ /* 0x001fc400078e40ff */ /*0640*/ IADD3 R0, R2, UR6, RZ ; /* 0x0000000602007c10 */ /* 0x002fc8000fffe0ff */ /*0650*/ LEA R0, R0, R3, 0xb ; /* 0x0000000300007211 */ /* 0x000fca00078e58ff */ /*0660*/ IMAD.WIDE.U32 R2, R0, R5, c[0x0][0x170] ; /* 0x00005c0000027625 */ /* 0x000fca00078e0005 */ /*0670*/ STG.E [R2.64], R26 ; /* 0x0000001a02007986 */ /* 0x000fe2000c101904 */ /*0680*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0690*/ BRA 0x690; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*06a0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06b0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*06f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0700*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0710*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0720*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0730*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0740*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0750*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0760*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0770*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z6matmulPfS_S_ii .globl _Z6matmulPfS_S_ii .p2align 8 .type _Z6matmulPfS_S_ii,@function _Z6matmulPfS_S_ii: s_load_b32 s3, s[0:1], 0x18 v_bfe_u32 v3, v0, 10, 10 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s3, 1 s_cbranch_scc1 .LBB0_5 s_load_b128 s[4:7], s[0:1], 0x0 v_dual_mov_b32 v2, 0 :: v_dual_and_b32 v5, 0x3ff, v0 v_add_lshl_u32 v6, s15, v3, 11 s_lshl_b32 s8, s14, 8 v_mov_b32_e32 v4, 0 s_delay_alu instid0(VALU_DEP_3) v_cmp_gt_u32_e32 vcc_lo, 16, v5 v_lshlrev_b32_e32 v7, 2, v5 s_mov_b32 s9, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s10 v_dual_mov_b32 v9, 0 :: v_dual_add_nc_u32 v8, s8, v5 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x800, v8 v_lshlrev_b64 v[10:11], 2, v[8:9] v_dual_mov_b32 v15, v9 :: v_dual_add_nc_u32 v14, 0x1000, v8 v_dual_mov_b32 v17, v9 :: v_dual_add_nc_u32 v16, 0x2000, v8 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_lshlrev_b64 v[14:15], 2, v[14:15] v_add_co_u32 v12, s2, s6, v12 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 global_load_b32 v1, v[10:11], off v_dual_mov_b32 v19, v9 :: v_dual_add_nc_u32 v18, 0x2800, v8 global_load_b32 v26, v[12:13], off v_dual_mov_b32 v11, v9 :: v_dual_add_nc_u32 v10, 0x1800, v8 v_add_co_u32 v14, s2, s6, v14 v_lshlrev_b64 v[16:17], 2, v[16:17] v_dual_mov_b32 v21, v9 :: v_dual_add_nc_u32 v20, 0x3000, v8 s_delay_alu instid0(VALU_DEP_4) | instskip(SKIP_3) | instid1(VALU_DEP_4) v_lshlrev_b64 v[10:11], 2, v[10:11] v_add_co_ci_u32_e64 v15, s2, s7, v15, s2 v_lshlrev_b64 v[18:19], 2, v[18:19] v_dual_mov_b32 v23, v9 :: v_dual_add_nc_u32 v22, 0x3800, v8 v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_co_u32 v16, s2, s6, v16 v_lshlrev_b64 v[20:21], 2, v[20:21] v_add_co_ci_u32_e64 v17, s2, s7, v17, s2 v_add_co_u32 v18, s2, s6, v18 v_lshlrev_b64 v[22:23], 2, v[22:23] v_add_co_ci_u32_e64 v19, s2, s7, v19, s2 v_add_co_u32 v20, s2, s6, v20 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v21, s2, s7, v21, s2 v_add_co_u32 v22, s2, s6, v22 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v23, s2, s7, v23, s2 s_clause 0x5 global_load_b32 v27, v[14:15], off global_load_b32 v28, v[10:11], off global_load_b32 v29, v[16:17], off global_load_b32 v30, v[18:19], off global_load_b32 v20, v[20:21], off global_load_b32 v21, v[22:23], off v_dual_mov_b32 v25, v9 :: v_dual_add_nc_u32 v24, 0x4000, v8 v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x4800, v8 v_dual_mov_b32 v17, v9 :: v_dual_add_nc_u32 v16, 0x5800, v8 v_mov_b32_e32 v19, v9 s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4) v_lshlrev_b64 v[24:25], 2, v[24:25] v_lshlrev_b64 v[10:11], 2, v[12:13] v_add_nc_u32_e32 v12, 0x5000, v8 v_add_nc_u32_e32 v18, 0x6000, v8 v_lshlrev_b64 v[16:17], 2, v[16:17] v_add_nc_u32_e32 v6, 16, v6 v_add_co_u32 v14, s2, s6, v24 v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_ci_u32_e64 v15, s2, s7, v25, s2 v_add_co_u32 v10, s2, s6, v10 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_4) v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_co_u32 v12, s2, s6, v12 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 s_clause 0x2 global_load_b32 v22, v[14:15], off global_load_b32 v23, v[10:11], off global_load_b32 v24, v[12:13], off v_dual_mov_b32 v13, v9 :: v_dual_add_nc_u32 v12, 0x6800, v8 v_lshlrev_b64 v[10:11], 2, v[18:19] v_add_co_u32 v14, s2, s6, v16 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v15, s2, s7, v17, s2 v_mov_b32_e32 v17, v9 v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_nc_u32_e32 v16, 0x7000, v8 v_add_co_u32 v10, s2, s6, v10 v_add_co_ci_u32_e64 v11, s2, s7, v11, s2 v_add_nc_u32_e32 v8, 0x7800, v8 v_add_co_u32 v12, s2, s6, v12 global_load_b32 v18, v[14:15], off v_lshlrev_b64 v[14:15], 2, v[16:17] v_add_co_ci_u32_e64 v13, s2, s7, v13, s2 v_lshlrev_b64 v[8:9], 2, v[8:9] s_clause 0x1 global_load_b32 v16, v[10:11], off global_load_b32 v17, v[12:13], off v_add_co_u32 v10, s2, s6, v14 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v11, s2, s7, v15, s2 v_add_co_u32 v8, s2, s6, v8 v_add_co_ci_u32_e64 v9, s2, s7, v9, s2 s_clause 0x1 global_load_b32 v19, v[10:11], off global_load_b32 v25, v[8:9], off ds_load_2addr_b32 v[8:9], v2 offset1:1 ds_load_2addr_b32 v[10:11], v2 offset0:2 offset1:3 ds_load_2addr_b32 v[12:13], v2 offset0:4 offset1:5 ds_load_2addr_b32 v[14:15], v2 offset0:6 offset1:7 s_add_i32 s9, s9, 16 s_add_i32 s8, s8, 0x8000 s_cmp_lt_i32 s9, s3 s_waitcnt vmcnt(15) lgkmcnt(3) v_fmac_f32_e32 v4, v8, v1 s_waitcnt vmcnt(14) s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v9, v26 ds_load_2addr_b32 v[8:9], v2 offset0:8 offset1:9 s_waitcnt vmcnt(13) lgkmcnt(3) v_fmac_f32_e32 v4, v10, v27 s_waitcnt vmcnt(12) v_fmac_f32_e32 v4, v11, v28 ds_load_2addr_b32 v[10:11], v2 offset0:10 offset1:11 s_waitcnt vmcnt(11) lgkmcnt(3) v_fmac_f32_e32 v4, v12, v29 s_waitcnt vmcnt(10) s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v13, v30 s_waitcnt vmcnt(9) lgkmcnt(2) v_fmac_f32_e32 v4, v14, v20 s_waitcnt vmcnt(8) s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v15, v21 ds_load_2addr_b32 v[12:13], v2 offset0:12 offset1:13 ds_load_2addr_b32 v[14:15], v2 offset0:14 offset1:15 s_waitcnt vmcnt(0) lgkmcnt(0) s_barrier buffer_gl0_inv v_fmac_f32_e32 v4, v8, v22 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v9, v23 v_fmac_f32_e32 v4, v10, v24 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v11, v18 v_fmac_f32_e32 v4, v12, v16 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fmac_f32_e32 v4, v13, v17 v_fmac_f32_e32 v4, v14, v19 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v15, v25 s_cbranch_scc0 .LBB0_6 .LBB0_3: s_and_saveexec_b32 s10, vcc_lo s_cbranch_execz .LBB0_2 v_add_nc_u32_e32 v1, v5, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[8:9], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v8, s2, s4, v8 s_delay_alu instid0(VALU_DEP_1) v_add_co_ci_u32_e64 v9, s2, s5, v9, s2 global_load_b32 v1, v[8:9], off s_waitcnt vmcnt(0) ds_store_b32 v7, v1 s_branch .LBB0_2 .LBB0_5: v_mov_b32_e32 v4, 0 .LBB0_6: s_load_b64 s[0:1], s[0:1], 0x10 v_add_lshl_u32 v2, s15, v3, 11 v_dual_mov_b32 v1, 0 :: v_dual_and_b32 v0, 0x3ff, v0 s_lshl_b32 s2, s14, 8 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_add3_u32 v0, s2, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[0:1] s_waitcnt lgkmcnt(0) v_add_co_u32 v0, vcc_lo, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v4, off s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z6matmulPfS_S_ii .amdhsa_group_segment_fixed_size 64 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 32 .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 31 .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 _Z6matmulPfS_S_ii, .Lfunc_end0-_Z6matmulPfS_S_ii .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 .group_segment_fixed_size: 64 .kernarg_segment_align: 8 .kernarg_segment_size: 32 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z6matmulPfS_S_ii .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z6matmulPfS_S_ii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 31 .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_0019ad65_00000000-6_gcompiler_matmul_THREADBLOCK_X_16__THREAD_Y_32_.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii .type _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii, @function _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii: .LFB2051: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movq %rsp, %rax movq %rax, 128(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z6matmulPfS_S_ii(%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 _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii, .-_Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii .globl _Z6matmulPfS_S_ii .type _Z6matmulPfS_S_ii, @function _Z6matmulPfS_S_ii: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z31__device_stub__Z6matmulPfS_S_iiPfS_S_ii addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z6matmulPfS_S_ii, .-_Z6matmulPfS_S_ii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z6matmulPfS_S_ii" .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 _Z6matmulPfS_S_ii(%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 "gcompiler_matmul_THREADBLOCK_X_16__THREAD_Y_32_.hip" .globl _Z21__device_stub__matmulPfS_S_ii # -- Begin function _Z21__device_stub__matmulPfS_S_ii .p2align 4, 0x90 .type _Z21__device_stub__matmulPfS_S_ii,@function _Z21__device_stub__matmulPfS_S_ii: # @_Z21__device_stub__matmulPfS_S_ii .cfi_startproc # %bb.0: subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) movl %ecx, 4(%rsp) movl %r8d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 4(%rsp), %rax movq %rax, 104(%rsp) movq %rsp, %rax movq %rax, 112(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z6matmulPfS_S_ii, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $136, %rsp .cfi_adjust_cfa_offset -136 retq .Lfunc_end0: .size _Z21__device_stub__matmulPfS_S_ii, .Lfunc_end0-_Z21__device_stub__matmulPfS_S_ii .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 $_Z6matmulPfS_S_ii, %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 _Z6matmulPfS_S_ii,@object # @_Z6matmulPfS_S_ii .section .rodata,"a",@progbits .globl _Z6matmulPfS_S_ii .p2align 3, 0x0 _Z6matmulPfS_S_ii: .quad _Z21__device_stub__matmulPfS_S_ii .size _Z6matmulPfS_S_ii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z6matmulPfS_S_ii" .size .L__unnamed_1, 18 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z21__device_stub__matmulPfS_S_ii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6matmulPfS_S_ii .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include<iostream> #include<cuda.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; cudaMalloc(&a,N*sizeof(int)); cudaMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; cudaMemcpy(a,temp,N*sizeof(int),cudaMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; cudaMemcpy(b,temp,N*sizeof(int),cudaMemcpyHostToDevice); add<<<N,1>>>(a,b); cudaMemcpy(temp,a,N*sizeof(int),cudaMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; cudaFree(a); cudaFree(b); }
code for sm_80 Function : _Z3addPiPKi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fc800078e0205 */ /*0050*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e0205 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*0080*/ IADD3 R7, R2, R7, RZ ; /* 0x0000000702077210 */ /* 0x004fca0007ffe0ff */ /*0090*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00b0*/ BRA 0xb0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include<iostream> #include<cuda.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; cudaMalloc(&a,N*sizeof(int)); cudaMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; cudaMemcpy(a,temp,N*sizeof(int),cudaMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; cudaMemcpy(b,temp,N*sizeof(int),cudaMemcpyHostToDevice); add<<<N,1>>>(a,b); cudaMemcpy(temp,a,N*sizeof(int),cudaMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; cudaFree(a); cudaFree(b); }
.file "tmpxft_00152e3b_00000000-6_test.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z25__device_stub__Z3addPiPKiPiPKi .type _Z25__device_stub__Z3addPiPKiPiPKi, @function _Z25__device_stub__Z3addPiPKiPiPKi: .LFB3694: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z3addPiPKi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z25__device_stub__Z3addPiPKiPiPKi, .-_Z25__device_stub__Z3addPiPKiPiPKi .globl _Z3addPiPKi .type _Z3addPiPKi, @function _Z3addPiPKi: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z3addPiPKiPiPKi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z3addPiPKi, .-_Z3addPiPKi .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $56, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $40, %edi call _Znam@PLT movq %rax, %r12 movq %rsp, %rdi movl $40, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $40, %esi call cudaMalloc@PLT movl $0, %eax .L12: movl %eax, (%r12,%rax,4) addq $1, %rax cmpq $10, %rax jne .L12 movl $1, %ecx movl $40, %edx movq %r12, %rsi movq (%rsp), %rdi call cudaMemcpy@PLT movl $0, %eax .L13: leal (%rax,%rax), %edx movl %edx, (%r12,%rax,4) addq $1, %rax cmpq $10, %rax jne .L13 movl $1, %ecx movl $40, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $10, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L14: movl $2, %ecx movl $40, %edx movq (%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq %r12, %rbp leaq 40(%r12), %r14 leaq _ZSt4cout(%rip), %r13 jmp .L19 .L25: movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z25__device_stub__Z3addPiPKiPiPKi jmp .L14 .L28: movq 40(%rsp), %rax subq %fs:40, %rax jne .L26 call _ZSt16__throw_bad_castv@PLT .L26: call __stack_chk_fail@PLT .L17: movq %r15, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%r15), %rax movl $10, %esi movq %r15, %rdi call *48(%rax) movl %eax, %esi .L18: movsbl %sil, %esi movq %rbx, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addq $4, %rbp cmpq %r14, %rbp je .L27 .L19: movl 0(%rbp), %esi movq %r13, %rdi call _ZNSolsEi@PLT movq %rax, %rbx movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %r15 testq %r15, %r15 je .L28 cmpb $0, 56(%r15) je .L17 movzbl 67(%r15), %esi jmp .L18 .L27: movq %r12, %rdi call _ZdaPv@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L29 movl $0, %eax 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 .L29: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3addPiPKi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3addPiPKi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3697: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include<iostream> #include<cuda.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; cudaMalloc(&a,N*sizeof(int)); cudaMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; cudaMemcpy(a,temp,N*sizeof(int),cudaMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; cudaMemcpy(b,temp,N*sizeof(int),cudaMemcpyHostToDevice); add<<<N,1>>>(a,b); cudaMemcpy(temp,a,N*sizeof(int),cudaMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; cudaFree(a); cudaFree(b); }
#include<iostream> #include<hip/hip_runtime.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; hipMalloc(&a,N*sizeof(int)); hipMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; hipMemcpy(a,temp,N*sizeof(int),hipMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; hipMemcpy(b,temp,N*sizeof(int),hipMemcpyHostToDevice); add<<<N,1>>>(a,b); hipMemcpy(temp,a,N*sizeof(int),hipMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; hipFree(a); hipFree(b); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include<iostream> #include<hip/hip_runtime.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; hipMalloc(&a,N*sizeof(int)); hipMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; hipMemcpy(a,temp,N*sizeof(int),hipMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; hipMemcpy(b,temp,N*sizeof(int),hipMemcpyHostToDevice); add<<<N,1>>>(a,b); hipMemcpy(temp,a,N*sizeof(int),hipMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; hipFree(a); hipFree(b); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addPiPKi .globl _Z3addPiPKi .p2align 8 .type _Z3addPiPKi,@function _Z3addPiPKi: s_load_b128 s[0:3], s[0:1], 0x0 s_mov_b32 s4, s15 s_ashr_i32 s5, s15, 31 s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b64 s[4:5], s[4:5], 2 s_waitcnt lgkmcnt(0) s_add_u32 s2, s2, s4 s_addc_u32 s3, s3, s5 s_add_u32 s0, s0, s4 s_addc_u32 s1, s1, s5 s_load_b32 s2, s[2:3], 0x0 s_load_b32 s3, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_add_i32 s2, s3, s2 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s2 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3addPiPKi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 16 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 2 .amdhsa_next_free_sgpr 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 _Z3addPiPKi, .Lfunc_end0-_Z3addPiPKi .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 16 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addPiPKi .private_segment_fixed_size: 0 .sgpr_count: 16 .sgpr_spill_count: 0 .symbol: _Z3addPiPKi.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<iostream> #include<hip/hip_runtime.h> using namespace std; __global__ void add(int *a,const int *b){ int i=blockIdx.x; a[i]+=b[i]; } int main(){ const int N=10; int *a,*b,*temp; temp=new int[N]; hipMalloc(&a,N*sizeof(int)); hipMalloc(&b,N*sizeof(int)); for(int i=0;i<N;i++) temp[i]=i; hipMemcpy(a,temp,N*sizeof(int),hipMemcpyHostToDevice); for(int i=0;i<N;i++) temp[i]=2*i; hipMemcpy(b,temp,N*sizeof(int),hipMemcpyHostToDevice); add<<<N,1>>>(a,b); hipMemcpy(temp,a,N*sizeof(int),hipMemcpyDeviceToHost); for(int i=0;i<N;i++) cout<<temp[i]<<endl; delete[] temp; hipFree(a); hipFree(b); }
.text .file "test.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__addPiPKi # -- Begin function _Z18__device_stub__addPiPKi .p2align 4, 0x90 .type _Z18__device_stub__addPiPKi,@function _Z18__device_stub__addPiPKi: # @_Z18__device_stub__addPiPKi .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 $_Z3addPiPKi, %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 _Z18__device_stub__addPiPKi, .Lfunc_end0-_Z18__device_stub__addPiPKi .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 subq $104, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $40, %edi callq _Znam movq %rax, %rbx movq %rsp, %rdi movl $40, %esi callq hipMalloc leaq 8(%rsp), %rdi movl $40, %esi callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $10, %rax jne .LBB1_1 # %bb.2: movq (%rsp), %rdi movl $40, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy xorl %eax, %eax .p2align 4, 0x90 .LBB1_3: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,2) addq $2, %rax cmpq $20, %rax jne .LBB1_3 # %bb.4: movq 8(%rsp), %rdi movl $40, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdx # imm = 0x100000001 leaq 9(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_6 # %bb.5: movq (%rsp), %rax movq 8(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z3addPiPKi, %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_6: movq (%rsp), %rsi movl $40, %edx movq %rbx, %rdi movl $2, %ecx callq hipMemcpy xorl %r15d, %r15d jmp .LBB1_7 .p2align 4, 0x90 .LBB1_10: # in Loop: Header=BB1_7 Depth=1 movq %r14, %rdi movq %rax, %r12 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r14), %rax movq %r14, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r12, %rax .LBB1_11: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit # in Loop: Header=BB1_7 Depth=1 movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r15 cmpq $10, %r15 je .LBB1_12 .LBB1_7: # =>This Inner Loop Header: Depth=1 movl (%rbx,%r15,4), %esi movl $_ZSt4cout, %edi callq _ZNSolsEi movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %r14 testq %r14, %r14 je .LBB1_13 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i # in Loop: Header=BB1_7 Depth=1 cmpb $0, 56(%r14) je .LBB1_10 # %bb.9: # in Loop: Header=BB1_7 Depth=1 movzbl 67(%r14), %ecx jmp .LBB1_11 .LBB1_12: movq %rbx, %rdi callq _ZdaPv movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree xorl %eax, %eax addq $104, %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 .LBB1_13: .cfi_def_cfa_offset 144 callq _ZSt16__throw_bad_castv .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z3addPiPKi, %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 _Z3addPiPKi,@object # @_Z3addPiPKi .section .rodata,"a",@progbits .globl _Z3addPiPKi .p2align 3, 0x0 _Z3addPiPKi: .quad _Z18__device_stub__addPiPKi .size _Z3addPiPKi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3addPiPKi" .size .L__unnamed_1, 12 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z18__device_stub__addPiPKi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addPiPKi .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting from CUDA device assembly to AMD device assembly.
code for sm_80 Function : _Z3addPiPKi .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.X ; /* 0x0000000000047919 */ /* 0x000e220000002500 */ /*0020*/ HFMA2.MMA R5, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff057435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R4, R5, c[0x0][0x168] ; /* 0x00005a0004027625 */ /* 0x001fc800078e0205 */ /*0050*/ IMAD.WIDE R4, R4, R5, c[0x0][0x160] ; /* 0x0000580004047625 */ /* 0x000fe400078e0205 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R7, [R4.64] ; /* 0x0000000404077981 */ /* 0x000ea4000c1e1900 */ /*0080*/ IADD3 R7, R2, R7, RZ ; /* 0x0000000702077210 */ /* 0x004fca0007ffe0ff */ /*0090*/ STG.E [R4.64], R7 ; /* 0x0000000704007986 */ /* 0x000fe2000c101904 */ /*00a0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00b0*/ BRA 0xb0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00c0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addPiPKi .globl _Z3addPiPKi .p2align 8 .type _Z3addPiPKi,@function _Z3addPiPKi: s_load_b128 s[0:3], s[0:1], 0x0 s_mov_b32 s4, s15 s_ashr_i32 s5, s15, 31 s_delay_alu instid0(SALU_CYCLE_1) s_lshl_b64 s[4:5], s[4:5], 2 s_waitcnt lgkmcnt(0) s_add_u32 s2, s2, s4 s_addc_u32 s3, s3, s5 s_add_u32 s0, s0, s4 s_addc_u32 s1, s1, s5 s_load_b32 s2, s[2:3], 0x0 s_load_b32 s3, s[0:1], 0x0 s_waitcnt lgkmcnt(0) s_add_i32 s2, s3, s2 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s2 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3addPiPKi .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 16 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 2 .amdhsa_next_free_sgpr 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 _Z3addPiPKi, .Lfunc_end0-_Z3addPiPKi .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 .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 16 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addPiPKi .private_segment_fixed_size: 0 .sgpr_count: 16 .sgpr_spill_count: 0 .symbol: _Z3addPiPKi.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_00152e3b_00000000-6_test.cudafe1.cpp" .text #APP .globl _ZSt21ios_base_library_initv #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3672: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3672: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z25__device_stub__Z3addPiPKiPiPKi .type _Z25__device_stub__Z3addPiPKiPiPKi, @function _Z25__device_stub__Z3addPiPKiPiPKi: .LFB3694: .cfi_startproc endbr64 subq $120, %rsp .cfi_def_cfa_offset 128 movq %rdi, 8(%rsp) movq %rsi, (%rsp) movq %fs:40, %rax movq %rax, 104(%rsp) xorl %eax, %eax leaq 8(%rsp), %rax movq %rax, 80(%rsp) movq %rsp, %rax movq %rax, 88(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) leaq 24(%rsp), %rcx leaq 16(%rsp), %rdx leaq 44(%rsp), %rsi leaq 32(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 104(%rsp), %rax subq %fs:40, %rax jne .L8 addq $120, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 24(%rsp) .cfi_def_cfa_offset 136 pushq 24(%rsp) .cfi_def_cfa_offset 144 leaq 96(%rsp), %r9 movq 60(%rsp), %rcx movl 68(%rsp), %r8d movq 48(%rsp), %rsi movl 56(%rsp), %edx leaq _Z3addPiPKi(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 128 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE3694: .size _Z25__device_stub__Z3addPiPKiPiPKi, .-_Z25__device_stub__Z3addPiPKiPiPKi .globl _Z3addPiPKi .type _Z3addPiPKi, @function _Z3addPiPKi: .LFB3695: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z25__device_stub__Z3addPiPKiPiPKi addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3695: .size _Z3addPiPKi, .-_Z3addPiPKi .globl main .type main, @function main: .LFB3669: .cfi_startproc endbr64 pushq %r15 .cfi_def_cfa_offset 16 .cfi_offset 15, -16 pushq %r14 .cfi_def_cfa_offset 24 .cfi_offset 14, -24 pushq %r13 .cfi_def_cfa_offset 32 .cfi_offset 13, -32 pushq %r12 .cfi_def_cfa_offset 40 .cfi_offset 12, -40 pushq %rbp .cfi_def_cfa_offset 48 .cfi_offset 6, -48 pushq %rbx .cfi_def_cfa_offset 56 .cfi_offset 3, -56 subq $56, %rsp .cfi_def_cfa_offset 112 movq %fs:40, %rax movq %rax, 40(%rsp) xorl %eax, %eax movl $40, %edi call _Znam@PLT movq %rax, %r12 movq %rsp, %rdi movl $40, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $40, %esi call cudaMalloc@PLT movl $0, %eax .L12: movl %eax, (%r12,%rax,4) addq $1, %rax cmpq $10, %rax jne .L12 movl $1, %ecx movl $40, %edx movq %r12, %rsi movq (%rsp), %rdi call cudaMemcpy@PLT movl $0, %eax .L13: leal (%rax,%rax), %edx movl %edx, (%r12,%rax,4) addq $1, %rax cmpq $10, %rax jne .L13 movl $1, %ecx movl $40, %edx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, 28(%rsp) movl $1, 32(%rsp) movl $1, 36(%rsp) movl $10, 16(%rsp) movl $1, 20(%rsp) movl $1, 24(%rsp) movl $0, %r9d movl $0, %r8d movq 28(%rsp), %rdx movl $1, %ecx movq 16(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L25 .L14: movl $2, %ecx movl $40, %edx movq (%rsp), %rsi movq %r12, %rdi call cudaMemcpy@PLT movq %r12, %rbp leaq 40(%r12), %r14 leaq _ZSt4cout(%rip), %r13 jmp .L19 .L25: movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z25__device_stub__Z3addPiPKiPiPKi jmp .L14 .L28: movq 40(%rsp), %rax subq %fs:40, %rax jne .L26 call _ZSt16__throw_bad_castv@PLT .L26: call __stack_chk_fail@PLT .L17: movq %r15, %rdi call _ZNKSt5ctypeIcE13_M_widen_initEv@PLT movq (%r15), %rax movl $10, %esi movq %r15, %rdi call *48(%rax) movl %eax, %esi .L18: movsbl %sil, %esi movq %rbx, %rdi call _ZNSo3putEc@PLT movq %rax, %rdi call _ZNSo5flushEv@PLT addq $4, %rbp cmpq %r14, %rbp je .L27 .L19: movl 0(%rbp), %esi movq %r13, %rdi call _ZNSolsEi@PLT movq %rax, %rbx movq (%rax), %rax movq -24(%rax), %rax movq 240(%rbx,%rax), %r15 testq %r15, %r15 je .L28 cmpb $0, 56(%r15) je .L17 movzbl 67(%r15), %esi jmp .L18 .L27: movq %r12, %rdi call _ZdaPv@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 40(%rsp), %rax subq %fs:40, %rax jne .L29 movl $0, %eax 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 .L29: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE3669: .size main, .-main .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z3addPiPKi" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3697: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC0(%rip), %rdx movq %rdx, %rcx leaq _Z3addPiPKi(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE3697: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
.text .file "test.hip" # Start of file scope inline assembly .globl _ZSt21ios_base_library_initv # End of file scope inline assembly .globl _Z18__device_stub__addPiPKi # -- Begin function _Z18__device_stub__addPiPKi .p2align 4, 0x90 .type _Z18__device_stub__addPiPKi,@function _Z18__device_stub__addPiPKi: # @_Z18__device_stub__addPiPKi .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 $_Z3addPiPKi, %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 _Z18__device_stub__addPiPKi, .Lfunc_end0-_Z18__device_stub__addPiPKi .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: pushq %r15 .cfi_def_cfa_offset 16 pushq %r14 .cfi_def_cfa_offset 24 pushq %r12 .cfi_def_cfa_offset 32 pushq %rbx .cfi_def_cfa_offset 40 subq $104, %rsp .cfi_def_cfa_offset 144 .cfi_offset %rbx, -40 .cfi_offset %r12, -32 .cfi_offset %r14, -24 .cfi_offset %r15, -16 movl $40, %edi callq _Znam movq %rax, %rbx movq %rsp, %rdi movl $40, %esi callq hipMalloc leaq 8(%rsp), %rdi movl $40, %esi callq hipMalloc xorl %eax, %eax .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,4) incq %rax cmpq $10, %rax jne .LBB1_1 # %bb.2: movq (%rsp), %rdi movl $40, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy xorl %eax, %eax .p2align 4, 0x90 .LBB1_3: # =>This Inner Loop Header: Depth=1 movl %eax, (%rbx,%rax,2) addq $2, %rax cmpq $20, %rax jne .LBB1_3 # %bb.4: movq 8(%rsp), %rdi movl $40, %edx movq %rbx, %rsi movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdx # imm = 0x100000001 leaq 9(%rdx), %rdi movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_6 # %bb.5: movq (%rsp), %rax movq 8(%rsp), %rcx movq %rax, 72(%rsp) movq %rcx, 64(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 48(%rsp), %rdi leaq 32(%rsp), %rsi leaq 24(%rsp), %rdx leaq 16(%rsp), %rcx callq __hipPopCallConfiguration movq 48(%rsp), %rsi movl 56(%rsp), %edx movq 32(%rsp), %rcx movl 40(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z3addPiPKi, %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_6: movq (%rsp), %rsi movl $40, %edx movq %rbx, %rdi movl $2, %ecx callq hipMemcpy xorl %r15d, %r15d jmp .LBB1_7 .p2align 4, 0x90 .LBB1_10: # in Loop: Header=BB1_7 Depth=1 movq %r14, %rdi movq %rax, %r12 callq _ZNKSt5ctypeIcE13_M_widen_initEv movq (%r14), %rax movq %r14, %rdi movl $10, %esi callq *48(%rax) movl %eax, %ecx movq %r12, %rax .LBB1_11: # %_ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_.exit # in Loop: Header=BB1_7 Depth=1 movsbl %cl, %esi movq %rax, %rdi callq _ZNSo3putEc movq %rax, %rdi callq _ZNSo5flushEv incq %r15 cmpq $10, %r15 je .LBB1_12 .LBB1_7: # =>This Inner Loop Header: Depth=1 movl (%rbx,%r15,4), %esi movl $_ZSt4cout, %edi callq _ZNSolsEi movq (%rax), %rcx movq -24(%rcx), %rcx movq 240(%rax,%rcx), %r14 testq %r14, %r14 je .LBB1_13 # %bb.8: # %_ZSt13__check_facetISt5ctypeIcEERKT_PS3_.exit.i.i # in Loop: Header=BB1_7 Depth=1 cmpb $0, 56(%r14) je .LBB1_10 # %bb.9: # in Loop: Header=BB1_7 Depth=1 movzbl 67(%r14), %ecx jmp .LBB1_11 .LBB1_12: movq %rbx, %rdi callq _ZdaPv movq (%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree xorl %eax, %eax addq $104, %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 .LBB1_13: .cfi_def_cfa_offset 144 callq _ZSt16__throw_bad_castv .Lfunc_end1: .size main, .Lfunc_end1-main .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_ctor .type __hip_module_ctor,@function __hip_module_ctor: # @__hip_module_ctor .cfi_startproc # %bb.0: subq $40, %rsp .cfi_def_cfa_offset 48 cmpq $0, __hip_gpubin_handle(%rip) jne .LBB2_2 # %bb.1: movl $__hip_fatbin_wrapper, %edi callq __hipRegisterFatBinary movq %rax, __hip_gpubin_handle(%rip) .LBB2_2: movq __hip_gpubin_handle(%rip), %rdi xorps %xmm0, %xmm0 movups %xmm0, 16(%rsp) movups %xmm0, (%rsp) movl $_Z3addPiPKi, %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 _Z3addPiPKi,@object # @_Z3addPiPKi .section .rodata,"a",@progbits .globl _Z3addPiPKi .p2align 3, 0x0 _Z3addPiPKi: .quad _Z18__device_stub__addPiPKi .size _Z3addPiPKi, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z3addPiPKi" .size .L__unnamed_1, 12 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z18__device_stub__addPiPKi .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z3addPiPKi .addrsig_sym _ZSt4cout .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" extern "C" { } #define TB 256 #define EPS 0.1 #undef MIN #define MIN(a, b) ((a) < (b) ? (a) : (b)) #undef MAX #define MAX(a, b) ((a) > (b) ? (a) : (b)) __global__ void avg_vote_kernel( float *A, float *B, int *corrAB, int patch, int c, int h, int w ) { int _id = blockIdx.x * blockDim.x + threadIdx.x; int size = h * w; int radius = patch / 2; if (_id < c * size) { // _id = dc * size + id int id = _id % size, dc = _id / size; int x1 = id % w, y1 = id / w; double sum = 0; int cnt = 0; for (int dx = -radius; dx <= radius; dx++) { for (int dy = -radius; dy <= radius; dy++) { int new_x1 = x1 + dx, new_y1 = y1 + dy; if (new_x1 >= 0 && new_x1 < w && new_y1 >= 0 && new_y1 < h) { int new_id1 = new_y1 * w + new_x1; int x2 = corrAB[new_id1 * 2 + 0]; int y2 = corrAB[new_id1 * 2 + 1]; int new_x2 = x2 - dx, new_y2 = y2 - dy; if (new_x2 >= 0 && new_x2 < w && new_y2 >= 0 && new_y2 < h) { int new_id2 = new_y2 * w + new_x2; sum += A[dc * size + new_id2]; cnt++; } } } } if (cnt != 0) B[dc * size + id] = sum / cnt; } return ; }
.file "tmpxft_0002a7e3_00000000-6_avg_vote_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 _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii .type _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii, @function _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii: .LFB2051: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 20(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z15avg_vote_kernelPfS_Piiiii(%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 _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii, .-_Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii .globl _Z15avg_vote_kernelPfS_Piiiii .type _Z15avg_vote_kernelPfS_Piiiii, @function _Z15avg_vote_kernelPfS_Piiiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z15avg_vote_kernelPfS_Piiiii, .-_Z15avg_vote_kernelPfS_Piiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z15avg_vote_kernelPfS_Piiiii" .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 _Z15avg_vote_kernelPfS_Piiiii(%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" extern "C" { } #define TB 256 #define EPS 0.1 #undef MIN #define MIN(a, b) ((a) < (b) ? (a) : (b)) #undef MAX #define MAX(a, b) ((a) > (b) ? (a) : (b)) __global__ void avg_vote_kernel( float *A, float *B, int *corrAB, int patch, int c, int h, int w ) { int _id = blockIdx.x * blockDim.x + threadIdx.x; int size = h * w; int radius = patch / 2; if (_id < c * size) { // _id = dc * size + id int id = _id % size, dc = _id / size; int x1 = id % w, y1 = id / w; double sum = 0; int cnt = 0; for (int dx = -radius; dx <= radius; dx++) { for (int dy = -radius; dy <= radius; dy++) { int new_x1 = x1 + dx, new_y1 = y1 + dy; if (new_x1 >= 0 && new_x1 < w && new_y1 >= 0 && new_y1 < h) { int new_id1 = new_y1 * w + new_x1; int x2 = corrAB[new_id1 * 2 + 0]; int y2 = corrAB[new_id1 * 2 + 1]; int new_x2 = x2 - dx, new_y2 = y2 - dy; if (new_x2 >= 0 && new_x2 < w && new_y2 >= 0 && new_y2 < h) { int new_id2 = new_y2 * w + new_x2; sum += A[dc * size + new_id2]; cnt++; } } } } if (cnt != 0) B[dc * size + id] = sum / cnt; } return ; }
#include <hip/hip_runtime.h> #include "includes.h" extern "C" { } #define TB 256 #define EPS 0.1 #undef MIN #define MIN(a, b) ((a) < (b) ? (a) : (b)) #undef MAX #define MAX(a, b) ((a) > (b) ? (a) : (b)) __global__ void avg_vote_kernel( float *A, float *B, int *corrAB, int patch, int c, int h, int w ) { int _id = blockIdx.x * blockDim.x + threadIdx.x; int size = h * w; int radius = patch / 2; if (_id < c * size) { // _id = dc * size + id int id = _id % size, dc = _id / size; int x1 = id % w, y1 = id / w; double sum = 0; int cnt = 0; for (int dx = -radius; dx <= radius; dx++) { for (int dy = -radius; dy <= radius; dy++) { int new_x1 = x1 + dx, new_y1 = y1 + dy; if (new_x1 >= 0 && new_x1 < w && new_y1 >= 0 && new_y1 < h) { int new_id1 = new_y1 * w + new_x1; int x2 = corrAB[new_id1 * 2 + 0]; int y2 = corrAB[new_id1 * 2 + 1]; int new_x2 = x2 - dx, new_y2 = y2 - dy; if (new_x2 >= 0 && new_x2 < w && new_y2 >= 0 && new_y2 < h) { int new_id2 = new_y2 * w + new_x2; sum += A[dc * size + new_id2]; cnt++; } } } } if (cnt != 0) B[dc * size + id] = sum / cnt; } return ; }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
#include <hip/hip_runtime.h> #include "includes.h" extern "C" { } #define TB 256 #define EPS 0.1 #undef MIN #define MIN(a, b) ((a) < (b) ? (a) : (b)) #undef MAX #define MAX(a, b) ((a) > (b) ? (a) : (b)) __global__ void avg_vote_kernel( float *A, float *B, int *corrAB, int patch, int c, int h, int w ) { int _id = blockIdx.x * blockDim.x + threadIdx.x; int size = h * w; int radius = patch / 2; if (_id < c * size) { // _id = dc * size + id int id = _id % size, dc = _id / size; int x1 = id % w, y1 = id / w; double sum = 0; int cnt = 0; for (int dx = -radius; dx <= radius; dx++) { for (int dy = -radius; dy <= radius; dy++) { int new_x1 = x1 + dx, new_y1 = y1 + dy; if (new_x1 >= 0 && new_x1 < w && new_y1 >= 0 && new_y1 < h) { int new_id1 = new_y1 * w + new_x1; int x2 = corrAB[new_id1 * 2 + 0]; int y2 = corrAB[new_id1 * 2 + 1]; int new_x2 = x2 - dx, new_y2 = y2 - dy; if (new_x2 >= 0 && new_x2 < w && new_y2 >= 0 && new_y2 < h) { int new_id2 = new_y2 * w + new_x2; sum += A[dc * size + new_id2]; cnt++; } } } } if (cnt != 0) B[dc * size + id] = sum / cnt; } return ; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z15avg_vote_kernelPfS_Piiiii .globl _Z15avg_vote_kernelPfS_Piiiii .p2align 8 .type _Z15avg_vote_kernelPfS_Piiiii,@function _Z15avg_vote_kernelPfS_Piiiii: s_clause 0x2 s_load_b32 s2, s[0:1], 0x34 s_load_b64 s[6:7], s[0:1], 0x1c s_load_b32 s13, s[0:1], 0x24 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_mad_u64_u32 v[3:4], null, s15, s2, v[0:1] s_mul_i32 s12, s13, s7 s_mul_i32 s2, s12, s6 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) v_cmp_gt_i32_e32 vcc_lo, s2, v3 s_and_saveexec_b32 s2, vcc_lo s_cbranch_execz .LBB0_17 s_ashr_i32 s2, s12, 31 v_ashrrev_i32_e32 v2, 31, v3 s_add_i32 s3, s12, s2 v_mov_b32_e32 v7, 0 s_xor_b32 s3, s3, s2 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_cvt_f32_u32_e32 v0, s3 s_sub_i32 s4, 0, s3 v_add_nc_u32_e32 v4, v3, v2 v_rcp_iflag_f32_e32 v0, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_xor_b32_e32 v4, v4, v2 v_xor_b32_e32 v2, s2, v2 s_load_b32 s2, s[0:1], 0x18 s_waitcnt_depctr 0xfff v_mul_f32_e32 v0, 0x4f7ffffe, v0 v_cvt_u32_f32_e32 v0, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_1) v_mul_lo_u32 v1, s4, v0 s_waitcnt lgkmcnt(0) s_cmp_lt_i32 s2, -1 v_mul_hi_u32 v1, v0, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v0, v0, v1 v_mul_hi_u32 v0, v4, v0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_lo_u32 v1, v0, s3 v_sub_nc_u32_e32 v1, v4, v1 v_add_nc_u32_e32 v4, 1, v0 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_subrev_nc_u32_e32 v5, s3, v1 v_cmp_le_u32_e32 vcc_lo, s3, v1 v_dual_cndmask_b32 v0, v0, v4 :: v_dual_cndmask_b32 v1, v1, v5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v4, 1, v0 v_cmp_le_u32_e32 vcc_lo, s3, v1 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e32 v0, v0, v4, vcc_lo v_xor_b32_e32 v0, v0, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_3) v_sub_nc_u32_e32 v5, v0, v2 v_mov_b32_e32 v1, 0 v_mov_b32_e32 v2, 0 v_mul_lo_u32 v6, v5, s12 s_delay_alu instid0(VALU_DEP_1) v_sub_nc_u32_e32 v0, v3, v6 s_cbranch_scc1 .LBB0_15 s_ashr_i32 s3, s13, 31 s_delay_alu instid0(VALU_DEP_1) v_ashrrev_i32_e32 v4, 31, v0 s_add_i32 s4, s13, s3 s_clause 0x1 s_load_b64 s[8:9], s[0:1], 0x0 s_load_b64 s[10:11], s[0:1], 0x10 s_xor_b32 s4, s4, s3 s_lshl_b32 s16, s13, 1 v_cvt_f32_u32_e32 v1, s4 s_sub_i32 s5, 0, s4 v_add_nc_u32_e32 v7, v0, v4 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_rcp_iflag_f32_e32 v1, v1 v_xor_b32_e32 v7, v7, v4 v_xor_b32_e32 v4, s3, v4 s_not_b32 s3, s13 s_waitcnt_depctr 0xfff v_mul_f32_e32 v1, 0x4f7ffffe, v1 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_cvt_u32_f32_e32 v1, v1 v_mul_lo_u32 v2, s5, v1 s_lshr_b32 s5, s2, 31 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s2, s2, s5 s_ashr_i32 s6, s2, 1 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) s_abs_i32 s14, s6 v_mul_hi_u32 v2, v1, v2 s_sub_i32 s17, 0, s6 s_not_b32 s15, s14 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v8, v1, v2 v_mad_u64_u32 v[1:2], null, v7, v8, 0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_mul_lo_u32 v1, v2, s4 v_sub_nc_u32_e32 v1, v7, v1 v_add_nc_u32_e32 v7, 1, v2 s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(VALU_DEP_2) v_subrev_nc_u32_e32 v8, s4, v1 v_cmp_le_u32_e32 vcc_lo, s4, v1 v_dual_cndmask_b32 v2, v2, v7 :: v_dual_cndmask_b32 v1, v1, v8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_nc_u32_e32 v7, 1, v2 v_cmp_le_u32_e32 vcc_lo, s4, v1 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_cndmask_b32_e32 v1, v2, v7, vcc_lo v_xor_b32_e32 v7, v1, v4 v_mad_u64_u32 v[1:2], null, s6, s3, v[3:4] s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2) v_sub_nc_u32_e32 v3, v7, v4 v_sub_nc_u32_e32 v7, v1, v6 s_delay_alu instid0(VALU_DEP_2) v_mul_lo_u32 v4, v3, s13 v_mov_b32_e32 v1, 0 v_mov_b32_e32 v2, 0 v_subrev_nc_u32_e32 v9, s6, v3 v_lshl_or_b32 v8, v7, 1, 1 v_mov_b32_e32 v7, 0 v_sub_nc_u32_e32 v10, v0, v4 .LBB0_3: s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_2) | instid1(VALU_DEP_2) v_dual_mov_b32 v3, v8 :: v_dual_add_nc_u32 v4, s17, v10 v_mov_b32_e32 v11, v9 s_mov_b32 s18, s6 v_cmp_lt_i32_e32 vcc_lo, -1, v4 v_cmp_gt_i32_e64 s2, s13, v4 s_branch .LBB0_8 .LBB0_4: s_or_b32 exec_lo, exec_lo, s4 .LBB0_5: s_delay_alu instid0(SALU_CYCLE_1) s_or_b32 exec_lo, exec_lo, s21 .LBB0_6: s_delay_alu instid0(SALU_CYCLE_1) s_or_b32 exec_lo, exec_lo, s20 .LBB0_7: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(SALU_CYCLE_1) s_or_b32 exec_lo, exec_lo, s19 v_add_nc_u32_e32 v11, 1, v11 v_add_nc_u32_e32 v3, s16, v3 s_add_i32 s18, s18, -1 s_cmp_eq_u32 s15, s18 s_cbranch_scc1 .LBB0_13 .LBB0_8: s_and_saveexec_b32 s19, vcc_lo s_cbranch_execz .LBB0_7 v_cmp_lt_i32_e64 s3, -1, v11 v_cmp_gt_i32_e64 s4, s7, v11 s_delay_alu instid0(VALU_DEP_2) s_and_b32 s3, s2, s3 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) s_and_b32 s3, s3, s4 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s20, s3 s_cbranch_execz .LBB0_6 v_add_nc_u32_e32 v12, -1, v3 s_mov_b32 s21, exec_lo s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v13, 31, v12 v_lshlrev_b64 v[12:13], 2, v[12:13] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add_co_u32 v12, s3, s10, v12 v_add_co_ci_u32_e64 v13, s3, s11, v13, s3 global_load_b32 v4, v[12:13], off s_waitcnt vmcnt(0) v_subrev_nc_u32_e32 v12, s17, v4 s_delay_alu instid0(VALU_DEP_1) v_cmpx_lt_i32_e32 -1, v12 s_cbranch_execz .LBB0_5 v_ashrrev_i32_e32 v4, 31, v3 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[13:14], 2, v[3:4] v_add_co_u32 v13, s3, s10, v13 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v14, s3, s11, v14, s3 v_cmp_gt_i32_e64 s3, s13, v12 global_load_b32 v4, v[13:14], off s_waitcnt vmcnt(0) v_add_nc_u32_e32 v4, s18, v4 v_cmp_gt_i32_e64 s4, s7, v4 v_cmp_lt_i32_e64 s5, -1, v4 s_delay_alu instid0(VALU_DEP_2) s_and_b32 s3, s3, s4 s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1) s_and_b32 s3, s3, s5 s_delay_alu instid0(SALU_CYCLE_1) s_and_saveexec_b32 s4, s3 s_cbranch_execz .LBB0_4 v_mul_lo_u32 v4, v4, s13 v_add_nc_u32_e32 v7, 1, v7 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1) v_add3_u32 v12, v12, v6, v4 v_ashrrev_i32_e32 v13, 31, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_lshlrev_b64 v[12:13], 2, v[12:13] v_add_co_u32 v12, s3, s8, v12 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_co_ci_u32_e64 v13, s3, s9, v13, s3 global_load_b32 v4, v[12:13], off s_waitcnt vmcnt(0) v_cvt_f64_f32_e32 v[12:13], v4 v_add_f64 v[1:2], v[1:2], v[12:13] s_branch .LBB0_4 .LBB0_13: v_add_nc_u32_e32 v8, 2, v8 s_add_i32 s2, s17, 1 s_cmp_eq_u32 s17, s14 s_cbranch_scc1 .LBB0_15 s_mov_b32 s17, s2 s_branch .LBB0_3 .LBB0_15: v_cmp_ne_u32_e32 vcc_lo, 0, v7 s_and_b32 exec_lo, exec_lo, vcc_lo s_cbranch_execz .LBB0_17 v_cvt_f64_i32_e32 v[3:4], v7 s_load_b64 s[0:1], s[0:1], 0x8 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_scale_f64 v[6:7], null, v[3:4], v[3:4], v[1:2] v_rcp_f64_e32 v[8:9], v[6:7] s_waitcnt_depctr 0xfff v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] v_fma_f64 v[10:11], -v[6:7], v[8:9], 1.0 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_fma_f64 v[8:9], v[8:9], v[10:11], v[8:9] v_div_scale_f64 v[10:11], vcc_lo, v[1:2], v[3:4], v[1:2] v_mul_f64 v[12:13], v[10:11], v[8:9] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_fma_f64 v[6:7], -v[6:7], v[12:13], v[10:11] v_div_fmas_f64 v[6:7], v[6:7], v[8:9], v[12:13] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_div_fixup_f64 v[1:2], v[6:7], v[3:4], v[1:2] v_cvt_f32_f64_e32 v3, v[1:2] v_mad_u64_u32 v[1:2], null, v5, s12, v[0:1] s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v2, 31, v1 v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v0, vcc_lo, s0, v0 v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v3, off .LBB0_17: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z15avg_vote_kernelPfS_Piiiii .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 296 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 22 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z15avg_vote_kernelPfS_Piiiii, .Lfunc_end0-_Z15avg_vote_kernelPfS_Piiiii .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .offset: 36 .size: 4 .value_kind: by_value - .offset: 40 .size: 4 .value_kind: hidden_block_count_x - .offset: 44 .size: 4 .value_kind: hidden_block_count_y - .offset: 48 .size: 4 .value_kind: hidden_block_count_z - .offset: 52 .size: 2 .value_kind: hidden_group_size_x - .offset: 54 .size: 2 .value_kind: hidden_group_size_y - .offset: 56 .size: 2 .value_kind: hidden_group_size_z - .offset: 58 .size: 2 .value_kind: hidden_remainder_x - .offset: 60 .size: 2 .value_kind: hidden_remainder_y - .offset: 62 .size: 2 .value_kind: hidden_remainder_z - .offset: 80 .size: 8 .value_kind: hidden_global_offset_x - .offset: 88 .size: 8 .value_kind: hidden_global_offset_y - .offset: 96 .size: 8 .value_kind: hidden_global_offset_z - .offset: 104 .size: 2 .value_kind: hidden_grid_dims .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: _Z15avg_vote_kernelPfS_Piiiii .private_segment_fixed_size: 0 .sgpr_count: 24 .sgpr_spill_count: 0 .symbol: _Z15avg_vote_kernelPfS_Piiiii.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" extern "C" { } #define TB 256 #define EPS 0.1 #undef MIN #define MIN(a, b) ((a) < (b) ? (a) : (b)) #undef MAX #define MAX(a, b) ((a) > (b) ? (a) : (b)) __global__ void avg_vote_kernel( float *A, float *B, int *corrAB, int patch, int c, int h, int w ) { int _id = blockIdx.x * blockDim.x + threadIdx.x; int size = h * w; int radius = patch / 2; if (_id < c * size) { // _id = dc * size + id int id = _id % size, dc = _id / size; int x1 = id % w, y1 = id / w; double sum = 0; int cnt = 0; for (int dx = -radius; dx <= radius; dx++) { for (int dy = -radius; dy <= radius; dy++) { int new_x1 = x1 + dx, new_y1 = y1 + dy; if (new_x1 >= 0 && new_x1 < w && new_y1 >= 0 && new_y1 < h) { int new_id1 = new_y1 * w + new_x1; int x2 = corrAB[new_id1 * 2 + 0]; int y2 = corrAB[new_id1 * 2 + 1]; int new_x2 = x2 - dx, new_y2 = y2 - dy; if (new_x2 >= 0 && new_x2 < w && new_y2 >= 0 && new_y2 < h) { int new_id2 = new_y2 * w + new_x2; sum += A[dc * size + new_id2]; cnt++; } } } } if (cnt != 0) B[dc * size + id] = sum / cnt; } return ; }
.text .file "avg_vote_kernel.hip" .globl _Z30__device_stub__avg_vote_kernelPfS_Piiiii # -- Begin function _Z30__device_stub__avg_vote_kernelPfS_Piiiii .p2align 4, 0x90 .type _Z30__device_stub__avg_vote_kernelPfS_Piiiii,@function _Z30__device_stub__avg_vote_kernelPfS_Piiiii: # @_Z30__device_stub__avg_vote_kernelPfS_Piiiii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 20(%rsp), %rax movq %rax, 120(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z15avg_vote_kernelPfS_Piiiii, %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 _Z30__device_stub__avg_vote_kernelPfS_Piiiii, .Lfunc_end0-_Z30__device_stub__avg_vote_kernelPfS_Piiiii .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 $_Z15avg_vote_kernelPfS_Piiiii, %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 _Z15avg_vote_kernelPfS_Piiiii,@object # @_Z15avg_vote_kernelPfS_Piiiii .section .rodata,"a",@progbits .globl _Z15avg_vote_kernelPfS_Piiiii .p2align 3, 0x0 _Z15avg_vote_kernelPfS_Piiiii: .quad _Z30__device_stub__avg_vote_kernelPfS_Piiiii .size _Z15avg_vote_kernelPfS_Piiiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z15avg_vote_kernelPfS_Piiiii" .size .L__unnamed_1, 30 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z30__device_stub__avg_vote_kernelPfS_Piiiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z15avg_vote_kernelPfS_Piiiii .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_0002a7e3_00000000-6_avg_vote_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 _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii .type _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii, @function _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii: .LFB2051: .cfi_startproc endbr64 subq $184, %rsp .cfi_def_cfa_offset 192 movq %rdi, 40(%rsp) movq %rsi, 32(%rsp) movq %rdx, 24(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%rsp) movq %fs:40, %rax movq %rax, 168(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 32(%rsp), %rax movq %rax, 120(%rsp) leaq 24(%rsp), %rax movq %rax, 128(%rsp) leaq 20(%rsp), %rax movq %rax, 136(%rsp) leaq 16(%rsp), %rax movq %rax, 144(%rsp) leaq 12(%rsp), %rax movq %rax, 152(%rsp) leaq 192(%rsp), %rax movq %rax, 160(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 168(%rsp), %rax subq %fs:40, %rax jne .L8 addq $184, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 200 pushq 56(%rsp) .cfi_def_cfa_offset 208 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z15avg_vote_kernelPfS_Piiiii(%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 _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii, .-_Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii .globl _Z15avg_vote_kernelPfS_Piiiii .type _Z15avg_vote_kernelPfS_Piiiii, @function _Z15avg_vote_kernelPfS_Piiiii: .LFB2052: .cfi_startproc endbr64 subq $16, %rsp .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z43__device_stub__Z15avg_vote_kernelPfS_PiiiiiPfS_Piiiii addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z15avg_vote_kernelPfS_Piiiii, .-_Z15avg_vote_kernelPfS_Piiiii .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "_Z15avg_vote_kernelPfS_Piiiii" .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 _Z15avg_vote_kernelPfS_Piiiii(%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 "avg_vote_kernel.hip" .globl _Z30__device_stub__avg_vote_kernelPfS_Piiiii # -- Begin function _Z30__device_stub__avg_vote_kernelPfS_Piiiii .p2align 4, 0x90 .type _Z30__device_stub__avg_vote_kernelPfS_Piiiii,@function _Z30__device_stub__avg_vote_kernelPfS_Piiiii: # @_Z30__device_stub__avg_vote_kernelPfS_Piiiii .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 88(%rsp) movq %rsi, 80(%rsp) movq %rdx, 72(%rsp) movl %ecx, 20(%rsp) movl %r8d, 16(%rsp) movl %r9d, 12(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 20(%rsp), %rax movq %rax, 120(%rsp) leaq 16(%rsp), %rax movq %rax, 128(%rsp) leaq 12(%rsp), %rax movq %rax, 136(%rsp) leaq 160(%rsp), %rax movq %rax, 144(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z15avg_vote_kernelPfS_Piiiii, %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 _Z30__device_stub__avg_vote_kernelPfS_Piiiii, .Lfunc_end0-_Z30__device_stub__avg_vote_kernelPfS_Piiiii .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 $_Z15avg_vote_kernelPfS_Piiiii, %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 _Z15avg_vote_kernelPfS_Piiiii,@object # @_Z15avg_vote_kernelPfS_Piiiii .section .rodata,"a",@progbits .globl _Z15avg_vote_kernelPfS_Piiiii .p2align 3, 0x0 _Z15avg_vote_kernelPfS_Piiiii: .quad _Z30__device_stub__avg_vote_kernelPfS_Piiiii .size _Z15avg_vote_kernelPfS_Piiiii, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z15avg_vote_kernelPfS_Piiiii" .size .L__unnamed_1, 30 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z30__device_stub__avg_vote_kernelPfS_Piiiii .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z15avg_vote_kernelPfS_Piiiii .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.
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #include "DropoutWithIndex.cuh" #include "../XDevice.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA __global__ /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> tData - the data pointer of the target tensor >> sIndex - mask index >> size - the size of the sIndex */ void KernelDropoutWithIndex1D(DTYPE * tData, int * sIndex, int size) { /* block id */ int i = blockDim.x * blockIdx.x + threadIdx.x; DTYPE * t = tData; if (i < size) { int id = sIndex[i]; t[id] = DTYPE(0.0F); } } /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> x - input tensor >> maskIndex - mask index tensor >> c - output tensor */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c) { int devID = c->devID; int blockNum = maskIndex->unitNum; int cudaGrids[3]; int cudaBlocks[3]; int devIDBackup; ProtectCudaDev(devID, devIDBackup); GDevs.GetCudaThread(devID, blockNum, cudaGrids, cudaBlocks); dim3 blocks(cudaGrids[0]); dim3 threads(cudaBlocks[0]); DTYPE * tData = (DTYPE*)c->data; int * sIndex = NULL; sIndex = (int *)maskIndex->data; KernelDropoutWithIndex1D <<<blocks, threads >>>(tData, sIndex, blockNum); BacktoCudaDev(devID, devIDBackup); } #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor)
code for sm_80
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #include "DropoutWithIndex.cuh" #include "../XDevice.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA __global__ /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> tData - the data pointer of the target tensor >> sIndex - mask index >> size - the size of the sIndex */ void KernelDropoutWithIndex1D(DTYPE * tData, int * sIndex, int size) { /* block id */ int i = blockDim.x * blockIdx.x + threadIdx.x; DTYPE * t = tData; if (i < size) { int id = sIndex[i]; t[id] = DTYPE(0.0F); } } /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> x - input tensor >> maskIndex - mask index tensor >> c - output tensor */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c) { int devID = c->devID; int blockNum = maskIndex->unitNum; int cudaGrids[3]; int cudaBlocks[3]; int devIDBackup; ProtectCudaDev(devID, devIDBackup); GDevs.GetCudaThread(devID, blockNum, cudaGrids, cudaBlocks); dim3 blocks(cudaGrids[0]); dim3 threads(cudaBlocks[0]); DTYPE * tData = (DTYPE*)c->data; int * sIndex = NULL; sIndex = (int *)maskIndex->data; KernelDropoutWithIndex1D <<<blocks, threads >>>(tData, sIndex, blockNum); BacktoCudaDev(devID, devIDBackup); } #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor)
.file "tmpxft_00030f30_00000000-6_DropoutWithIndex.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3094: .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 .LFE3094: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3117: .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 .LFE3117: .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.
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #include "DropoutWithIndex.cuh" #include "../XDevice.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA __global__ /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> tData - the data pointer of the target tensor >> sIndex - mask index >> size - the size of the sIndex */ void KernelDropoutWithIndex1D(DTYPE * tData, int * sIndex, int size) { /* block id */ int i = blockDim.x * blockIdx.x + threadIdx.x; DTYPE * t = tData; if (i < size) { int id = sIndex[i]; t[id] = DTYPE(0.0F); } } /* This is a special implementation of "dropout" to reduce memory with maskIndex. >> x - input tensor >> maskIndex - mask index tensor >> c - output tensor */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c) { int devID = c->devID; int blockNum = maskIndex->unitNum; int cudaGrids[3]; int cudaBlocks[3]; int devIDBackup; ProtectCudaDev(devID, devIDBackup); GDevs.GetCudaThread(devID, blockNum, cudaGrids, cudaBlocks); dim3 blocks(cudaGrids[0]); dim3 threads(cudaBlocks[0]); DTYPE * tData = (DTYPE*)c->data; int * sIndex = NULL; sIndex = (int *)maskIndex->data; KernelDropoutWithIndex1D <<<blocks, threads >>>(tData, sIndex, blockNum); BacktoCudaDev(devID, devIDBackup); } #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor)
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #ifndef __DROPOUTWITHINDEX_CUH__ #define __DROPOUTWITHINDEX_CUH__ #include "../XTensor.h" #include "DropoutWithIndex.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA /* dropout with index (cuda version) */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c); #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor) #endif // __DROPOUTWITHINDEX_CUH__
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #ifndef __DROPOUTWITHINDEX_CUH__ #define __DROPOUTWITHINDEX_CUH__ #include "../XTensor.h" #include "DropoutWithIndex.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA /* dropout with index (cuda version) */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c); #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor) #endif // __DROPOUTWITHINDEX_CUH__
.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.
/* NiuTrans.Tensor - an open-source tensor library * Copyright (C) 2017, Natural Language Processing Lab, Northeastern University. * All rights reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * $Created by: Jiang Yufan (email: jiangyufan2018@outlook.com) 2019-03-20 */ #ifndef __DROPOUTWITHINDEX_CUH__ #define __DROPOUTWITHINDEX_CUH__ #include "../XTensor.h" #include "DropoutWithIndex.h" namespace nts { // namespace nts(NiuTrans.Tensor) #ifdef USE_CUDA /* dropout with index (cuda version) */ void _CudaDropoutWithIndex(const XTensor * x, XTensor * maskIndex, XTensor * c); #endif // USE_CUDA } // namespace nts(NiuTrans.Tensor) #endif // __DROPOUTWITHINDEX_CUH__
.text .file "DropoutWithIndex.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_00030f30_00000000-6_DropoutWithIndex.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB3094: .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 .LFE3094: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB3117: .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 .LFE3117: .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 "DropoutWithIndex.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> #define N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device cudaMalloc((void**)&DA, size); cudaMalloc((void**)&DB, size); cudaMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente cudaMemcpy(DA, HA, size, cudaMemcpyHostToDevice); cudaMemcpy(DB, HB, size, cudaMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) cudaMemcpy(HC, DC, size, cudaMemcpyDeviceToHost); // liberamos la memoria reservada en el device cudaFree(DA); cudaFree(DB); cudaFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} return 0; }
code for sm_80 Function : _Z6VecAddPiS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R6, SR_TID.X ; /* 0x0000000000067919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x001fc800078e0207 */ /*0050*/ IMAD.WIDE R4, R6.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x0c0fe400078e0207 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*0080*/ IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fe200078e0207 */ /*0090*/ IADD3 R9, R2, R5, RZ ; /* 0x0000000502097210 */ /* 0x004fca0007ffe0ff */ /*00a0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include <stdio.h> #define N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device cudaMalloc((void**)&DA, size); cudaMalloc((void**)&DB, size); cudaMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente cudaMemcpy(DA, HA, size, cudaMemcpyHostToDevice); cudaMemcpy(DB, HB, size, cudaMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) cudaMemcpy(HC, DC, size, cudaMemcpyDeviceToHost); // liberamos la memoria reservada en el device cudaFree(DA); cudaFree(DB); cudaFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} return 0; }
.file "tmpxft_00131469_00000000-6_suma-vectores1.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 _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ .type _Z29__device_stub__Z6VecAddPiS_S_PiS_S_, @function _Z29__device_stub__Z6VecAddPiS_S_PiS_S_: .LFB2082: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z6VecAddPiS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z29__device_stub__Z6VecAddPiS_S_PiS_S_, .-_Z29__device_stub__Z6VecAddPiS_S_PiS_S_ .globl _Z6VecAddPiS_S_ .type _Z6VecAddPiS_S_, @function _Z6VecAddPiS_S_: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z6VecAddPiS_S_, .-_Z6VecAddPiS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "error en componente %d\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $4096, %rsp .cfi_def_cfa_offset 4104 orq $0, (%rsp) subq $1976, %rsp .cfi_def_cfa_offset 6080 movq %fs:40, %rax movq %rax, 6056(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $2000, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $2000, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $2000, %esi call cudaMalloc@PLT movl $0, %eax .L12: movl %eax, %edx negl %edx movl %edx, 48(%rsp,%rax,4) leal (%rax,%rax,2), %edx movl %edx, 2048(%rsp,%rax,4) addq $1, %rax cmpq $500, %rax jne .L12 leaq 48(%rsp), %rsi movl $1, %ecx movl $2000, %edx movq (%rsp), %rdi call cudaMemcpy@PLT leaq 2048(%rsp), %rsi movl $1, %ecx movl $2000, %edx movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $500, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 36(%rsp), %rdx movl $1, %ecx movq 24(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L21 .L13: leaq 4048(%rsp), %rdi movl $2, %ecx movl $2000, %edx movq 16(%rsp), %rsi call cudaMemcpy@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movl $0, %edx leaq 4048(%rsp), %rcx .L16: movl 2048(%rsp,%rdx,4), %eax addl 48(%rsp,%rdx,4), %eax cmpl %eax, (%rcx,%rdx,4) jne .L22 addq $1, %rdx cmpq $500, %rdx jne .L16 jmp .L15 .L21: movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ jmp .L13 .L22: leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L15: movq 6056(%rsp), %rax subq %fs:40, %rax jne .L23 movl $0, %eax addq $6072, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z6VecAddPiS_S_" .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 _Z6VecAddPiS_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 .LFE2085: .size _ZL24__sti____cudaRegisterAllv, .-_ZL24__sti____cudaRegisterAllv .section .init_array,"aw" .align 8 .quad _ZL24__sti____cudaRegisterAllv .section .nvFatBinSegment,"aw" .align 8 .type _ZL15__fatDeviceText, @object .size _ZL15__fatDeviceText, 24 _ZL15__fatDeviceText: .long 1180844977 .long 1 .quad fatbinData .quad 0 .local _ZL20__cudaFatCubinHandle .comm _ZL20__cudaFatCubinHandle,8,8 .ident "GCC: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0" .section .note.GNU-stack,"",@progbits .section .note.gnu.property,"a" .align 8 .long 1f - 0f .long 4f - 1f .long 5 0: .string "GNU" 1: .align 8 .long 0xc0000002 .long 3f - 2f 2: .long 0x3 3: .align 8 4:
You are a helpful coding assistant specialized in converting CUDA source code to HIP/AMD source code.
#include <stdio.h> #define N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device cudaMalloc((void**)&DA, size); cudaMalloc((void**)&DB, size); cudaMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente cudaMemcpy(DA, HA, size, cudaMemcpyHostToDevice); cudaMemcpy(DB, HB, size, cudaMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) cudaMemcpy(HC, DC, size, cudaMemcpyDeviceToHost); // liberamos la memoria reservada en el device cudaFree(DA); cudaFree(DB); cudaFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} return 0; }
#include <hip/hip_runtime.h> #include <stdio.h> #define N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device hipMalloc((void**)&DA, size); hipMalloc((void**)&DB, size); hipMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente hipMemcpy(DA, HA, size, hipMemcpyHostToDevice); hipMemcpy(DB, HB, size, hipMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) hipMemcpy(HC, DC, size, hipMemcpyDeviceToHost); // liberamos la memoria reservada en el device hipFree(DA); hipFree(DB); hipFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} 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 N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device hipMalloc((void**)&DA, size); hipMalloc((void**)&DB, size); hipMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente hipMemcpy(DA, HA, size, hipMemcpyHostToDevice); hipMemcpy(DB, HB, size, hipMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) hipMemcpy(HC, DC, size, hipMemcpyDeviceToHost); // liberamos la memoria reservada en el device hipFree(DA); hipFree(DB); hipFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} return 0; }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z6VecAddPiS_S_ .globl _Z6VecAddPiS_S_ .p2align 8 .type _Z6VecAddPiS_S_,@function _Z6VecAddPiS_S_: s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_load_b64 s[0:1], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_clause 0x1 global_load_b32 v1, v0, s[4:5] global_load_b32 v2, v0, s[6:7] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, v2, v1 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z6VecAddPiS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z6VecAddPiS_S_, .Lfunc_end0-_Z6VecAddPiS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z6VecAddPiS_S_ .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z6VecAddPiS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include <stdio.h> #define N 500 __global__ void VecAdd(int* DA, int* DB, int* DC) { int i = threadIdx.x; DC[i] = DA[i] + DB[i]; } int main() { int HA[N], HB[N], HC[N]; int *DA, *DB, *DC; int i; int size = N*sizeof(int); // reservamos espacio en la memoria global del device hipMalloc((void**)&DA, size); hipMalloc((void**)&DB, size); hipMalloc((void**)&DC, size); // inicializamos HA y HB for (i=0; i<N; i++) {HA[i]=-i; HB[i] = 3*i;} // copiamos HA y HB del host a DA y DB en el device, respectivamente hipMemcpy(DA, HA, size, hipMemcpyHostToDevice); hipMemcpy(DB, HB, size, hipMemcpyHostToDevice); // llamamos al kernel (1 bloque de N hilos) VecAdd <<<1, N>>>(DA, DB, DC); // N hilos ejecutan el kernel en paralelo // copiamos el resultado, que está en la memoria global del device, (DC) al host (a HC) hipMemcpy(HC, DC, size, hipMemcpyDeviceToHost); // liberamos la memoria reservada en el device hipFree(DA); hipFree(DB); hipFree(DC); // una vez que tenemos los resultados en el host, comprobamos que son correctos // esta comprobación debe quitarse una vez que el programa es correcto (p. ej., para medir el tiempo de ejecución) for (i = 0; i < N; i++) // printf("%d + %d = %d\n",HA[i],HB[i],HC[i]); if (HC[i]!= (HA[i]+HB[i])) {printf("error en componente %d\n", i); break;} return 0; }
.text .file "suma-vectores1.hip" .globl _Z21__device_stub__VecAddPiS_S_ # -- Begin function _Z21__device_stub__VecAddPiS_S_ .p2align 4, 0x90 .type _Z21__device_stub__VecAddPiS_S_,@function _Z21__device_stub__VecAddPiS_S_: # @_Z21__device_stub__VecAddPiS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z6VecAddPiS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z21__device_stub__VecAddPiS_S_, .Lfunc_end0-_Z21__device_stub__VecAddPiS_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $6104, %rsp # imm = 0x17D8 .cfi_def_cfa_offset 6112 leaq 16(%rsp), %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc leaq 8(%rsp), %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc movq %rsp, %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc xorl %eax, %eax xorl %ecx, %ecx xorl %edx, %edx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, 4096(%rsp,%rdx,4) movl %ecx, 2096(%rsp,%rdx,4) incq %rdx addl $3, %ecx decl %eax cmpq $500, %rdx # imm = 0x1F4 jne .LBB1_1 # %bb.2: movq 16(%rsp), %rdi leaq 4096(%rsp), %rsi movl $2000, %edx # imm = 0x7D0 movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi leaq 2096(%rsp), %rsi movl $2000, %edx # imm = 0x7D0 movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdi # imm = 0x100000001 leaq 499(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 16(%rsp), %rax movq 8(%rsp), %rcx movq (%rsp), %rdx movq %rax, 88(%rsp) movq %rcx, 80(%rsp) movq %rdx, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z6VecAddPiS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: movq (%rsp), %rsi leaq 96(%rsp), %rdi movl $2000, %edx # imm = 0x7D0 movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree xorl %esi, %esi .p2align 4, 0x90 .LBB1_5: # =>This Inner Loop Header: Depth=1 movl 2096(%rsp,%rsi,4), %eax addl 4096(%rsp,%rsi,4), %eax cmpl %eax, 96(%rsp,%rsi,4) jne .LBB1_6 # %bb.7: # in Loop: Header=BB1_5 Depth=1 incq %rsi cmpq $500, %rsi # imm = 0x1F4 jne .LBB1_5 jmp .LBB1_8 .LBB1_6: movl $.L.str, %edi # kill: def $esi killed $esi killed $rsi xorl %eax, %eax callq printf .LBB1_8: # %.loopexit xorl %eax, %eax addq $6104, %rsp # imm = 0x17D8 .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 $_Z6VecAddPiS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z6VecAddPiS_S_,@object # @_Z6VecAddPiS_S_ .section .rodata,"a",@progbits .globl _Z6VecAddPiS_S_ .p2align 3, 0x0 _Z6VecAddPiS_S_: .quad _Z21__device_stub__VecAddPiS_S_ .size _Z6VecAddPiS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "error en componente %d\n" .size .L.str, 24 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z6VecAddPiS_S_" .size .L__unnamed_1, 16 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .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__VecAddPiS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6VecAddPiS_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 : _Z6VecAddPiS_S_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R6, SR_TID.X ; /* 0x0000000000067919 */ /* 0x000e220000002100 */ /*0020*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0030*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0040*/ IMAD.WIDE R2, R6, R7, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x001fc800078e0207 */ /*0050*/ IMAD.WIDE R4, R6.reuse, R7.reuse, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x0c0fe400078e0207 */ /*0060*/ LDG.E R2, [R2.64] ; /* 0x0000000402027981 */ /* 0x000ea8000c1e1900 */ /*0070*/ LDG.E R5, [R4.64] ; /* 0x0000000404057981 */ /* 0x000ea2000c1e1900 */ /*0080*/ IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fe200078e0207 */ /*0090*/ IADD3 R9, R2, R5, RZ ; /* 0x0000000502097210 */ /* 0x004fca0007ffe0ff */ /*00a0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*00b0*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*00c0*/ BRA 0xc0; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*00d0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00e0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*00f0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0100*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0110*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0120*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0130*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0140*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0150*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0160*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0170*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z6VecAddPiS_S_ .globl _Z6VecAddPiS_S_ .p2align 8 .type _Z6VecAddPiS_S_,@function _Z6VecAddPiS_S_: s_load_b128 s[4:7], s[0:1], 0x0 v_lshlrev_b32_e32 v0, 2, v0 s_load_b64 s[0:1], s[0:1], 0x10 s_waitcnt lgkmcnt(0) s_clause 0x1 global_load_b32 v1, v0, s[4:5] global_load_b32 v2, v0, s[6:7] s_waitcnt vmcnt(0) v_add_nc_u32_e32 v1, v2, v1 global_store_b32 v0, v1, s[0:1] s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z6VecAddPiS_S_ .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 24 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 3 .amdhsa_next_free_sgpr 8 .amdhsa_reserve_vcc 0 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z6VecAddPiS_S_, .Lfunc_end0-_Z6VecAddPiS_S_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 24 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z6VecAddPiS_S_ .private_segment_fixed_size: 0 .sgpr_count: 8 .sgpr_spill_count: 0 .symbol: _Z6VecAddPiS_S_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 3 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00131469_00000000-6_suma-vectores1.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 _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ .type _Z29__device_stub__Z6VecAddPiS_S_PiS_S_, @function _Z29__device_stub__Z6VecAddPiS_S_PiS_S_: .LFB2082: .cfi_startproc endbr64 subq $136, %rsp .cfi_def_cfa_offset 144 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movq %fs:40, %rax movq %rax, 120(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 120(%rsp), %rax subq %fs:40, %rax jne .L8 addq $136, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 152 pushq 40(%rsp) .cfi_def_cfa_offset 160 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z6VecAddPiS_S_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 144 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z29__device_stub__Z6VecAddPiS_S_PiS_S_, .-_Z29__device_stub__Z6VecAddPiS_S_PiS_S_ .globl _Z6VecAddPiS_S_ .type _Z6VecAddPiS_S_, @function _Z6VecAddPiS_S_: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z6VecAddPiS_S_, .-_Z6VecAddPiS_S_ .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "error en componente %d\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 subq $4096, %rsp .cfi_def_cfa_offset 4104 orq $0, (%rsp) subq $1976, %rsp .cfi_def_cfa_offset 6080 movq %fs:40, %rax movq %rax, 6056(%rsp) xorl %eax, %eax movq %rsp, %rdi movl $2000, %esi call cudaMalloc@PLT leaq 8(%rsp), %rdi movl $2000, %esi call cudaMalloc@PLT leaq 16(%rsp), %rdi movl $2000, %esi call cudaMalloc@PLT movl $0, %eax .L12: movl %eax, %edx negl %edx movl %edx, 48(%rsp,%rax,4) leal (%rax,%rax,2), %edx movl %edx, 2048(%rsp,%rax,4) addq $1, %rax cmpq $500, %rax jne .L12 leaq 48(%rsp), %rsi movl $1, %ecx movl $2000, %edx movq (%rsp), %rdi call cudaMemcpy@PLT leaq 2048(%rsp), %rsi movl $1, %ecx movl $2000, %edx movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $500, 36(%rsp) movl $1, 40(%rsp) movl $1, 44(%rsp) movl $1, 24(%rsp) movl $1, 28(%rsp) movl $1, 32(%rsp) movl $0, %r9d movl $0, %r8d movq 36(%rsp), %rdx movl $1, %ecx movq 24(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L21 .L13: leaq 4048(%rsp), %rdi movl $2, %ecx movl $2000, %edx movq 16(%rsp), %rsi call cudaMemcpy@PLT movq (%rsp), %rdi call cudaFree@PLT movq 8(%rsp), %rdi call cudaFree@PLT movq 16(%rsp), %rdi call cudaFree@PLT movl $0, %edx leaq 4048(%rsp), %rcx .L16: movl 2048(%rsp,%rdx,4), %eax addl 48(%rsp,%rdx,4), %eax cmpl %eax, (%rcx,%rdx,4) jne .L22 addq $1, %rdx cmpq $500, %rdx jne .L16 jmp .L15 .L21: movq 16(%rsp), %rdx movq 8(%rsp), %rsi movq (%rsp), %rdi call _Z29__device_stub__Z6VecAddPiS_S_PiS_S_ jmp .L13 .L22: leaq .LC0(%rip), %rsi movl $2, %edi movl $0, %eax call __printf_chk@PLT .L15: movq 6056(%rsp), %rax subq %fs:40, %rax jne .L23 movl $0, %eax addq $6072, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L23: .cfi_restore_state call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC1: .string "_Z6VecAddPiS_S_" .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 _Z6VecAddPiS_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 .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 "suma-vectores1.hip" .globl _Z21__device_stub__VecAddPiS_S_ # -- Begin function _Z21__device_stub__VecAddPiS_S_ .p2align 4, 0x90 .type _Z21__device_stub__VecAddPiS_S_,@function _Z21__device_stub__VecAddPiS_S_: # @_Z21__device_stub__VecAddPiS_S_ .cfi_startproc # %bb.0: subq $104, %rsp .cfi_def_cfa_offset 112 movq %rdi, 72(%rsp) movq %rsi, 64(%rsp) movq %rdx, 56(%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 64(%rsp), %rax movq %rax, 88(%rsp) leaq 56(%rsp), %rax movq %rax, 96(%rsp) leaq 40(%rsp), %rdi leaq 24(%rsp), %rsi leaq 16(%rsp), %rdx leaq 8(%rsp), %rcx callq __hipPopCallConfiguration movq 40(%rsp), %rsi movl 48(%rsp), %edx movq 24(%rsp), %rcx movl 32(%rsp), %r8d leaq 80(%rsp), %r9 movl $_Z6VecAddPiS_S_, %edi pushq 8(%rsp) .cfi_adjust_cfa_offset 8 pushq 24(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $120, %rsp .cfi_adjust_cfa_offset -120 retq .Lfunc_end0: .size _Z21__device_stub__VecAddPiS_S_, .Lfunc_end0-_Z21__device_stub__VecAddPiS_S_ .cfi_endproc # -- End function .globl main # -- Begin function main .p2align 4, 0x90 .type main,@function main: # @main .cfi_startproc # %bb.0: subq $6104, %rsp # imm = 0x17D8 .cfi_def_cfa_offset 6112 leaq 16(%rsp), %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc leaq 8(%rsp), %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc movq %rsp, %rdi movl $2000, %esi # imm = 0x7D0 callq hipMalloc xorl %eax, %eax xorl %ecx, %ecx xorl %edx, %edx .p2align 4, 0x90 .LBB1_1: # =>This Inner Loop Header: Depth=1 movl %eax, 4096(%rsp,%rdx,4) movl %ecx, 2096(%rsp,%rdx,4) incq %rdx addl $3, %ecx decl %eax cmpq $500, %rdx # imm = 0x1F4 jne .LBB1_1 # %bb.2: movq 16(%rsp), %rdi leaq 4096(%rsp), %rsi movl $2000, %edx # imm = 0x7D0 movl $1, %ecx callq hipMemcpy movq 8(%rsp), %rdi leaq 2096(%rsp), %rsi movl $2000, %edx # imm = 0x7D0 movl $1, %ecx callq hipMemcpy movabsq $4294967297, %rdi # imm = 0x100000001 leaq 499(%rdi), %rdx movl $1, %esi movl $1, %ecx xorl %r8d, %r8d xorl %r9d, %r9d callq __hipPushCallConfiguration testl %eax, %eax jne .LBB1_4 # %bb.3: movq 16(%rsp), %rax movq 8(%rsp), %rcx movq (%rsp), %rdx movq %rax, 88(%rsp) movq %rcx, 80(%rsp) movq %rdx, 72(%rsp) leaq 88(%rsp), %rax movq %rax, 96(%rsp) leaq 80(%rsp), %rax movq %rax, 104(%rsp) leaq 72(%rsp), %rax movq %rax, 112(%rsp) leaq 56(%rsp), %rdi leaq 40(%rsp), %rsi leaq 32(%rsp), %rdx leaq 24(%rsp), %rcx callq __hipPopCallConfiguration movq 56(%rsp), %rsi movl 64(%rsp), %edx movq 40(%rsp), %rcx movl 48(%rsp), %r8d leaq 96(%rsp), %r9 movl $_Z6VecAddPiS_S_, %edi pushq 24(%rsp) .cfi_adjust_cfa_offset 8 pushq 40(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $16, %rsp .cfi_adjust_cfa_offset -16 .LBB1_4: movq (%rsp), %rsi leaq 96(%rsp), %rdi movl $2000, %edx # imm = 0x7D0 movl $2, %ecx callq hipMemcpy movq 16(%rsp), %rdi callq hipFree movq 8(%rsp), %rdi callq hipFree movq (%rsp), %rdi callq hipFree xorl %esi, %esi .p2align 4, 0x90 .LBB1_5: # =>This Inner Loop Header: Depth=1 movl 2096(%rsp,%rsi,4), %eax addl 4096(%rsp,%rsi,4), %eax cmpl %eax, 96(%rsp,%rsi,4) jne .LBB1_6 # %bb.7: # in Loop: Header=BB1_5 Depth=1 incq %rsi cmpq $500, %rsi # imm = 0x1F4 jne .LBB1_5 jmp .LBB1_8 .LBB1_6: movl $.L.str, %edi # kill: def $esi killed $esi killed $rsi xorl %eax, %eax callq printf .LBB1_8: # %.loopexit xorl %eax, %eax addq $6104, %rsp # imm = 0x17D8 .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 $_Z6VecAddPiS_S_, %esi movl $.L__unnamed_1, %edx movl $.L__unnamed_1, %ecx movl $-1, %r8d xorl %r9d, %r9d callq __hipRegisterFunction movl $__hip_module_dtor, %edi addq $40, %rsp .cfi_def_cfa_offset 8 jmp atexit # TAILCALL .Lfunc_end2: .size __hip_module_ctor, .Lfunc_end2-__hip_module_ctor .cfi_endproc # -- End function .p2align 4, 0x90 # -- Begin function __hip_module_dtor .type __hip_module_dtor,@function __hip_module_dtor: # @__hip_module_dtor .cfi_startproc # %bb.0: movq __hip_gpubin_handle(%rip), %rdi testq %rdi, %rdi je .LBB3_2 # %bb.1: pushq %rax .cfi_def_cfa_offset 16 callq __hipUnregisterFatBinary movq $0, __hip_gpubin_handle(%rip) addq $8, %rsp .cfi_def_cfa_offset 8 .LBB3_2: retq .Lfunc_end3: .size __hip_module_dtor, .Lfunc_end3-__hip_module_dtor .cfi_endproc # -- End function .type _Z6VecAddPiS_S_,@object # @_Z6VecAddPiS_S_ .section .rodata,"a",@progbits .globl _Z6VecAddPiS_S_ .p2align 3, 0x0 _Z6VecAddPiS_S_: .quad _Z21__device_stub__VecAddPiS_S_ .size _Z6VecAddPiS_S_, 8 .type .L.str,@object # @.str .section .rodata.str1.1,"aMS",@progbits,1 .L.str: .asciz "error en componente %d\n" .size .L.str, 24 .type .L__unnamed_1,@object # @0 .L__unnamed_1: .asciz "_Z6VecAddPiS_S_" .size .L__unnamed_1, 16 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .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__VecAddPiS_S_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z6VecAddPiS_S_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
#include "includes.h" __global__ void compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
code for sm_80 Function : _Z17compute_distancesPfiiS_iiiS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.Y ; /* 0x0000000000047919 */ /* 0x000e220000002600 */ /*0020*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*0030*/ MOV R6, c[0x0][0x16c] ; /* 0x00005b0000067a02 */ /* 0x000fe20000000f00 */ /*0040*/ UIADD3 UR4, UR4, -0x1, URZ ; /* 0xffffffff04047890 */ /* 0x000fe2000fffe03f */ /*0050*/ S2R R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e620000002500 */ /*0060*/ ULDC UR5, c[0x0][0x16c] ; /* 0x00005b0000057ab9 */ /* 0x000fe20000000800 */ /*0070*/ MOV R7, c[0x0][0x17c] ; /* 0x00005f0000077a02 */ /* 0x000fe20000000f00 */ /*0080*/ UIMAD UR4, UR4, UR5, URZ ; /* 0x00000005040472a4 */ /* 0x000fe2000f8e023f */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000ea20000002200 */ /*00a0*/ IMAD.SHL.U32 R6, R6, 0x10, RZ ; /* 0x0000001006067824 */ /* 0x000fe200078e00ff */ /*00b0*/ SHF.L.U32 R7, R7, 0x4, RZ ; /* 0x0000000407077819 */ /* 0x000fe200000006ff */ /*00c0*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fe200000001ff */ /*00d0*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000ee20000002100 */ /*00e0*/ ISETP.LE.AND P1, PT, RZ, UR4, PT ; /* 0x00000004ff007c0c */ /* 0x000fc4000bf23270 */ /*00f0*/ SHF.L.U32 R4, R4, 0x4, RZ ; /* 0x0000000404047819 */ /* 0x001fe400000006ff */ /*0100*/ SHF.L.U32 R5, R5, 0x4, RZ ; /* 0x0000000405057819 */ /* 0x002fe400000006ff */ /*0110*/ IADD3 R8, R4.reuse, UR4, RZ ; /* 0x0000000404087c10 */ /* 0x040fe2000fffe0ff */ /*0120*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0130*/ IADD3 R2, R4, R3, RZ ; /* 0x0000000304027210 */ /* 0x004fe20007ffe0ff */ /*0140*/ STS.128 [0x800], R4 ; /* 0x00080004ff007388 */ /* 0x0001e60000000c00 */ /*0150*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x168], PT ; /* 0x00005a0002007a0c */ /* 0x000fe20003f06270 */ /*0160*/ STS [0x810], R8 ; /* 0x00081008ff007388 */ /* 0x0001e20000000800 */ /*0170*/ IADD3 R2, R5, R0, RZ ; /* 0x0000000005027210 */ /* 0x008fc80007ffe0ff */ /*0180*/ ISETP.LT.AND P0, PT, R2, c[0x0][0x178], !P0 ; /* 0x00005e0002007a0c */ /* 0x000fe20004701270 */ /*0190*/ @!P1 BRA 0x990 ; /* 0x000007f000009947 */ /* 0x000fd80003800000 */ /*01a0*/ IABS R6, c[0x0][0x16c] ; /* 0x00005b0000067a13 */ /* 0x001fe20000000000 */ /*01b0*/ IMAD.MOV.U32 R15, RZ, RZ, RZ ; /* 0x000000ffff0f7224 */ /* 0x000fe200078e00ff */ /*01c0*/ IADD3 R8, R4, R0, RZ ; /* 0x0000000004087210 */ /* 0x000fe40007ffe0ff */ /*01d0*/ I2F.RP R10, R6 ; /* 0x00000006000a7306 */ /* 0x000e220000209400 */ /*01e0*/ SHF.L.U32 R12, R0, 0x2, RZ ; /* 0x00000002000c7819 */ /* 0x000fe400000006ff */ /*01f0*/ ISETP.GE.AND P1, PT, R8, c[0x0][0x168], PT ; /* 0x00005a0008007a0c */ /* 0x000fe40003f26270 */ /*0200*/ MOV R8, RZ ; /* 0x000000ff00087202 */ /* 0x000fc40000000f00 */ /*0210*/ MOV R7, R5 ; /* 0x0000000500077202 */ /* 0x000fe40000000f00 */ /*0220*/ LEA R5, R3.reuse, R12, 0x6 ; /* 0x0000000c03057211 */ /* 0x040fe200078e30ff */ /*0230*/ MUFU.RCP R10, R10 ; /* 0x0000000a000a7308 */ /* 0x001e240000001000 */ /*0240*/ IADD3 R9, R10, 0xffffffe, RZ ; /* 0x0ffffffe0a097810 */ /* 0x001fe20007ffe0ff */ /*0250*/ IMAD R10, R3, c[0x0][0x17c], R0 ; /* 0x00005f00030a7a24 */ /* 0x000fca00078e0200 */ /*0260*/ F2I.FTZ.U32.TRUNC.NTZ R9, R9 ; /* 0x0000000900097305 */ /* 0x000e24000021f000 */ /*0270*/ IMAD.MOV R11, RZ, RZ, -R9 ; /* 0x000000ffff0b7224 */ /* 0x001fc800078e0a09 */ /*0280*/ IMAD R11, R11, R6, RZ ; /* 0x000000060b0b7224 */ /* 0x000fc800078e02ff */ /*0290*/ IMAD.HI.U32 R8, R9, R11, R8 ; /* 0x0000000b09087227 */ /* 0x000fc800078e0008 */ /*02a0*/ IMAD R9, R3, c[0x0][0x16c], R0 ; /* 0x00005b0003097a24 */ /* 0x000fe400078e0200 */ /*02b0*/ IABS R13, R4 ; /* 0x00000004000d7213 */ /* 0x000fe20000000000 */ /*02c0*/ BSSY B0, 0x4d0 ; /* 0x0000020000007945 */ /* 0x000fe20003800000 */ /*02d0*/ IABS R14, c[0x0][0x16c] ; /* 0x00005b00000e7a13 */ /* 0x002fc60000000000 */ /*02e0*/ IMAD.HI.U32 R11, R8, R13, RZ ; /* 0x0000000d080b7227 */ /* 0x000fca00078e00ff */ /*02f0*/ IADD3 R12, -R11, RZ, RZ ; /* 0x000000ff0b0c7210 */ /* 0x000fca0007ffe1ff */ /*0300*/ IMAD R13, R14, R12, R13 ; /* 0x0000000c0e0d7224 */ /* 0x000fe200078e020d */ /*0310*/ LOP3.LUT R12, R4, c[0x0][0x16c], RZ, 0x3c, !PT ; /* 0x00005b00040c7a12 */ /* 0x000fc800078e3cff */ /*0320*/ ISETP.GT.U32.AND P4, PT, R6, R13, PT ; /* 0x0000000d0600720c */ /* 0x000fe40003f84070 */ /*0330*/ ISETP.GE.AND P3, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x000fd60003f66270 */ /*0340*/ @!P4 IADD3 R13, R13, -R14, RZ ; /* 0x8000000e0d0dc210 */ /* 0x000fe40007ffe0ff */ /*0350*/ @!P4 IADD3 R11, R11, 0x1, RZ ; /* 0x000000010b0bc810 */ /* 0x000fe40007ffe0ff */ /*0360*/ ISETP.GE.U32.AND P2, PT, R13, R6, PT ; /* 0x000000060d00720c */ /* 0x000fe40003f46070 */ /*0370*/ ISETP.NE.AND P4, PT, RZ, c[0x0][0x16c], PT ; /* 0x00005b00ff007a0c */ /* 0x000fd60003f85270 */ /*0380*/ @P2 IADD3 R11, R11, 0x1, RZ ; /* 0x000000010b0b2810 */ /* 0x000fc80007ffe0ff */ /*0390*/ @!P3 IADD3 R11, -R11, RZ, RZ ; /* 0x000000ff0b0bb210 */ /* 0x000fe40007ffe1ff */ /*03a0*/ @!P4 LOP3.LUT R11, RZ, c[0x0][0x16c], RZ, 0x33, !PT ; /* 0x00005b00ff0bca12 */ /* 0x000fc800078e33ff */ /*03b0*/ IADD3 R11, R11, R3, RZ ; /* 0x000000030b0b7210 */ /* 0x000fc80007ffe0ff */ /*03c0*/ ISETP.GE.AND P2, PT, R11, c[0x0][0x180], PT ; /* 0x000060000b007a0c */ /* 0x000fda0003f46270 */ /*03d0*/ @P2 STS [R5], RZ ; /* 0x000000ff05002388 */ /* 0x0011e20000000800 */ /*03e0*/ @P2 IMAD.MOV.U32 R18, RZ, RZ, RZ ; /* 0x000000ffff122224 */ /* 0x000fe200078e00ff */ /*03f0*/ @P2 BRA 0x4c0 ; /* 0x000000c000002947 */ /* 0x000fea0003800000 */ /*0400*/ @!P1 IADD3 R12, R9, R4, RZ ; /* 0x00000004090c9210 */ /* 0x000fe20007ffe0ff */ /*0410*/ HFMA2.MMA R14, -RZ, RZ, 0, 0 ; /* 0x00000000ff0e7435 */ /* 0x000fe200000001ff */ /*0420*/ @!P1 MOV R13, 0x4 ; /* 0x00000004000d9802 */ /* 0x000fe40000000f00 */ /*0430*/ ISETP.GE.AND P2, PT, R2, c[0x0][0x178], PT ; /* 0x00005e0002007a0c */ /* 0x000fc60003f46270 */ /*0440*/ @!P1 IMAD.WIDE R12, R12, R13, c[0x0][0x160] ; /* 0x000058000c0c9625 */ /* 0x000fca00078e020d */ /*0450*/ @!P1 LDG.E R14, [R12.64] ; /* 0x000000040c0e9981 */ /* 0x000ea2000c1e1900 */ /*0460*/ IMAD.MOV.U32 R18, RZ, RZ, RZ ; /* 0x000000ffff127224 */ /* 0x000fc800078e00ff */ /*0470*/ @!P2 IADD3 R16, R10, R7, RZ ; /* 0x000000070a10a210 */ /* 0x000fe40007ffe0ff */ /*0480*/ @!P2 MOV R17, 0x4 ; /* 0x000000040011a802 */ /* 0x000fca0000000f00 */ /*0490*/ @!P2 IMAD.WIDE R16, R16, R17, c[0x0][0x170] ; /* 0x00005c001010a625 */ /* 0x000fca00078e0211 */ /*04a0*/ @!P2 LDG.E R18, [R16.64] ; /* 0x000000041012a981 */ /* 0x000368000c1e1900 */ /*04b0*/ STS [R5], R14 ; /* 0x0000000e05007388 */ /* 0x0043e40000000800 */ /*04c0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*04d0*/ STS [R5+0x400], R18 ; /* 0x0004001205007388 */ /* 0x0205e20000000800 */ /*04e0*/ BSSY B0, 0x920 ; /* 0x0000043000007945 */ /* 0x000fe60003800000 */ /*04f0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0500*/ @!P0 BRA 0x910 ; /* 0x0000040000008947 */ /* 0x000fea0003800000 */ /*0510*/ LDS R13, [R3.X4] ; /* 0x00000000030d7984 */ /* 0x004fe80000004800 */ /*0520*/ LDS R14, [R0.X4+0x400] ; /* 0x00040000000e7984 */ /* 0x002e680000004800 */ /*0530*/ LDS R26, [R0.X4+0x440] ; /* 0x00044000001a7984 */ /* 0x000fe80000004800 */ /*0540*/ LDS R27, [R3.X4+0x40] ; /* 0x00004000031b7984 */ /* 0x000ea80000004800 */ /*0550*/ LDS R17, [R0.X4+0x480] ; /* 0x0004800000117984 */ /* 0x000fe80000004800 */ /*0560*/ LDS R18, [R3.X4+0x80] ; /* 0x0000800003127984 */ /* 0x000ee80000004800 */ /*0570*/ LDS R23, [R0.X4+0x4c0] ; /* 0x0004c00000177984 */ /* 0x000fe80000004800 */ /*0580*/ LDS R24, [R3.X4+0xc0] ; /* 0x0000c00003187984 */ /* 0x000f280000004800 */ /*0590*/ LDS R19, [R0.X4+0x500] ; /* 0x0005000000137984 */ /* 0x000fe80000004800 */ /*05a0*/ LDS R20, [R3.X4+0x100] ; /* 0x0001000003147984 */ /* 0x000f680000004800 */ /*05b0*/ LDS R21, [R0.X4+0x540] ; /* 0x0005400000157984 */ /* 0x000fe80000004800 */ /*05c0*/ LDS R22, [R3.X4+0x140] ; /* 0x0001400003167984 */ /* 0x000e220000004800 */ /*05d0*/ FADD R16, R13, -R14 ; /* 0x8000000e0d107221 */ /* 0x002fc60000000000 */ /*05e0*/ LDS R11, [R0.X4+0x580] ; /* 0x00058000000b7984 */ /* 0x000fe20000004800 */ /*05f0*/ FFMA R25, R16, R16, R15 ; /* 0x0000001010197223 */ /* 0x000fc6000000000f */ /*0600*/ LDS R12, [R3.X4+0x180] ; /* 0x00018000030c7984 */ /* 0x000e620000004800 */ /*0610*/ FADD R26, -R26, R27 ; /* 0x0000001b1a1a7221 */ /* 0x004fc60000000100 */ /*0620*/ LDS R13, [R0.X4+0x5c0] ; /* 0x0005c000000d7984 */ /* 0x000fe20000004800 */ /*0630*/ FFMA R25, R26, R26, R25 ; /* 0x0000001a1a197223 */ /* 0x000fc60000000019 */ /*0640*/ LDS R14, [R3.X4+0x1c0] ; /* 0x0001c000030e7984 */ /* 0x000ea20000004800 */ /*0650*/ FADD R18, -R17, R18 ; /* 0x0000001211127221 */ /* 0x008fc60000000100 */ /*0660*/ LDS R15, [R0.X4+0x600] ; /* 0x00060000000f7984 */ /* 0x000fe20000004800 */ /*0670*/ FFMA R25, R18, R18, R25 ; /* 0x0000001212197223 */ /* 0x000fc60000000019 */ /*0680*/ LDS R16, [R3.X4+0x200] ; /* 0x0002000003107984 */ /* 0x000ee20000004800 */ /*0690*/ FADD R24, -R23, R24 ; /* 0x0000001817187221 */ /* 0x010fc60000000100 */ /*06a0*/ LDS R18, [R0.X4+0x640] ; /* 0x0006400000127984 */ /* 0x000fe20000004800 */ /*06b0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe40000000019 */ /*06c0*/ FADD R24, -R19, R20 ; /* 0x0000001413187221 */ /* 0x020fe20000000100 */ /*06d0*/ LDS R17, [R3.X4+0x240] ; /* 0x0002400003117984 */ /* 0x000f260000004800 */ /*06e0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe20000000019 */ /*06f0*/ LDS R20, [R0.X4+0x680] ; /* 0x0006800000147984 */ /* 0x000fe20000004800 */ /*0700*/ FADD R24, -R21, R22 ; /* 0x0000001615187221 */ /* 0x001fc60000000100 */ /*0710*/ LDS R19, [R3.X4+0x280] ; /* 0x0002800003137984 */ /* 0x000e220000004800 */ /*0720*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*0730*/ LDS R22, [R0.X4+0x6c0] ; /* 0x0006c00000167984 */ /* 0x000fe20000004800 */ /*0740*/ FADD R24, -R11, R12 ; /* 0x0000000c0b187221 */ /* 0x002fc60000000100 */ /*0750*/ LDS R21, [R3.X4+0x2c0] ; /* 0x0002c00003157984 */ /* 0x000e620000004800 */ /*0760*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*0770*/ LDS R11, [R0.X4+0x700] ; /* 0x00070000000b7984 */ /* 0x000fe20000004800 */ /*0780*/ FADD R24, -R13, R14 ; /* 0x0000000e0d187221 */ /* 0x004fc60000000100 */ /*0790*/ LDS R12, [R3.X4+0x300] ; /* 0x00030000030c7984 */ /* 0x000ea20000004800 */ /*07a0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*07b0*/ LDS R14, [R0.X4+0x740] ; /* 0x00074000000e7984 */ /* 0x000fe20000004800 */ /*07c0*/ FADD R26, -R15, R16 ; /* 0x000000100f1a7221 */ /* 0x008fc60000000100 */ /*07d0*/ LDS R13, [R3.X4+0x340] ; /* 0x00034000030d7984 */ /* 0x000ee20000004800 */ /*07e0*/ FFMA R25, R26, R26, R25 ; /* 0x0000001a1a197223 */ /* 0x000fc60000000019 */ /*07f0*/ LDS R24, [R0.X4+0x780] ; /* 0x0007800000187984 */ /* 0x000fe80000004800 */ /*0800*/ LDS R23, [R3.X4+0x380] ; /* 0x0003800003177984 */ /* 0x000f620000004800 */ /*0810*/ FADD R18, -R18, R17 ; /* 0x0000001112127221 */ /* 0x010fc60000000100 */ /*0820*/ LDS R16, [R0.X4+0x7c0] ; /* 0x0007c00000107984 */ /* 0x000fe20000004800 */ /*0830*/ FFMA R25, R18, R18, R25 ; /* 0x0000001212197223 */ /* 0x000fc60000000019 */ /*0840*/ LDS R15, [R3.X4+0x3c0] ; /* 0x0003c000030f7984 */ /* 0x000f220000004800 */ /*0850*/ FADD R20, -R20, R19 ; /* 0x0000001314147221 */ /* 0x001fc80000000100 */ /*0860*/ FFMA R25, R20, R20, R25 ; /* 0x0000001414197223 */ /* 0x000fe40000000019 */ /*0870*/ FADD R22, -R22, R21 ; /* 0x0000001516167221 */ /* 0x002fc80000000100 */ /*0880*/ FFMA R25, R22, R22, R25 ; /* 0x0000001616197223 */ /* 0x000fe40000000019 */ /*0890*/ FADD R12, -R11, R12 ; /* 0x0000000c0b0c7221 */ /* 0x004fc80000000100 */ /*08a0*/ FFMA R25, R12, R12, R25 ; /* 0x0000000c0c197223 */ /* 0x000fe40000000019 */ /*08b0*/ FADD R14, -R14, R13 ; /* 0x0000000d0e0e7221 */ /* 0x008fc80000000100 */ /*08c0*/ FFMA R25, R14, R14, R25 ; /* 0x0000000e0e197223 */ /* 0x000fe40000000019 */ /*08d0*/ FADD R24, -R24, R23 ; /* 0x0000001718187221 */ /* 0x020fc80000000100 */ /*08e0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe40000000019 */ /*08f0*/ FADD R16, -R16, R15 ; /* 0x0000000f10107221 */ /* 0x010fc80000000100 */ /*0900*/ FFMA R15, R16, R16, R25 ; /* 0x00000010100f7223 */ /* 0x000fe40000000019 */ /*0910*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x004fea0003800000 */ /*0920*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0930*/ LDS.64 R12, [0x808] ; /* 0x00080800ff0c7984 */ /* 0x000ea80000000a00 */ /*0940*/ LDS R11, [0x810] ; /* 0x00081000ff0b7984 */ /* 0x000ee20000000800 */ /*0950*/ IADD3 R4, R12, R4, RZ ; /* 0x000000040c047210 */ /* 0x004fc40007ffe0ff */ /*0960*/ IADD3 R7, R7, R13, RZ ; /* 0x0000000d07077210 */ /* 0x000fe40007ffe0ff */ /*0970*/ ISETP.GT.AND P2, PT, R4, R11, PT ; /* 0x0000000b0400720c */ /* 0x008fda0003f44270 */ /*0980*/ @!P2 BRA 0x2b0 ; /* 0xfffff9200000a947 */ /* 0x000fea000383ffff */ /*0990*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*09a0*/ LDS.64 R4, [0x800] ; /* 0x00080000ff047984 */ /* 0x003e220000000a00 */ /*09b0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fe200000001ff */ /*09c0*/ IADD3 R3, R3, R4, RZ ; /* 0x0000000403037210 */ /* 0x001fe40007ffe0ff */ /*09d0*/ IADD3 R0, R0, R5, RZ ; /* 0x0000000500007210 */ /* 0x000fca0007ffe0ff */ /*09e0*/ IMAD R3, R3, c[0x0][0x17c], R0 ; /* 0x00005f0003037a24 */ /* 0x000fc800078e0200 */ /*09f0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x188] ; /* 0x0000620003027625 */ /* 0x000fca00078e0202 */ /*0a00*/ STG.E [R2.64], R15 ; /* 0x0000000f02007986 */ /* 0x000fe2000c101904 */ /*0a10*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0a20*/ BRA 0xa20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
You are a helpful coding assistant specialized in converting CUDA source code to CUDA host assembly.
#include "includes.h" __global__ void compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
.file "tmpxft_00130023_00000000-6_compute_distances.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ .type _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_, @function _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movq %rcx, 24(%rsp) movl %r8d, 20(%rsp) movl %r9d, 16(%rsp) movq 216(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 8(%rsp), %rax movq %rax, 168(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z17compute_distancesPfiiS_iiiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_, .-_Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ .globl _Z17compute_distancesPfiiS_iiiS_ .type _Z17compute_distancesPfiiS_iiiS_, @function _Z17compute_distancesPfiiS_iiiS_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z17compute_distancesPfiiS_iiiS_, .-_Z17compute_distancesPfiiS_iiiS_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z17compute_distancesPfiiS_iiiS_" .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 _Z17compute_distancesPfiiS_iiiS_(%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 compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
#include <hip/hip_runtime.h> #include "includes.h" __global__ void compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
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 compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z17compute_distancesPfiiS_iiiS_ .globl _Z17compute_distancesPfiiS_iiiS_ .p2align 8 .type _Z17compute_distancesPfiiS_iiiS_,@function _Z17compute_distancesPfiiS_iiiS_: s_clause 0x2 s_load_b64 s[10:11], s[0:1], 0x8 s_load_b32 s12, s[0:1], 0x20 s_load_b64 s[4:5], s[0:1], 0x18 v_dual_mov_b32 v4, 0 :: v_dual_and_b32 v1, 0x3ff, v0 s_movk_i32 s2, 0x800 s_movk_i32 s3, 0x800 v_bfe_u32 v0, v0, 10, 10 s_lshl_b32 s13, s15, 4 s_lshl_b32 s16, s14, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(SALU_CYCLE_1) v_dual_mov_b32 v2, s13 :: v_dual_mov_b32 v3, s16 v_add_nc_u32_e32 v7, s16, v1 ds_store_b32 v4, v3 offset:2064 s_waitcnt lgkmcnt(0) s_lshl_b32 s6, s5, 4 v_dual_mov_b32 v10, s6 :: v_dual_add_nc_u32 v5, s2, v4 s_lshl_b32 s2, s11, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(SALU_CYCLE_1) v_dual_mov_b32 v9, s2 :: v_dual_add_nc_u32 v6, s3, v4 s_add_i32 s3, s12, -1 v_cmp_gt_i32_e64 s2, s4, v7 s_mul_i32 s3, s3, s11 s_add_i32 s7, s3, s13 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v11, s7 :: v_dual_add_nc_u32 v8, s13, v0 s_cmp_gt_i32 s13, s7 ds_store_2addr_b32 v5, v10, v2 offset0:2 offset1:3 ds_store_2addr_b32 v6, v11, v9 offset1:1 v_cmp_gt_i32_e64 s3, s10, v8 s_cbranch_scc1 .LBB0_14 s_ashr_i32 s14, s11, 31 s_clause 0x1 s_load_b64 s[6:7], s[0:1], 0x0 s_load_b64 s[8:9], s[0:1], 0x10 s_add_i32 s4, s11, s14 v_lshlrev_b32_e32 v10, 2, v1 s_xor_b32 s15, s4, s14 v_dual_mov_b32 v6, s16 :: v_dual_add_nc_u32 v9, s13, v1 v_cvt_f32_u32_e32 v2, s15 v_lshlrev_b32_e32 v11, 6, v0 v_add_nc_u32_e32 v7, 0x400, v10 v_lshlrev_b32_e32 v5, 2, v0 v_cmp_gt_i32_e64 s4, s10, v9 v_rcp_iflag_f32_e32 v8, v2 v_mad_u64_u32 v[2:3], null, v0, s11, v[1:2] s_and_b32 s10, s3, s2 s_sub_i32 s11, 0, s15 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_3) v_mad_u64_u32 v[3:4], null, v0, s5, v[1:2] s_waitcnt_depctr 0xfff v_mul_f32_e32 v4, 0x4f7ffffe, v8 v_dual_mov_b32 v8, 0 :: v_dual_add_nc_u32 v9, v11, v10 v_add_nc_u32_e32 v10, v7, v11 v_cvt_u32_f32_e32 v11, v4 v_mov_b32_e32 v4, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s16 s_movk_i32 s16, 0x800 s_delay_alu instid0(SALU_CYCLE_1) v_add_nc_u32_e32 v12, s16, v8 s_barrier buffer_gl0_inv ds_load_2addr_b32 v[12:13], v12 offset1:1 ds_load_b32 v14, v8 offset:2056 s_waitcnt lgkmcnt(1) v_readfirstlane_b32 s16, v13 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v6, v14, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s13, s16, s13 v_cmp_gt_i32_e32 vcc_lo, s13, v12 s_cbranch_vccnz .LBB0_14 .LBB0_3: s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_readfirstlane_b32 s16, v11 s_ashr_i32 s17, s13, 31 s_add_i32 s19, s13, s17 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) s_xor_b32 s19, s19, s17 s_mul_i32 s18, s11, s16 s_xor_b32 s17, s17, s14 s_mul_hi_u32 s18, s16, s18 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s16, s16, s18 s_mul_hi_u32 s16, s19, s16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_mul_i32 s18, s16, s15 s_sub_i32 s18, s19, s18 s_add_i32 s19, s16, 1 s_sub_i32 s20, s18, s15 s_cmp_ge_u32 s18, s15 s_cselect_b32 s16, s19, s16 s_cselect_b32 s18, s20, s18 s_add_i32 s19, s16, 1 s_cmp_ge_u32 s18, s15 s_cselect_b32 s16, s19, s16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_xor_b32 s16, s16, s17 s_sub_i32 s16, s16, s17 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v12, s16, v0 s_mov_b32 s16, exec_lo v_cmpx_le_i32_e64 s12, v12 s_xor_b32 s16, exec_lo, s16 s_cbranch_execz .LBB0_5 ds_store_b32 v9, v8 ds_store_b32 v10, v8 .LBB0_5: s_and_not1_saveexec_b32 s16, s16 s_cbranch_execz .LBB0_11 v_dual_mov_b32 v12, 0 :: v_dual_mov_b32 v13, 0 s_and_saveexec_b32 s17, s4 s_cbranch_execz .LBB0_8 v_add_nc_u32_e32 v13, s13, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v14, 31, v13 v_lshlrev_b64 v[13:14], 2, v[13:14] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v13, vcc_lo, s6, v13 v_add_co_ci_u32_e32 v14, vcc_lo, s7, v14, vcc_lo global_load_b32 v13, v[13:14], off .LBB0_8: s_or_b32 exec_lo, exec_lo, s17 s_waitcnt vmcnt(0) ds_store_b32 v9, v13 s_and_saveexec_b32 s17, s2 s_cbranch_execz .LBB0_10 v_add_nc_u32_e32 v12, v3, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v13, 31, v12 v_lshlrev_b64 v[12:13], 2, v[12:13] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v12, vcc_lo, s8, v12 v_add_co_ci_u32_e32 v13, vcc_lo, s9, v13, vcc_lo global_load_b32 v12, v[12:13], off .LBB0_10: s_or_b32 exec_lo, exec_lo, s17 s_waitcnt vmcnt(0) ds_store_b32 v10, v12 .LBB0_11: s_or_b32 exec_lo, exec_lo, s16 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_and_saveexec_b32 s16, s10 s_cbranch_execz .LBB0_2 s_mov_b32 s17, 0 .LBB0_13: s_delay_alu instid0(SALU_CYCLE_1) v_add_nc_u32_e32 v12, s17, v5 v_add_nc_u32_e32 v13, s17, v7 s_add_i32 s17, s17, 64 ds_load_b32 v12, v12 ds_load_b32 v13, v13 s_cmpk_lg_i32 s17, 0x400 s_waitcnt lgkmcnt(0) v_sub_f32_e32 v12, v12, v13 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v12, v12 s_cbranch_scc1 .LBB0_13 s_branch .LBB0_2 .LBB0_14: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, s3, s2 s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_16 v_mov_b32_e32 v2, 0 s_movk_i32 s2, 0x800 s_load_b64 s[0:1], s[0:1], 0x28 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v2, s2, v2 ds_load_2addr_b32 v[2:3], v2 offset0:3 offset1:4 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v0, v2, v0 v_mul_lo_u32 v0, v0, s5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add3_u32 v0, v3, v1, v0 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, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v4, off .LBB0_16: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z17compute_distancesPfiiS_iiiS_ .amdhsa_group_segment_fixed_size 2068 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 48 .amdhsa_user_sgpr_count 14 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 21 .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 _Z17compute_distancesPfiiS_iiiS_, .Lfunc_end0-_Z17compute_distancesPfiiS_iiiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 2068 .kernarg_segment_align: 8 .kernarg_segment_size: 48 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z17compute_distancesPfiiS_iiiS_ .private_segment_fixed_size: 0 .sgpr_count: 23 .sgpr_spill_count: 0 .symbol: _Z17compute_distancesPfiiS_iiiS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting HIP source code to AMD host assembly.
#include <hip/hip_runtime.h> #include "includes.h" __global__ void compute_distances(float * ref, int ref_width, int ref_pitch, float * query, int query_width, int query_pitch, int height, float * dist) { // Declaration of the shared memory arrays As and Bs used to store the sub-matrix of A and B __shared__ float shared_A[BLOCK_DIM][BLOCK_DIM]; __shared__ float shared_B[BLOCK_DIM][BLOCK_DIM]; // Sub-matrix of A (begin, step, end) and Sub-matrix of B (begin, step) __shared__ int begin_A; __shared__ int begin_B; __shared__ int step_A; __shared__ int step_B; __shared__ int end_A; // Thread index int tx = threadIdx.x; int ty = threadIdx.y; // Initializarion of the SSD for the current thread float ssd = 0.f; // Loop parameters begin_A = BLOCK_DIM * blockIdx.y; begin_B = BLOCK_DIM * blockIdx.x; step_A = BLOCK_DIM * ref_pitch; step_B = BLOCK_DIM * query_pitch; end_A = begin_A + (height-1) * ref_pitch; // Conditions int cond0 = (begin_A + tx < ref_width); // used to write in shared memory int cond1 = (begin_B + tx < query_width); // used to write in shared memory & to computations and to write in output array int cond2 = (begin_A + ty < ref_width); // used to computations and to write in output matrix // Loop over all the sub-matrices of A and B required to compute the block sub-matrix for (int a = begin_A, b = begin_B; a <= end_A; a += step_A, b += step_B) { // Load the matrices from device memory to shared memory; each thread loads one element of each matrix if (a/ref_pitch + ty < height) { shared_A[ty][tx] = (cond0)? ref[a + ref_pitch * ty + tx] : 0; shared_B[ty][tx] = (cond1)? query[b + query_pitch * ty + tx] : 0; } else { shared_A[ty][tx] = 0; shared_B[ty][tx] = 0; } // Synchronize to make sure the matrices are loaded __syncthreads(); // Compute the difference between the two matrixes; each thread computes one element of the block sub-matrix if (cond2 && cond1) { for (int k = 0; k < BLOCK_DIM; ++k){ float tmp = shared_A[k][ty] - shared_B[k][tx]; ssd += tmp*tmp; } } // Synchronize to make sure that the preceeding computation is done before loading two new sub-matrices of A and B in the next iteration __syncthreads(); } // Write the block sub-matrix to device memory; each thread writes one element if (cond2 && cond1) { dist[ (begin_A + ty) * query_pitch + begin_B + tx ] = ssd; } }
.text .file "compute_distances.hip" .globl _Z32__device_stub__compute_distancesPfiiS_iiiS_ # -- Begin function _Z32__device_stub__compute_distancesPfiiS_iiiS_ .p2align 4, 0x90 .type _Z32__device_stub__compute_distancesPfiiS_iiiS_,@function _Z32__device_stub__compute_distancesPfiiS_iiiS_: # @_Z32__device_stub__compute_distancesPfiiS_iiiS_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) leaq 168(%rsp), %rax movq %rax, 136(%rsp) leaq 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 $_Z17compute_distancesPfiiS_iiiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z32__device_stub__compute_distancesPfiiS_iiiS_, .Lfunc_end0-_Z32__device_stub__compute_distancesPfiiS_iiiS_ .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 $_Z17compute_distancesPfiiS_iiiS_, %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 _Z17compute_distancesPfiiS_iiiS_,@object # @_Z17compute_distancesPfiiS_iiiS_ .section .rodata,"a",@progbits .globl _Z17compute_distancesPfiiS_iiiS_ .p2align 3, 0x0 _Z17compute_distancesPfiiS_iiiS_: .quad _Z32__device_stub__compute_distancesPfiiS_iiiS_ .size _Z17compute_distancesPfiiS_iiiS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z17compute_distancesPfiiS_iiiS_" .size .L__unnamed_1, 33 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z32__device_stub__compute_distancesPfiiS_iiiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17compute_distancesPfiiS_iiiS_ .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 : _Z17compute_distancesPfiiS_iiiS_ .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R4, SR_CTAID.Y ; /* 0x0000000000047919 */ /* 0x000e220000002600 */ /*0020*/ ULDC UR4, c[0x0][0x180] ; /* 0x0000600000047ab9 */ /* 0x000fe20000000800 */ /*0030*/ MOV R6, c[0x0][0x16c] ; /* 0x00005b0000067a02 */ /* 0x000fe20000000f00 */ /*0040*/ UIADD3 UR4, UR4, -0x1, URZ ; /* 0xffffffff04047890 */ /* 0x000fe2000fffe03f */ /*0050*/ S2R R5, SR_CTAID.X ; /* 0x0000000000057919 */ /* 0x000e620000002500 */ /*0060*/ ULDC UR5, c[0x0][0x16c] ; /* 0x00005b0000057ab9 */ /* 0x000fe20000000800 */ /*0070*/ MOV R7, c[0x0][0x17c] ; /* 0x00005f0000077a02 */ /* 0x000fe20000000f00 */ /*0080*/ UIMAD UR4, UR4, UR5, URZ ; /* 0x00000005040472a4 */ /* 0x000fe2000f8e023f */ /*0090*/ S2R R3, SR_TID.Y ; /* 0x0000000000037919 */ /* 0x000ea20000002200 */ /*00a0*/ IMAD.SHL.U32 R6, R6, 0x10, RZ ; /* 0x0000001006067824 */ /* 0x000fe200078e00ff */ /*00b0*/ SHF.L.U32 R7, R7, 0x4, RZ ; /* 0x0000000407077819 */ /* 0x000fe200000006ff */ /*00c0*/ HFMA2.MMA R15, -RZ, RZ, 0, 0 ; /* 0x00000000ff0f7435 */ /* 0x000fe200000001ff */ /*00d0*/ S2R R0, SR_TID.X ; /* 0x0000000000007919 */ /* 0x000ee20000002100 */ /*00e0*/ ISETP.LE.AND P1, PT, RZ, UR4, PT ; /* 0x00000004ff007c0c */ /* 0x000fc4000bf23270 */ /*00f0*/ SHF.L.U32 R4, R4, 0x4, RZ ; /* 0x0000000404047819 */ /* 0x001fe400000006ff */ /*0100*/ SHF.L.U32 R5, R5, 0x4, RZ ; /* 0x0000000405057819 */ /* 0x002fe400000006ff */ /*0110*/ IADD3 R8, R4.reuse, UR4, RZ ; /* 0x0000000404087c10 */ /* 0x040fe2000fffe0ff */ /*0120*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fe20000000a00 */ /*0130*/ IADD3 R2, R4, R3, RZ ; /* 0x0000000304027210 */ /* 0x004fe20007ffe0ff */ /*0140*/ STS.128 [0x800], R4 ; /* 0x00080004ff007388 */ /* 0x0001e60000000c00 */ /*0150*/ ISETP.GE.AND P0, PT, R2, c[0x0][0x168], PT ; /* 0x00005a0002007a0c */ /* 0x000fe20003f06270 */ /*0160*/ STS [0x810], R8 ; /* 0x00081008ff007388 */ /* 0x0001e20000000800 */ /*0170*/ IADD3 R2, R5, R0, RZ ; /* 0x0000000005027210 */ /* 0x008fc80007ffe0ff */ /*0180*/ ISETP.LT.AND P0, PT, R2, c[0x0][0x178], !P0 ; /* 0x00005e0002007a0c */ /* 0x000fe20004701270 */ /*0190*/ @!P1 BRA 0x990 ; /* 0x000007f000009947 */ /* 0x000fd80003800000 */ /*01a0*/ IABS R6, c[0x0][0x16c] ; /* 0x00005b0000067a13 */ /* 0x001fe20000000000 */ /*01b0*/ IMAD.MOV.U32 R15, RZ, RZ, RZ ; /* 0x000000ffff0f7224 */ /* 0x000fe200078e00ff */ /*01c0*/ IADD3 R8, R4, R0, RZ ; /* 0x0000000004087210 */ /* 0x000fe40007ffe0ff */ /*01d0*/ I2F.RP R10, R6 ; /* 0x00000006000a7306 */ /* 0x000e220000209400 */ /*01e0*/ SHF.L.U32 R12, R0, 0x2, RZ ; /* 0x00000002000c7819 */ /* 0x000fe400000006ff */ /*01f0*/ ISETP.GE.AND P1, PT, R8, c[0x0][0x168], PT ; /* 0x00005a0008007a0c */ /* 0x000fe40003f26270 */ /*0200*/ MOV R8, RZ ; /* 0x000000ff00087202 */ /* 0x000fc40000000f00 */ /*0210*/ MOV R7, R5 ; /* 0x0000000500077202 */ /* 0x000fe40000000f00 */ /*0220*/ LEA R5, R3.reuse, R12, 0x6 ; /* 0x0000000c03057211 */ /* 0x040fe200078e30ff */ /*0230*/ MUFU.RCP R10, R10 ; /* 0x0000000a000a7308 */ /* 0x001e240000001000 */ /*0240*/ IADD3 R9, R10, 0xffffffe, RZ ; /* 0x0ffffffe0a097810 */ /* 0x001fe20007ffe0ff */ /*0250*/ IMAD R10, R3, c[0x0][0x17c], R0 ; /* 0x00005f00030a7a24 */ /* 0x000fca00078e0200 */ /*0260*/ F2I.FTZ.U32.TRUNC.NTZ R9, R9 ; /* 0x0000000900097305 */ /* 0x000e24000021f000 */ /*0270*/ IMAD.MOV R11, RZ, RZ, -R9 ; /* 0x000000ffff0b7224 */ /* 0x001fc800078e0a09 */ /*0280*/ IMAD R11, R11, R6, RZ ; /* 0x000000060b0b7224 */ /* 0x000fc800078e02ff */ /*0290*/ IMAD.HI.U32 R8, R9, R11, R8 ; /* 0x0000000b09087227 */ /* 0x000fc800078e0008 */ /*02a0*/ IMAD R9, R3, c[0x0][0x16c], R0 ; /* 0x00005b0003097a24 */ /* 0x000fe400078e0200 */ /*02b0*/ IABS R13, R4 ; /* 0x00000004000d7213 */ /* 0x000fe20000000000 */ /*02c0*/ BSSY B0, 0x4d0 ; /* 0x0000020000007945 */ /* 0x000fe20003800000 */ /*02d0*/ IABS R14, c[0x0][0x16c] ; /* 0x00005b00000e7a13 */ /* 0x002fc60000000000 */ /*02e0*/ IMAD.HI.U32 R11, R8, R13, RZ ; /* 0x0000000d080b7227 */ /* 0x000fca00078e00ff */ /*02f0*/ IADD3 R12, -R11, RZ, RZ ; /* 0x000000ff0b0c7210 */ /* 0x000fca0007ffe1ff */ /*0300*/ IMAD R13, R14, R12, R13 ; /* 0x0000000c0e0d7224 */ /* 0x000fe200078e020d */ /*0310*/ LOP3.LUT R12, R4, c[0x0][0x16c], RZ, 0x3c, !PT ; /* 0x00005b00040c7a12 */ /* 0x000fc800078e3cff */ /*0320*/ ISETP.GT.U32.AND P4, PT, R6, R13, PT ; /* 0x0000000d0600720c */ /* 0x000fe40003f84070 */ /*0330*/ ISETP.GE.AND P3, PT, R12, RZ, PT ; /* 0x000000ff0c00720c */ /* 0x000fd60003f66270 */ /*0340*/ @!P4 IADD3 R13, R13, -R14, RZ ; /* 0x8000000e0d0dc210 */ /* 0x000fe40007ffe0ff */ /*0350*/ @!P4 IADD3 R11, R11, 0x1, RZ ; /* 0x000000010b0bc810 */ /* 0x000fe40007ffe0ff */ /*0360*/ ISETP.GE.U32.AND P2, PT, R13, R6, PT ; /* 0x000000060d00720c */ /* 0x000fe40003f46070 */ /*0370*/ ISETP.NE.AND P4, PT, RZ, c[0x0][0x16c], PT ; /* 0x00005b00ff007a0c */ /* 0x000fd60003f85270 */ /*0380*/ @P2 IADD3 R11, R11, 0x1, RZ ; /* 0x000000010b0b2810 */ /* 0x000fc80007ffe0ff */ /*0390*/ @!P3 IADD3 R11, -R11, RZ, RZ ; /* 0x000000ff0b0bb210 */ /* 0x000fe40007ffe1ff */ /*03a0*/ @!P4 LOP3.LUT R11, RZ, c[0x0][0x16c], RZ, 0x33, !PT ; /* 0x00005b00ff0bca12 */ /* 0x000fc800078e33ff */ /*03b0*/ IADD3 R11, R11, R3, RZ ; /* 0x000000030b0b7210 */ /* 0x000fc80007ffe0ff */ /*03c0*/ ISETP.GE.AND P2, PT, R11, c[0x0][0x180], PT ; /* 0x000060000b007a0c */ /* 0x000fda0003f46270 */ /*03d0*/ @P2 STS [R5], RZ ; /* 0x000000ff05002388 */ /* 0x0011e20000000800 */ /*03e0*/ @P2 IMAD.MOV.U32 R18, RZ, RZ, RZ ; /* 0x000000ffff122224 */ /* 0x000fe200078e00ff */ /*03f0*/ @P2 BRA 0x4c0 ; /* 0x000000c000002947 */ /* 0x000fea0003800000 */ /*0400*/ @!P1 IADD3 R12, R9, R4, RZ ; /* 0x00000004090c9210 */ /* 0x000fe20007ffe0ff */ /*0410*/ HFMA2.MMA R14, -RZ, RZ, 0, 0 ; /* 0x00000000ff0e7435 */ /* 0x000fe200000001ff */ /*0420*/ @!P1 MOV R13, 0x4 ; /* 0x00000004000d9802 */ /* 0x000fe40000000f00 */ /*0430*/ ISETP.GE.AND P2, PT, R2, c[0x0][0x178], PT ; /* 0x00005e0002007a0c */ /* 0x000fc60003f46270 */ /*0440*/ @!P1 IMAD.WIDE R12, R12, R13, c[0x0][0x160] ; /* 0x000058000c0c9625 */ /* 0x000fca00078e020d */ /*0450*/ @!P1 LDG.E R14, [R12.64] ; /* 0x000000040c0e9981 */ /* 0x000ea2000c1e1900 */ /*0460*/ IMAD.MOV.U32 R18, RZ, RZ, RZ ; /* 0x000000ffff127224 */ /* 0x000fc800078e00ff */ /*0470*/ @!P2 IADD3 R16, R10, R7, RZ ; /* 0x000000070a10a210 */ /* 0x000fe40007ffe0ff */ /*0480*/ @!P2 MOV R17, 0x4 ; /* 0x000000040011a802 */ /* 0x000fca0000000f00 */ /*0490*/ @!P2 IMAD.WIDE R16, R16, R17, c[0x0][0x170] ; /* 0x00005c001010a625 */ /* 0x000fca00078e0211 */ /*04a0*/ @!P2 LDG.E R18, [R16.64] ; /* 0x000000041012a981 */ /* 0x000368000c1e1900 */ /*04b0*/ STS [R5], R14 ; /* 0x0000000e05007388 */ /* 0x0043e40000000800 */ /*04c0*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x000fea0003800000 */ /*04d0*/ STS [R5+0x400], R18 ; /* 0x0004001205007388 */ /* 0x0205e20000000800 */ /*04e0*/ BSSY B0, 0x920 ; /* 0x0000043000007945 */ /* 0x000fe60003800000 */ /*04f0*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0500*/ @!P0 BRA 0x910 ; /* 0x0000040000008947 */ /* 0x000fea0003800000 */ /*0510*/ LDS R13, [R3.X4] ; /* 0x00000000030d7984 */ /* 0x004fe80000004800 */ /*0520*/ LDS R14, [R0.X4+0x400] ; /* 0x00040000000e7984 */ /* 0x002e680000004800 */ /*0530*/ LDS R26, [R0.X4+0x440] ; /* 0x00044000001a7984 */ /* 0x000fe80000004800 */ /*0540*/ LDS R27, [R3.X4+0x40] ; /* 0x00004000031b7984 */ /* 0x000ea80000004800 */ /*0550*/ LDS R17, [R0.X4+0x480] ; /* 0x0004800000117984 */ /* 0x000fe80000004800 */ /*0560*/ LDS R18, [R3.X4+0x80] ; /* 0x0000800003127984 */ /* 0x000ee80000004800 */ /*0570*/ LDS R23, [R0.X4+0x4c0] ; /* 0x0004c00000177984 */ /* 0x000fe80000004800 */ /*0580*/ LDS R24, [R3.X4+0xc0] ; /* 0x0000c00003187984 */ /* 0x000f280000004800 */ /*0590*/ LDS R19, [R0.X4+0x500] ; /* 0x0005000000137984 */ /* 0x000fe80000004800 */ /*05a0*/ LDS R20, [R3.X4+0x100] ; /* 0x0001000003147984 */ /* 0x000f680000004800 */ /*05b0*/ LDS R21, [R0.X4+0x540] ; /* 0x0005400000157984 */ /* 0x000fe80000004800 */ /*05c0*/ LDS R22, [R3.X4+0x140] ; /* 0x0001400003167984 */ /* 0x000e220000004800 */ /*05d0*/ FADD R16, R13, -R14 ; /* 0x8000000e0d107221 */ /* 0x002fc60000000000 */ /*05e0*/ LDS R11, [R0.X4+0x580] ; /* 0x00058000000b7984 */ /* 0x000fe20000004800 */ /*05f0*/ FFMA R25, R16, R16, R15 ; /* 0x0000001010197223 */ /* 0x000fc6000000000f */ /*0600*/ LDS R12, [R3.X4+0x180] ; /* 0x00018000030c7984 */ /* 0x000e620000004800 */ /*0610*/ FADD R26, -R26, R27 ; /* 0x0000001b1a1a7221 */ /* 0x004fc60000000100 */ /*0620*/ LDS R13, [R0.X4+0x5c0] ; /* 0x0005c000000d7984 */ /* 0x000fe20000004800 */ /*0630*/ FFMA R25, R26, R26, R25 ; /* 0x0000001a1a197223 */ /* 0x000fc60000000019 */ /*0640*/ LDS R14, [R3.X4+0x1c0] ; /* 0x0001c000030e7984 */ /* 0x000ea20000004800 */ /*0650*/ FADD R18, -R17, R18 ; /* 0x0000001211127221 */ /* 0x008fc60000000100 */ /*0660*/ LDS R15, [R0.X4+0x600] ; /* 0x00060000000f7984 */ /* 0x000fe20000004800 */ /*0670*/ FFMA R25, R18, R18, R25 ; /* 0x0000001212197223 */ /* 0x000fc60000000019 */ /*0680*/ LDS R16, [R3.X4+0x200] ; /* 0x0002000003107984 */ /* 0x000ee20000004800 */ /*0690*/ FADD R24, -R23, R24 ; /* 0x0000001817187221 */ /* 0x010fc60000000100 */ /*06a0*/ LDS R18, [R0.X4+0x640] ; /* 0x0006400000127984 */ /* 0x000fe20000004800 */ /*06b0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe40000000019 */ /*06c0*/ FADD R24, -R19, R20 ; /* 0x0000001413187221 */ /* 0x020fe20000000100 */ /*06d0*/ LDS R17, [R3.X4+0x240] ; /* 0x0002400003117984 */ /* 0x000f260000004800 */ /*06e0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe20000000019 */ /*06f0*/ LDS R20, [R0.X4+0x680] ; /* 0x0006800000147984 */ /* 0x000fe20000004800 */ /*0700*/ FADD R24, -R21, R22 ; /* 0x0000001615187221 */ /* 0x001fc60000000100 */ /*0710*/ LDS R19, [R3.X4+0x280] ; /* 0x0002800003137984 */ /* 0x000e220000004800 */ /*0720*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*0730*/ LDS R22, [R0.X4+0x6c0] ; /* 0x0006c00000167984 */ /* 0x000fe20000004800 */ /*0740*/ FADD R24, -R11, R12 ; /* 0x0000000c0b187221 */ /* 0x002fc60000000100 */ /*0750*/ LDS R21, [R3.X4+0x2c0] ; /* 0x0002c00003157984 */ /* 0x000e620000004800 */ /*0760*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*0770*/ LDS R11, [R0.X4+0x700] ; /* 0x00070000000b7984 */ /* 0x000fe20000004800 */ /*0780*/ FADD R24, -R13, R14 ; /* 0x0000000e0d187221 */ /* 0x004fc60000000100 */ /*0790*/ LDS R12, [R3.X4+0x300] ; /* 0x00030000030c7984 */ /* 0x000ea20000004800 */ /*07a0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fc60000000019 */ /*07b0*/ LDS R14, [R0.X4+0x740] ; /* 0x00074000000e7984 */ /* 0x000fe20000004800 */ /*07c0*/ FADD R26, -R15, R16 ; /* 0x000000100f1a7221 */ /* 0x008fc60000000100 */ /*07d0*/ LDS R13, [R3.X4+0x340] ; /* 0x00034000030d7984 */ /* 0x000ee20000004800 */ /*07e0*/ FFMA R25, R26, R26, R25 ; /* 0x0000001a1a197223 */ /* 0x000fc60000000019 */ /*07f0*/ LDS R24, [R0.X4+0x780] ; /* 0x0007800000187984 */ /* 0x000fe80000004800 */ /*0800*/ LDS R23, [R3.X4+0x380] ; /* 0x0003800003177984 */ /* 0x000f620000004800 */ /*0810*/ FADD R18, -R18, R17 ; /* 0x0000001112127221 */ /* 0x010fc60000000100 */ /*0820*/ LDS R16, [R0.X4+0x7c0] ; /* 0x0007c00000107984 */ /* 0x000fe20000004800 */ /*0830*/ FFMA R25, R18, R18, R25 ; /* 0x0000001212197223 */ /* 0x000fc60000000019 */ /*0840*/ LDS R15, [R3.X4+0x3c0] ; /* 0x0003c000030f7984 */ /* 0x000f220000004800 */ /*0850*/ FADD R20, -R20, R19 ; /* 0x0000001314147221 */ /* 0x001fc80000000100 */ /*0860*/ FFMA R25, R20, R20, R25 ; /* 0x0000001414197223 */ /* 0x000fe40000000019 */ /*0870*/ FADD R22, -R22, R21 ; /* 0x0000001516167221 */ /* 0x002fc80000000100 */ /*0880*/ FFMA R25, R22, R22, R25 ; /* 0x0000001616197223 */ /* 0x000fe40000000019 */ /*0890*/ FADD R12, -R11, R12 ; /* 0x0000000c0b0c7221 */ /* 0x004fc80000000100 */ /*08a0*/ FFMA R25, R12, R12, R25 ; /* 0x0000000c0c197223 */ /* 0x000fe40000000019 */ /*08b0*/ FADD R14, -R14, R13 ; /* 0x0000000d0e0e7221 */ /* 0x008fc80000000100 */ /*08c0*/ FFMA R25, R14, R14, R25 ; /* 0x0000000e0e197223 */ /* 0x000fe40000000019 */ /*08d0*/ FADD R24, -R24, R23 ; /* 0x0000001718187221 */ /* 0x020fc80000000100 */ /*08e0*/ FFMA R25, R24, R24, R25 ; /* 0x0000001818197223 */ /* 0x000fe40000000019 */ /*08f0*/ FADD R16, -R16, R15 ; /* 0x0000000f10107221 */ /* 0x010fc80000000100 */ /*0900*/ FFMA R15, R16, R16, R25 ; /* 0x00000010100f7223 */ /* 0x000fe40000000019 */ /*0910*/ BSYNC B0 ; /* 0x0000000000007941 */ /* 0x004fea0003800000 */ /*0920*/ BAR.SYNC.DEFER_BLOCKING 0x0 ; /* 0x0000000000007b1d */ /* 0x000fec0000010000 */ /*0930*/ LDS.64 R12, [0x808] ; /* 0x00080800ff0c7984 */ /* 0x000ea80000000a00 */ /*0940*/ LDS R11, [0x810] ; /* 0x00081000ff0b7984 */ /* 0x000ee20000000800 */ /*0950*/ IADD3 R4, R12, R4, RZ ; /* 0x000000040c047210 */ /* 0x004fc40007ffe0ff */ /*0960*/ IADD3 R7, R7, R13, RZ ; /* 0x0000000d07077210 */ /* 0x000fe40007ffe0ff */ /*0970*/ ISETP.GT.AND P2, PT, R4, R11, PT ; /* 0x0000000b0400720c */ /* 0x008fda0003f44270 */ /*0980*/ @!P2 BRA 0x2b0 ; /* 0xfffff9200000a947 */ /* 0x000fea000383ffff */ /*0990*/ @!P0 EXIT ; /* 0x000000000000894d */ /* 0x000fea0003800000 */ /*09a0*/ LDS.64 R4, [0x800] ; /* 0x00080000ff047984 */ /* 0x003e220000000a00 */ /*09b0*/ HFMA2.MMA R2, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff027435 */ /* 0x000fe200000001ff */ /*09c0*/ IADD3 R3, R3, R4, RZ ; /* 0x0000000403037210 */ /* 0x001fe40007ffe0ff */ /*09d0*/ IADD3 R0, R0, R5, RZ ; /* 0x0000000500007210 */ /* 0x000fca0007ffe0ff */ /*09e0*/ IMAD R3, R3, c[0x0][0x17c], R0 ; /* 0x00005f0003037a24 */ /* 0x000fc800078e0200 */ /*09f0*/ IMAD.WIDE R2, R3, R2, c[0x0][0x188] ; /* 0x0000620003027625 */ /* 0x000fca00078e0202 */ /*0a00*/ STG.E [R2.64], R15 ; /* 0x0000000f02007986 */ /* 0x000fe2000c101904 */ /*0a10*/ EXIT ; /* 0x000000000000794d */ /* 0x000fea0003800000 */ /*0a20*/ BRA 0xa20; /* 0xfffffff000007947 */ /* 0x000fc0000383ffff */ /*0a30*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a40*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a50*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a60*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a70*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a80*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0a90*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0aa0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ab0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ac0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ad0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0ae0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ /*0af0*/ NOP; /* 0x0000000000007918 */ /* 0x000fc00000000000 */ ..........
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z17compute_distancesPfiiS_iiiS_ .globl _Z17compute_distancesPfiiS_iiiS_ .p2align 8 .type _Z17compute_distancesPfiiS_iiiS_,@function _Z17compute_distancesPfiiS_iiiS_: s_clause 0x2 s_load_b64 s[10:11], s[0:1], 0x8 s_load_b32 s12, s[0:1], 0x20 s_load_b64 s[4:5], s[0:1], 0x18 v_dual_mov_b32 v4, 0 :: v_dual_and_b32 v1, 0x3ff, v0 s_movk_i32 s2, 0x800 s_movk_i32 s3, 0x800 v_bfe_u32 v0, v0, 10, 10 s_lshl_b32 s13, s15, 4 s_lshl_b32 s16, s14, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(SALU_CYCLE_1) v_dual_mov_b32 v2, s13 :: v_dual_mov_b32 v3, s16 v_add_nc_u32_e32 v7, s16, v1 ds_store_b32 v4, v3 offset:2064 s_waitcnt lgkmcnt(0) s_lshl_b32 s6, s5, 4 v_dual_mov_b32 v10, s6 :: v_dual_add_nc_u32 v5, s2, v4 s_lshl_b32 s2, s11, 4 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(SALU_CYCLE_1) v_dual_mov_b32 v9, s2 :: v_dual_add_nc_u32 v6, s3, v4 s_add_i32 s3, s12, -1 v_cmp_gt_i32_e64 s2, s4, v7 s_mul_i32 s3, s3, s11 s_add_i32 s7, s3, s13 s_delay_alu instid0(SALU_CYCLE_1) v_dual_mov_b32 v11, s7 :: v_dual_add_nc_u32 v8, s13, v0 s_cmp_gt_i32 s13, s7 ds_store_2addr_b32 v5, v10, v2 offset0:2 offset1:3 ds_store_2addr_b32 v6, v11, v9 offset1:1 v_cmp_gt_i32_e64 s3, s10, v8 s_cbranch_scc1 .LBB0_14 s_ashr_i32 s14, s11, 31 s_clause 0x1 s_load_b64 s[6:7], s[0:1], 0x0 s_load_b64 s[8:9], s[0:1], 0x10 s_add_i32 s4, s11, s14 v_lshlrev_b32_e32 v10, 2, v1 s_xor_b32 s15, s4, s14 v_dual_mov_b32 v6, s16 :: v_dual_add_nc_u32 v9, s13, v1 v_cvt_f32_u32_e32 v2, s15 v_lshlrev_b32_e32 v11, 6, v0 v_add_nc_u32_e32 v7, 0x400, v10 v_lshlrev_b32_e32 v5, 2, v0 v_cmp_gt_i32_e64 s4, s10, v9 v_rcp_iflag_f32_e32 v8, v2 v_mad_u64_u32 v[2:3], null, v0, s11, v[1:2] s_and_b32 s10, s3, s2 s_sub_i32 s11, 0, s15 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_3) v_mad_u64_u32 v[3:4], null, v0, s5, v[1:2] s_waitcnt_depctr 0xfff v_mul_f32_e32 v4, 0x4f7ffffe, v8 v_dual_mov_b32 v8, 0 :: v_dual_add_nc_u32 v9, v11, v10 v_add_nc_u32_e32 v10, v7, v11 v_cvt_u32_f32_e32 v11, v4 v_mov_b32_e32 v4, 0 s_branch .LBB0_3 .LBB0_2: s_or_b32 exec_lo, exec_lo, s16 s_movk_i32 s16, 0x800 s_delay_alu instid0(SALU_CYCLE_1) v_add_nc_u32_e32 v12, s16, v8 s_barrier buffer_gl0_inv ds_load_2addr_b32 v[12:13], v12 offset1:1 ds_load_b32 v14, v8 offset:2056 s_waitcnt lgkmcnt(1) v_readfirstlane_b32 s16, v13 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v6, v14, v6 s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s13, s16, s13 v_cmp_gt_i32_e32 vcc_lo, s13, v12 s_cbranch_vccnz .LBB0_14 .LBB0_3: s_delay_alu instid0(VALU_DEP_2) | instskip(SKIP_1) | instid1(SALU_CYCLE_1) v_readfirstlane_b32 s16, v11 s_ashr_i32 s17, s13, 31 s_add_i32 s19, s13, s17 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(VALU_DEP_1) s_xor_b32 s19, s19, s17 s_mul_i32 s18, s11, s16 s_xor_b32 s17, s17, s14 s_mul_hi_u32 s18, s16, s18 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_add_i32 s16, s16, s18 s_mul_hi_u32 s16, s19, s16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_mul_i32 s18, s16, s15 s_sub_i32 s18, s19, s18 s_add_i32 s19, s16, 1 s_sub_i32 s20, s18, s15 s_cmp_ge_u32 s18, s15 s_cselect_b32 s16, s19, s16 s_cselect_b32 s18, s20, s18 s_add_i32 s19, s16, 1 s_cmp_ge_u32 s18, s15 s_cselect_b32 s16, s19, s16 s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_xor_b32 s16, s16, s17 s_sub_i32 s16, s16, s17 s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v12, s16, v0 s_mov_b32 s16, exec_lo v_cmpx_le_i32_e64 s12, v12 s_xor_b32 s16, exec_lo, s16 s_cbranch_execz .LBB0_5 ds_store_b32 v9, v8 ds_store_b32 v10, v8 .LBB0_5: s_and_not1_saveexec_b32 s16, s16 s_cbranch_execz .LBB0_11 v_dual_mov_b32 v12, 0 :: v_dual_mov_b32 v13, 0 s_and_saveexec_b32 s17, s4 s_cbranch_execz .LBB0_8 v_add_nc_u32_e32 v13, s13, v2 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v14, 31, v13 v_lshlrev_b64 v[13:14], 2, v[13:14] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v13, vcc_lo, s6, v13 v_add_co_ci_u32_e32 v14, vcc_lo, s7, v14, vcc_lo global_load_b32 v13, v[13:14], off .LBB0_8: s_or_b32 exec_lo, exec_lo, s17 s_waitcnt vmcnt(0) ds_store_b32 v9, v13 s_and_saveexec_b32 s17, s2 s_cbranch_execz .LBB0_10 v_add_nc_u32_e32 v12, v3, v6 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_ashrrev_i32_e32 v13, 31, v12 v_lshlrev_b64 v[12:13], 2, v[12:13] s_waitcnt lgkmcnt(0) s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2) v_add_co_u32 v12, vcc_lo, s8, v12 v_add_co_ci_u32_e32 v13, vcc_lo, s9, v13, vcc_lo global_load_b32 v12, v[12:13], off .LBB0_10: s_or_b32 exec_lo, exec_lo, s17 s_waitcnt vmcnt(0) ds_store_b32 v10, v12 .LBB0_11: s_or_b32 exec_lo, exec_lo, s16 s_waitcnt lgkmcnt(0) s_barrier buffer_gl0_inv s_and_saveexec_b32 s16, s10 s_cbranch_execz .LBB0_2 s_mov_b32 s17, 0 .LBB0_13: s_delay_alu instid0(SALU_CYCLE_1) v_add_nc_u32_e32 v12, s17, v5 v_add_nc_u32_e32 v13, s17, v7 s_add_i32 s17, s17, 64 ds_load_b32 v12, v12 ds_load_b32 v13, v13 s_cmpk_lg_i32 s17, 0x400 s_waitcnt lgkmcnt(0) v_sub_f32_e32 v12, v12, v13 s_delay_alu instid0(VALU_DEP_1) v_fmac_f32_e32 v4, v12, v12 s_cbranch_scc1 .LBB0_13 s_branch .LBB0_2 .LBB0_14: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(SALU_CYCLE_1) s_and_b32 s2, s3, s2 s_and_saveexec_b32 s3, s2 s_cbranch_execz .LBB0_16 v_mov_b32_e32 v2, 0 s_movk_i32 s2, 0x800 s_load_b64 s[0:1], s[0:1], 0x28 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_3) | instid1(VALU_DEP_1) v_add_nc_u32_e32 v2, s2, v2 ds_load_2addr_b32 v[2:3], v2 offset0:3 offset1:4 s_waitcnt lgkmcnt(0) v_add_nc_u32_e32 v0, v2, v0 v_mul_lo_u32 v0, v0, s5 s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1) v_add3_u32 v0, v3, v1, v0 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, s0, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo global_store_b32 v[0:1], v4, off .LBB0_16: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z17compute_distancesPfiiS_iiiS_ .amdhsa_group_segment_fixed_size 2068 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 48 .amdhsa_user_sgpr_count 14 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 1 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 1 .amdhsa_next_free_vgpr 15 .amdhsa_next_free_sgpr 21 .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 _Z17compute_distancesPfiiS_iiiS_, .Lfunc_end0-_Z17compute_distancesPfiiS_iiiS_ .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .offset: 8 .size: 4 .value_kind: by_value - .offset: 12 .size: 4 .value_kind: by_value - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 28 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: by_value - .address_space: global .offset: 40 .size: 8 .value_kind: global_buffer .group_segment_fixed_size: 2068 .kernarg_segment_align: 8 .kernarg_segment_size: 48 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z17compute_distancesPfiiS_iiiS_ .private_segment_fixed_size: 0 .sgpr_count: 23 .sgpr_spill_count: 0 .symbol: _Z17compute_distancesPfiiS_iiiS_.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 15 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata
You are a helpful coding assistant specialized in converting from CUDA host assembly to AMD host assembly.
.file "tmpxft_00130023_00000000-6_compute_distances.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2029: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2029: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ .type _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_, @function _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_: .LFB2051: .cfi_startproc endbr64 subq $200, %rsp .cfi_def_cfa_offset 208 movq %rdi, 40(%rsp) movl %esi, 36(%rsp) movl %edx, 32(%rsp) movq %rcx, 24(%rsp) movl %r8d, 20(%rsp) movl %r9d, 16(%rsp) movq 216(%rsp), %rax movq %rax, 8(%rsp) movq %fs:40, %rax movq %rax, 184(%rsp) xorl %eax, %eax leaq 40(%rsp), %rax movq %rax, 112(%rsp) leaq 36(%rsp), %rax movq %rax, 120(%rsp) leaq 32(%rsp), %rax movq %rax, 128(%rsp) leaq 24(%rsp), %rax movq %rax, 136(%rsp) leaq 20(%rsp), %rax movq %rax, 144(%rsp) leaq 16(%rsp), %rax movq %rax, 152(%rsp) leaq 208(%rsp), %rax movq %rax, 160(%rsp) leaq 8(%rsp), %rax movq %rax, 168(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl $1, 72(%rsp) movl $1, 76(%rsp) movl $1, 80(%rsp) movl $1, 84(%rsp) leaq 56(%rsp), %rcx leaq 48(%rsp), %rdx leaq 76(%rsp), %rsi leaq 64(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 184(%rsp), %rax subq %fs:40, %rax jne .L8 addq $200, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 56(%rsp) .cfi_def_cfa_offset 216 pushq 56(%rsp) .cfi_def_cfa_offset 224 leaq 128(%rsp), %r9 movq 92(%rsp), %rcx movl 100(%rsp), %r8d movq 80(%rsp), %rsi movl 88(%rsp), %edx leaq _Z17compute_distancesPfiiS_iiiS_(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 208 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2051: .size _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_, .-_Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ .globl _Z17compute_distancesPfiiS_iiiS_ .type _Z17compute_distancesPfiiS_iiiS_, @function _Z17compute_distancesPfiiS_iiiS_: .LFB2052: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 pushq 24(%rsp) .cfi_def_cfa_offset 24 movl 24(%rsp), %eax pushq %rax .cfi_def_cfa_offset 32 call _Z46__device_stub__Z17compute_distancesPfiiS_iiiS_PfiiS_iiiS_ addq $24, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2052: .size _Z17compute_distancesPfiiS_iiiS_, .-_Z17compute_distancesPfiiS_iiiS_ .section .rodata.str1.8,"aMS",@progbits,1 .align 8 .LC0: .string "_Z17compute_distancesPfiiS_iiiS_" .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 _Z17compute_distancesPfiiS_iiiS_(%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 "compute_distances.hip" .globl _Z32__device_stub__compute_distancesPfiiS_iiiS_ # -- Begin function _Z32__device_stub__compute_distancesPfiiS_iiiS_ .p2align 4, 0x90 .type _Z32__device_stub__compute_distancesPfiiS_iiiS_,@function _Z32__device_stub__compute_distancesPfiiS_iiiS_: # @_Z32__device_stub__compute_distancesPfiiS_iiiS_ .cfi_startproc # %bb.0: subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 72(%rsp) movl %esi, 12(%rsp) movl %edx, 8(%rsp) movq %rcx, 64(%rsp) movl %r8d, 4(%rsp) movl %r9d, (%rsp) leaq 72(%rsp), %rax movq %rax, 80(%rsp) leaq 12(%rsp), %rax movq %rax, 88(%rsp) leaq 8(%rsp), %rax movq %rax, 96(%rsp) leaq 64(%rsp), %rax movq %rax, 104(%rsp) leaq 4(%rsp), %rax movq %rax, 112(%rsp) movq %rsp, %rax movq %rax, 120(%rsp) leaq 160(%rsp), %rax movq %rax, 128(%rsp) leaq 168(%rsp), %rax movq %rax, 136(%rsp) leaq 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 $_Z17compute_distancesPfiiS_iiiS_, %edi pushq 16(%rsp) .cfi_adjust_cfa_offset 8 pushq 32(%rsp) .cfi_adjust_cfa_offset 8 callq hipLaunchKernel addq $168, %rsp .cfi_adjust_cfa_offset -168 retq .Lfunc_end0: .size _Z32__device_stub__compute_distancesPfiiS_iiiS_, .Lfunc_end0-_Z32__device_stub__compute_distancesPfiiS_iiiS_ .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 $_Z17compute_distancesPfiiS_iiiS_, %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 _Z17compute_distancesPfiiS_iiiS_,@object # @_Z17compute_distancesPfiiS_iiiS_ .section .rodata,"a",@progbits .globl _Z17compute_distancesPfiiS_iiiS_ .p2align 3, 0x0 _Z17compute_distancesPfiiS_iiiS_: .quad _Z32__device_stub__compute_distancesPfiiS_iiiS_ .size _Z17compute_distancesPfiiS_iiiS_, 8 .type .L__unnamed_1,@object # @0 .section .rodata.str1.1,"aMS",@progbits,1 .L__unnamed_1: .asciz "_Z17compute_distancesPfiiS_iiiS_" .size .L__unnamed_1, 33 .type __hip_fatbin_wrapper,@object # @__hip_fatbin_wrapper .section .hipFatBinSegment,"a",@progbits .p2align 3, 0x0 __hip_fatbin_wrapper: .long 1212764230 # 0x48495046 .long 1 # 0x1 .quad __hip_fatbin .quad 0 .size __hip_fatbin_wrapper, 24 .type __hip_gpubin_handle,@object # @__hip_gpubin_handle .local __hip_gpubin_handle .comm __hip_gpubin_handle,8,8 .section .init_array,"aw",@init_array .p2align 3, 0x0 .quad __hip_module_ctor .type __hip_cuid_,@object # @__hip_cuid_ .bss .globl __hip_cuid_ __hip_cuid_: .byte 0 # 0x0 .size __hip_cuid_, 1 .section ".linker-options","e",@llvm_linker_options .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym _Z32__device_stub__compute_distancesPfiiS_iiiS_ .addrsig_sym __hip_module_ctor .addrsig_sym __hip_module_dtor .addrsig_sym _Z17compute_distancesPfiiS_iiiS_ .addrsig_sym __hip_fatbin .addrsig_sym __hip_fatbin_wrapper .addrsig_sym __hip_cuid_
You are a helpful coding assistant specialized in converting CUDA source code to CUDA device assembly.
/* Compute the sum of two vectors using CUDA * Vishwas S */ #include <stdio.h> #include <stdlib.h> __global__ void add(int *a, int *b, int *c, int n) { int id = blockIdx.x*blockDim.x + threadIdx.x; if(id<n) c[id] = a[id] + b[id]; } int main() { int N; int *a, *b, *c, *da, *db, *dc; scanf("%d",&N); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); a = (int *)calloc(N,sizeof(int)); b = (int *)calloc(N,sizeof(int)); c = (int *)calloc(N,sizeof(int)); for(int i = 0; i < N; i++) { a[i] = rand()%48; b[i] = rand()%50; } int size = N*sizeof(int); cudaMalloc(&da,size); cudaMalloc(&db,size); cudaMalloc(&dc,size); cudaMemcpy(da,a,size,cudaMemcpyHostToDevice); cudaMemcpy(db,b,size,cudaMemcpyHostToDevice); cudaEventRecord(start); add<<<(N+511)/512,512>>>(da,db,dc,N); //block count, threads per block cudaEventRecord(stop); cudaMemcpy(c,dc,size,cudaMemcpyDeviceToHost); cudaEventSynchronize(stop); float ms; cudaEventElapsedTime(&ms,start,stop); printf("%f\n",ms); }
code for sm_80 Function : _Z3addPiS_S_i .headerflags @"EF_CUDA_TEXMODE_UNIFIED EF_CUDA_64BIT_ADDRESS EF_CUDA_SM80 EF_CUDA_VIRTUAL_SM(EF_CUDA_SM80)" /*0000*/ MOV R1, c[0x0][0x28] ; /* 0x00000a0000017a02 */ /* 0x000fe40000000f00 */ /*0010*/ S2R R6, SR_CTAID.X ; /* 0x0000000000067919 */ /* 0x000e280000002500 */ /*0020*/ S2R R3, SR_TID.X ; /* 0x0000000000037919 */ /* 0x000e240000002100 */ /*0030*/ IMAD R6, R6, c[0x0][0x0], R3 ; /* 0x0000000006067a24 */ /* 0x001fca00078e0203 */ /*0040*/ ISETP.GE.AND P0, PT, R6, c[0x0][0x178], PT ; /* 0x00005e0006007a0c */ /* 0x000fda0003f06270 */ /*0050*/ @P0 EXIT ; /* 0x000000000000094d */ /* 0x000fea0003800000 */ /*0060*/ HFMA2.MMA R7, -RZ, RZ, 0, 2.384185791015625e-07 ; /* 0x00000004ff077435 */ /* 0x000fe200000001ff */ /*0070*/ ULDC.64 UR4, c[0x0][0x118] ; /* 0x0000460000047ab9 */ /* 0x000fd20000000a00 */ /*0080*/ IMAD.WIDE R4, R6, R7, c[0x0][0x168] ; /* 0x00005a0006047625 */ /* 0x000fc800078e0207 */ /*0090*/ IMAD.WIDE R2, R6.reuse, R7.reuse, c[0x0][0x160] ; /* 0x0000580006027625 */ /* 0x0c0fe400078e0207 */ /*00a0*/ LDG.E R4, [R4.64] ; /* 0x0000000404047981 */ /* 0x000ea8000c1e1900 */ /*00b0*/ LDG.E R3, [R2.64] ; /* 0x0000000402037981 */ /* 0x000ea2000c1e1900 */ /*00c0*/ IMAD.WIDE R6, R6, R7, c[0x0][0x170] ; /* 0x00005c0006067625 */ /* 0x000fe200078e0207 */ /*00d0*/ IADD3 R9, R4, R3, RZ ; /* 0x0000000304097210 */ /* 0x004fca0007ffe0ff */ /*00e0*/ STG.E [R6.64], R9 ; /* 0x0000000906007986 */ /* 0x000fe2000c101904 */ /*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.
/* Compute the sum of two vectors using CUDA * Vishwas S */ #include <stdio.h> #include <stdlib.h> __global__ void add(int *a, int *b, int *c, int n) { int id = blockIdx.x*blockDim.x + threadIdx.x; if(id<n) c[id] = a[id] + b[id]; } int main() { int N; int *a, *b, *c, *da, *db, *dc; scanf("%d",&N); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); a = (int *)calloc(N,sizeof(int)); b = (int *)calloc(N,sizeof(int)); c = (int *)calloc(N,sizeof(int)); for(int i = 0; i < N; i++) { a[i] = rand()%48; b[i] = rand()%50; } int size = N*sizeof(int); cudaMalloc(&da,size); cudaMalloc(&db,size); cudaMalloc(&dc,size); cudaMemcpy(da,a,size,cudaMemcpyHostToDevice); cudaMemcpy(db,b,size,cudaMemcpyHostToDevice); cudaEventRecord(start); add<<<(N+511)/512,512>>>(da,db,dc,N); //block count, threads per block cudaEventRecord(stop); cudaMemcpy(c,dc,size,cudaMemcpyDeviceToHost); cudaEventSynchronize(stop); float ms; cudaEventElapsedTime(&ms,start,stop); printf("%f\n",ms); }
.file "tmpxft_0004a228_00000000-6_13b.cudafe1.cpp" .text #APP #NO_APP .type _ZL26__cudaUnregisterBinaryUtilv, @function _ZL26__cudaUnregisterBinaryUtilv: .LFB2060: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaUnregisterFatBinary@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2060: .size _ZL26__cudaUnregisterBinaryUtilv, .-_ZL26__cudaUnregisterBinaryUtilv .globl _Z27__device_stub__Z3addPiS_S_iPiS_S_i .type _Z27__device_stub__Z3addPiS_S_iPiS_S_i, @function _Z27__device_stub__Z3addPiS_S_iPiS_S_i: .LFB2082: .cfi_startproc endbr64 subq $152, %rsp .cfi_def_cfa_offset 160 movq %rdi, 24(%rsp) movq %rsi, 16(%rsp) movq %rdx, 8(%rsp) movl %ecx, 4(%rsp) movq %fs:40, %rax movq %rax, 136(%rsp) xorl %eax, %eax leaq 24(%rsp), %rax movq %rax, 96(%rsp) leaq 16(%rsp), %rax movq %rax, 104(%rsp) leaq 8(%rsp), %rax movq %rax, 112(%rsp) leaq 4(%rsp), %rax movq %rax, 120(%rsp) movl $1, 48(%rsp) movl $1, 52(%rsp) movl $1, 56(%rsp) movl $1, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) leaq 40(%rsp), %rcx leaq 32(%rsp), %rdx leaq 60(%rsp), %rsi leaq 48(%rsp), %rdi call __cudaPopCallConfiguration@PLT testl %eax, %eax je .L7 .L3: movq 136(%rsp), %rax subq %fs:40, %rax jne .L8 addq $152, %rsp .cfi_remember_state .cfi_def_cfa_offset 8 ret .L7: .cfi_restore_state pushq 40(%rsp) .cfi_def_cfa_offset 168 pushq 40(%rsp) .cfi_def_cfa_offset 176 leaq 112(%rsp), %r9 movq 76(%rsp), %rcx movl 84(%rsp), %r8d movq 64(%rsp), %rsi movl 72(%rsp), %edx leaq _Z3addPiS_S_i(%rip), %rdi call cudaLaunchKernel@PLT addq $16, %rsp .cfi_def_cfa_offset 160 jmp .L3 .L8: call __stack_chk_fail@PLT .cfi_endproc .LFE2082: .size _Z27__device_stub__Z3addPiS_S_iPiS_S_i, .-_Z27__device_stub__Z3addPiS_S_iPiS_S_i .globl _Z3addPiS_S_i .type _Z3addPiS_S_i, @function _Z3addPiS_S_i: .LFB2083: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 call _Z27__device_stub__Z3addPiS_S_iPiS_S_i addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE2083: .size _Z3addPiS_S_i, .-_Z3addPiS_S_i .section .rodata.str1.1,"aMS",@progbits,1 .LC0: .string "%d" .LC1: .string "%f\n" .text .globl main .type main, @function main: .LFB2057: .cfi_startproc endbr64 pushq %r14 .cfi_def_cfa_offset 16 .cfi_offset 14, -16 pushq %r13 .cfi_def_cfa_offset 24 .cfi_offset 13, -24 pushq %r12 .cfi_def_cfa_offset 32 .cfi_offset 12, -32 pushq %rbp .cfi_def_cfa_offset 40 .cfi_offset 6, -40 pushq %rbx .cfi_def_cfa_offset 48 .cfi_offset 3, -48 subq $80, %rsp .cfi_def_cfa_offset 128 movq %fs:40, %rax movq %rax, 72(%rsp) xorl %eax, %eax leaq 4(%rsp), %rsi leaq .LC0(%rip), %rdi call __isoc23_scanf@PLT leaq 32(%rsp), %rdi call cudaEventCreate@PLT leaq 40(%rsp), %rdi call cudaEventCreate@PLT movl 4(%rsp), %r14d movslq %r14d, %rbx movl $4, %esi movq %rbx, %rdi call calloc@PLT movq %rax, %r12 movl $4, %esi movq %rbx, %rdi call calloc@PLT movq %rax, %rbp movl $4, %esi movq %rbx, %rdi call calloc@PLT movq %rax, %r13 testl %r14d, %r14d jle .L12 movl $0, %ebx .L13: call rand@PLT movslq %eax, %rdx imulq $715827883, %rdx, %rdx sarq $35, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx leal (%rdx,%rdx,2), %edx sall $4, %edx subl %edx, %eax movl %eax, (%r12,%rbx,4) call rand@PLT movslq %eax, %rdx imulq $1374389535, %rdx, %rdx sarq $36, %rdx movl %eax, %ecx sarl $31, %ecx subl %ecx, %edx imull $50, %edx, %edx subl %edx, %eax movl %eax, 0(%rbp,%rbx,4) movl 4(%rsp), %r14d addq $1, %rbx cmpl %ebx, %r14d jg .L13 .L12: sall $2, %r14d movslq %r14d, %r14 leaq 8(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT leaq 16(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT leaq 24(%rsp), %rdi movq %r14, %rsi call cudaMalloc@PLT movl $1, %ecx movq %r14, %rdx movq %r12, %rsi movq 8(%rsp), %rdi call cudaMemcpy@PLT movl $1, %ecx movq %r14, %rdx movq %rbp, %rsi movq 16(%rsp), %rdi call cudaMemcpy@PLT movl $0, %esi movq 32(%rsp), %rdi call cudaEventRecord@PLT movl $512, 60(%rsp) movl $1, 64(%rsp) movl $1, 68(%rsp) movl 4(%rsp), %edx leal 1022(%rdx), %eax addl $511, %edx cmovns %edx, %eax sarl $9, %eax movl %eax, 48(%rsp) movl $1, 52(%rsp) movl $0, %r9d movl $0, %r8d movq 60(%rsp), %rdx movl $1, %ecx movq 48(%rsp), %rdi movl $1, %esi call __cudaPushCallConfiguration@PLT testl %eax, %eax je .L18 .L14: movl $0, %esi movq 40(%rsp), %rdi call cudaEventRecord@PLT movl $2, %ecx movq %r14, %rdx movq 24(%rsp), %rsi movq %r13, %rdi call cudaMemcpy@PLT movq 40(%rsp), %rdi call cudaEventSynchronize@PLT leaq 60(%rsp), %rdi movq 40(%rsp), %rdx movq 32(%rsp), %rsi call cudaEventElapsedTime@PLT pxor %xmm0, %xmm0 cvtss2sd 60(%rsp), %xmm0 leaq .LC1(%rip), %rsi movl $2, %edi movl $1, %eax call __printf_chk@PLT movq 72(%rsp), %rax subq %fs:40, %rax jne .L19 movl $0, %eax addq $80, %rsp .cfi_remember_state .cfi_def_cfa_offset 48 popq %rbx .cfi_def_cfa_offset 40 popq %rbp .cfi_def_cfa_offset 32 popq %r12 .cfi_def_cfa_offset 24 popq %r13 .cfi_def_cfa_offset 16 popq %r14 .cfi_def_cfa_offset 8 ret .L18: .cfi_restore_state movl 4(%rsp), %ecx movq 24(%rsp), %rdx movq 16(%rsp), %rsi movq 8(%rsp), %rdi call _Z27__device_stub__Z3addPiS_S_iPiS_S_i jmp .L14 .L19: call __stack_chk_fail@PLT .cfi_endproc .LFE2057: .size main, .-main .section .rodata.str1.1 .LC2: .string "_Z3addPiS_S_i" .text .type _ZL24__sti____cudaRegisterAllv, @function _ZL24__sti____cudaRegisterAllv: .LFB2085: .cfi_startproc endbr64 subq $8, %rsp .cfi_def_cfa_offset 16 leaq _ZL15__fatDeviceText(%rip), %rdi call __cudaRegisterFatBinary@PLT movq %rax, %rdi movq %rax, _ZL20__cudaFatCubinHandle(%rip) pushq $0 .cfi_def_cfa_offset 24 pushq $0 .cfi_def_cfa_offset 32 pushq $0 .cfi_def_cfa_offset 40 pushq $0 .cfi_def_cfa_offset 48 movl $0, %r9d movl $-1, %r8d leaq .LC2(%rip), %rdx movq %rdx, %rcx leaq _Z3addPiS_S_i(%rip), %rsi call __cudaRegisterFunction@PLT addq $32, %rsp .cfi_def_cfa_offset 16 movq _ZL20__cudaFatCubinHandle(%rip), %rdi call __cudaRegisterFatBinaryEnd@PLT leaq _ZL26__cudaUnregisterBinaryUtilv(%rip), %rdi call atexit@PLT addq $8, %rsp .cfi_def_cfa_offset 8 ret .cfi_endproc .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.
/* Compute the sum of two vectors using CUDA * Vishwas S */ #include <stdio.h> #include <stdlib.h> __global__ void add(int *a, int *b, int *c, int n) { int id = blockIdx.x*blockDim.x + threadIdx.x; if(id<n) c[id] = a[id] + b[id]; } int main() { int N; int *a, *b, *c, *da, *db, *dc; scanf("%d",&N); cudaEvent_t start, stop; cudaEventCreate(&start); cudaEventCreate(&stop); a = (int *)calloc(N,sizeof(int)); b = (int *)calloc(N,sizeof(int)); c = (int *)calloc(N,sizeof(int)); for(int i = 0; i < N; i++) { a[i] = rand()%48; b[i] = rand()%50; } int size = N*sizeof(int); cudaMalloc(&da,size); cudaMalloc(&db,size); cudaMalloc(&dc,size); cudaMemcpy(da,a,size,cudaMemcpyHostToDevice); cudaMemcpy(db,b,size,cudaMemcpyHostToDevice); cudaEventRecord(start); add<<<(N+511)/512,512>>>(da,db,dc,N); //block count, threads per block cudaEventRecord(stop); cudaMemcpy(c,dc,size,cudaMemcpyDeviceToHost); cudaEventSynchronize(stop); float ms; cudaEventElapsedTime(&ms,start,stop); printf("%f\n",ms); }
/* Compute the sum of two vectors using CUDA * Vishwas S */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void add(int *a, int *b, int *c, int n) { int id = blockIdx.x*blockDim.x + threadIdx.x; if(id<n) c[id] = a[id] + b[id]; } int main() { int N; int *a, *b, *c, *da, *db, *dc; scanf("%d",&N); hipEvent_t start, stop; hipEventCreate(&start); hipEventCreate(&stop); a = (int *)calloc(N,sizeof(int)); b = (int *)calloc(N,sizeof(int)); c = (int *)calloc(N,sizeof(int)); for(int i = 0; i < N; i++) { a[i] = rand()%48; b[i] = rand()%50; } int size = N*sizeof(int); hipMalloc(&da,size); hipMalloc(&db,size); hipMalloc(&dc,size); hipMemcpy(da,a,size,hipMemcpyHostToDevice); hipMemcpy(db,b,size,hipMemcpyHostToDevice); hipEventRecord(start); add<<<(N+511)/512,512>>>(da,db,dc,N); //block count, threads per block hipEventRecord(stop); hipMemcpy(c,dc,size,hipMemcpyDeviceToHost); hipEventSynchronize(stop); float ms; hipEventElapsedTime(&ms,start,stop); printf("%f\n",ms); }
You are a helpful coding assistant specialized in converting HIP source code to AMD device assembly.
/* Compute the sum of two vectors using CUDA * Vishwas S */ #include <hip/hip_runtime.h> #include <stdio.h> #include <stdlib.h> __global__ void add(int *a, int *b, int *c, int n) { int id = blockIdx.x*blockDim.x + threadIdx.x; if(id<n) c[id] = a[id] + b[id]; } int main() { int N; int *a, *b, *c, *da, *db, *dc; scanf("%d",&N); hipEvent_t start, stop; hipEventCreate(&start); hipEventCreate(&stop); a = (int *)calloc(N,sizeof(int)); b = (int *)calloc(N,sizeof(int)); c = (int *)calloc(N,sizeof(int)); for(int i = 0; i < N; i++) { a[i] = rand()%48; b[i] = rand()%50; } int size = N*sizeof(int); hipMalloc(&da,size); hipMalloc(&db,size); hipMalloc(&dc,size); hipMemcpy(da,a,size,hipMemcpyHostToDevice); hipMemcpy(db,b,size,hipMemcpyHostToDevice); hipEventRecord(start); add<<<(N+511)/512,512>>>(da,db,dc,N); //block count, threads per block hipEventRecord(stop); hipMemcpy(c,dc,size,hipMemcpyDeviceToHost); hipEventSynchronize(stop); float ms; hipEventElapsedTime(&ms,start,stop); printf("%f\n",ms); }
.text .amdgcn_target "amdgcn-amd-amdhsa--gfx1100" .protected _Z3addPiS_S_i .globl _Z3addPiS_S_i .p2align 8 .type _Z3addPiS_S_i,@function _Z3addPiS_S_i: s_clause 0x1 s_load_b32 s2, s[0:1], 0x2c s_load_b32 s3, s[0:1], 0x18 s_waitcnt lgkmcnt(0) s_and_b32 s2, s2, 0xffff s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_mad_u64_u32 v[1:2], null, s15, s2, v[0:1] s_mov_b32 s2, exec_lo v_cmpx_gt_i32_e64 s3, v1 s_cbranch_execz .LBB0_2 s_load_b128 s[4:7], s[0:1], 0x0 v_ashrrev_i32_e32 v2, 31, v1 s_load_b64 s[0:1], s[0:1], 0x10 s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(VALU_DEP_1) v_lshlrev_b64 v[0:1], 2, v[1:2] s_waitcnt lgkmcnt(0) v_add_co_u32 v2, vcc_lo, s4, v0 s_delay_alu instid0(VALU_DEP_2) v_add_co_ci_u32_e32 v3, vcc_lo, s5, v1, vcc_lo v_add_co_u32 v4, vcc_lo, s6, v0 v_add_co_ci_u32_e32 v5, vcc_lo, s7, v1, vcc_lo v_add_co_u32 v0, vcc_lo, s0, v0 global_load_b32 v2, v[2:3], off global_load_b32 v3, v[4:5], off v_add_co_ci_u32_e32 v1, vcc_lo, s1, v1, vcc_lo s_waitcnt vmcnt(0) v_add_nc_u32_e32 v2, v3, v2 global_store_b32 v[0:1], v2, off .LBB0_2: s_nop 0 s_sendmsg sendmsg(MSG_DEALLOC_VGPRS) s_endpgm .section .rodata,"a",@progbits .p2align 6, 0x0 .amdhsa_kernel _Z3addPiS_S_i .amdhsa_group_segment_fixed_size 0 .amdhsa_private_segment_fixed_size 0 .amdhsa_kernarg_size 288 .amdhsa_user_sgpr_count 15 .amdhsa_user_sgpr_dispatch_ptr 0 .amdhsa_user_sgpr_queue_ptr 0 .amdhsa_user_sgpr_kernarg_segment_ptr 1 .amdhsa_user_sgpr_dispatch_id 0 .amdhsa_user_sgpr_private_segment_size 0 .amdhsa_wavefront_size32 1 .amdhsa_uses_dynamic_stack 0 .amdhsa_enable_private_segment 0 .amdhsa_system_sgpr_workgroup_id_x 1 .amdhsa_system_sgpr_workgroup_id_y 0 .amdhsa_system_sgpr_workgroup_id_z 0 .amdhsa_system_sgpr_workgroup_info 0 .amdhsa_system_vgpr_workitem_id 0 .amdhsa_next_free_vgpr 6 .amdhsa_next_free_sgpr 16 .amdhsa_float_round_mode_32 0 .amdhsa_float_round_mode_16_64 0 .amdhsa_float_denorm_mode_32 3 .amdhsa_float_denorm_mode_16_64 3 .amdhsa_dx10_clamp 1 .amdhsa_ieee_mode 1 .amdhsa_fp16_overflow 0 .amdhsa_workgroup_processor_mode 1 .amdhsa_memory_ordered 1 .amdhsa_forward_progress 0 .amdhsa_shared_vgpr_count 0 .amdhsa_exception_fp_ieee_invalid_op 0 .amdhsa_exception_fp_denorm_src 0 .amdhsa_exception_fp_ieee_div_zero 0 .amdhsa_exception_fp_ieee_overflow 0 .amdhsa_exception_fp_ieee_underflow 0 .amdhsa_exception_fp_ieee_inexact 0 .amdhsa_exception_int_div_zero 0 .end_amdhsa_kernel .text .Lfunc_end0: .size _Z3addPiS_S_i, .Lfunc_end0-_Z3addPiS_S_i .section .AMDGPU.csdata,"",@progbits .text .p2alignl 7, 3214868480 .fill 96, 4, 3214868480 .type __hip_cuid_,@object .section .bss,"aw",@nobits .globl __hip_cuid_ __hip_cuid_: .byte 0 .size __hip_cuid_, 1 .ident "AMD clang version 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.2 25012 e5bf7e55c91490b07c49d8960fa7983d864936c4)" .section ".note.GNU-stack","",@progbits .addrsig .addrsig_sym __hip_cuid_ .amdgpu_metadata --- amdhsa.kernels: - .args: - .address_space: global .offset: 0 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 8 .size: 8 .value_kind: global_buffer - .address_space: global .offset: 16 .size: 8 .value_kind: global_buffer - .offset: 24 .size: 4 .value_kind: by_value - .offset: 32 .size: 4 .value_kind: hidden_block_count_x - .offset: 36 .size: 4 .value_kind: hidden_block_count_y - .offset: 40 .size: 4 .value_kind: hidden_block_count_z - .offset: 44 .size: 2 .value_kind: hidden_group_size_x - .offset: 46 .size: 2 .value_kind: hidden_group_size_y - .offset: 48 .size: 2 .value_kind: hidden_group_size_z - .offset: 50 .size: 2 .value_kind: hidden_remainder_x - .offset: 52 .size: 2 .value_kind: hidden_remainder_y - .offset: 54 .size: 2 .value_kind: hidden_remainder_z - .offset: 72 .size: 8 .value_kind: hidden_global_offset_x - .offset: 80 .size: 8 .value_kind: hidden_global_offset_y - .offset: 88 .size: 8 .value_kind: hidden_global_offset_z - .offset: 96 .size: 2 .value_kind: hidden_grid_dims .group_segment_fixed_size: 0 .kernarg_segment_align: 8 .kernarg_segment_size: 288 .language: OpenCL C .language_version: - 2 - 0 .max_flat_workgroup_size: 1024 .name: _Z3addPiS_S_i .private_segment_fixed_size: 0 .sgpr_count: 18 .sgpr_spill_count: 0 .symbol: _Z3addPiS_S_i.kd .uniform_work_group_size: 1 .uses_dynamic_stack: false .vgpr_count: 6 .vgpr_spill_count: 0 .wavefront_size: 32 .workgroup_processor_mode: 1 amdhsa.target: amdgcn-amd-amdhsa--gfx1100 amdhsa.version: - 1 - 2 ... .end_amdgpu_metadata