diff options
Diffstat (limited to 'polly/tools/GPURuntime/GPUJIT.c')
| -rw-r--r-- | polly/tools/GPURuntime/GPUJIT.c | 133 |
1 files changed, 63 insertions, 70 deletions
diff --git a/polly/tools/GPURuntime/GPUJIT.c b/polly/tools/GPURuntime/GPUJIT.c index a863ea4e485..a03ebe9a654 100644 --- a/polly/tools/GPURuntime/GPUJIT.c +++ b/polly/tools/GPURuntime/GPUJIT.c @@ -1,4 +1,5 @@ -/******************** GPUJIT.cpp - GPUJIT Execution Engine ********************/ +/******************** GPUJIT.cpp - GPUJIT Execution Engine \ + * ********************/ /* */ /* The LLVM Compiler Infrastructure */ /* */ @@ -53,8 +54,8 @@ static CuMemAllocFcnTy *CuMemAllocFcnPtr; typedef CUresult CUDAAPI CuFuncSetBlockShapeFcnTy(CUfunction, int, int, int); static CuFuncSetBlockShapeFcnTy *CuFuncSetBlockShapeFcnPtr; -typedef CUresult CUDAAPI CuParamSetvFcnTy(CUfunction, int, void *, - unsigned int); +typedef CUresult CUDAAPI +CuParamSetvFcnTy(CUfunction, int, void *, unsigned int); static CuParamSetvFcnTy *CuParamSetvFcnPtr; typedef CUresult CUDAAPI CuParamSetSizeFcnTy(CUfunction, unsigned int); @@ -90,13 +91,12 @@ static CuCtxCreateFcnTy *CuCtxCreateFcnPtr; typedef CUresult CUDAAPI CuDeviceGetFcnTy(CUdevice *, int); static CuDeviceGetFcnTy *CuDeviceGetFcnPtr; -typedef CUresult CUDAAPI CuModuleLoadDataExFcnTy(CUmodule *, const void *, - unsigned int, CUjit_option *, - void **); +typedef CUresult CUDAAPI CuModuleLoadDataExFcnTy( + CUmodule *, const void *, unsigned int, CUjit_option *, void **); static CuModuleLoadDataExFcnTy *CuModuleLoadDataExFcnPtr; -typedef CUresult CUDAAPI CuModuleGetFunctionFcnTy(CUfunction *, CUmodule, - const char *); +typedef CUresult CUDAAPI +CuModuleGetFunctionFcnTy(CUfunction *, CUmodule, const char *); static CuModuleGetFunctionFcnTy *CuModuleGetFunctionFcnPtr; typedef CUresult CUDAAPI CuDeviceComputeCapabilityFcnTy(int *, int *, CUdevice); @@ -109,15 +109,14 @@ static CuDeviceGetNameFcnTy *CuDeviceGetNameFcnPtr; typedef cudaError_t CUDARTAPI CudaEventCreateFcnTy(cudaEvent_t *); static CudaEventCreateFcnTy *CudaEventCreateFcnPtr; -typedef cudaError_t CUDARTAPI CudaEventRecordFcnTy(cudaEvent_t, - cudaStream_t); +typedef cudaError_t CUDARTAPI CudaEventRecordFcnTy(cudaEvent_t, cudaStream_t); static CudaEventRecordFcnTy *CudaEventRecordFcnPtr; typedef cudaError_t CUDARTAPI CudaEventSynchronizeFcnTy(cudaEvent_t); static CudaEventSynchronizeFcnTy *CudaEventSynchronizeFcnPtr; -typedef cudaError_t CUDARTAPI CudaEventElapsedTimeFcnTy(float *, cudaEvent_t, - cudaEvent_t); +typedef cudaError_t CUDARTAPI +CudaEventElapsedTimeFcnTy(float *, cudaEvent_t, cudaEvent_t); static CudaEventElapsedTimeFcnTy *CudaEventElapsedTimeFcnPtr; typedef cudaError_t CUDARTAPI CudaEventDestroyFcnTy(cudaEvent_t); @@ -166,81 +165,77 @@ static int initialDeviceAPIs() { * of this kind of cast may not be emitted by clang and new versions of gcc * as it is valid on POSIX 2008. */ - CuFuncSetBlockShapeFcnPtr = - (CuFuncSetBlockShapeFcnTy *) getAPIHandle(HandleCuda, - "cuFuncSetBlockShape"); + CuFuncSetBlockShapeFcnPtr = (CuFuncSetBlockShapeFcnTy *)getAPIHandle( + HandleCuda, "cuFuncSetBlockShape"); - CuParamSetvFcnPtr = (CuParamSetvFcnTy *) getAPIHandle(HandleCuda, - "cuParamSetv"); + CuParamSetvFcnPtr = + (CuParamSetvFcnTy *)getAPIHandle(HandleCuda, "cuParamSetv"); - CuParamSetSizeFcnPtr = (CuParamSetSizeFcnTy *) getAPIHandle(HandleCuda, - "cuParamSetSize"); + CuParamSetSizeFcnPtr = + (CuParamSetSizeFcnTy *)getAPIHandle(HandleCuda, "cuParamSetSize"); - CuLaunchGridFcnPtr = (CuLaunchGridFcnTy *) getAPIHandle(HandleCuda, - "cuLaunchGrid"); + CuLaunchGridFcnPtr = + (CuLaunchGridFcnTy *)getAPIHandle(HandleCuda, "cuLaunchGrid"); - CuMemAllocFcnPtr = (CuMemAllocFcnTy *) getAPIHandle(HandleCuda, - "cuMemAlloc_v2"); + CuMemAllocFcnPtr = + (CuMemAllocFcnTy *)getAPIHandle(HandleCuda, "cuMemAlloc_v2"); - CuMemFreeFcnPtr = (CuMemFreeFcnTy *) getAPIHandle(HandleCuda, "cuMemFree_v2"); + CuMemFreeFcnPtr = (CuMemFreeFcnTy *)getAPIHandle(HandleCuda, "cuMemFree_v2"); - CuMemcpyDtoHFcnPtr = (CuMemcpyDtoHFcnTy *) getAPIHandle(HandleCuda, - "cuMemcpyDtoH_v2"); + CuMemcpyDtoHFcnPtr = + (CuMemcpyDtoHFcnTy *)getAPIHandle(HandleCuda, "cuMemcpyDtoH_v2"); - CuMemcpyHtoDFcnPtr = (CuMemcpyHtoDFcnTy *) getAPIHandle(HandleCuda, - "cuMemcpyHtoD_v2"); + CuMemcpyHtoDFcnPtr = + (CuMemcpyHtoDFcnTy *)getAPIHandle(HandleCuda, "cuMemcpyHtoD_v2"); - CuModuleUnloadFcnPtr = (CuModuleUnloadFcnTy *) getAPIHandle(HandleCuda, - "cuModuleUnload"); + CuModuleUnloadFcnPtr = + (CuModuleUnloadFcnTy *)getAPIHandle(HandleCuda, "cuModuleUnload"); - CuCtxDestroyFcnPtr = (CuCtxDestroyFcnTy *) getAPIHandle(HandleCuda, - "cuCtxDestroy"); + CuCtxDestroyFcnPtr = + (CuCtxDestroyFcnTy *)getAPIHandle(HandleCuda, "cuCtxDestroy"); - CuInitFcnPtr = (CuInitFcnTy *) getAPIHandle(HandleCuda, "cuInit"); + CuInitFcnPtr = (CuInitFcnTy *)getAPIHandle(HandleCuda, "cuInit"); - CuDeviceGetCountFcnPtr = (CuDeviceGetCountFcnTy *) getAPIHandle(HandleCuda, - "cuDeviceGetCount"); + CuDeviceGetCountFcnPtr = + (CuDeviceGetCountFcnTy *)getAPIHandle(HandleCuda, "cuDeviceGetCount"); - CuDeviceGetFcnPtr = (CuDeviceGetFcnTy *) getAPIHandle(HandleCuda, - "cuDeviceGet"); + CuDeviceGetFcnPtr = + (CuDeviceGetFcnTy *)getAPIHandle(HandleCuda, "cuDeviceGet"); - CuCtxCreateFcnPtr = (CuCtxCreateFcnTy *) getAPIHandle(HandleCuda, - "cuCtxCreate_v2"); + CuCtxCreateFcnPtr = + (CuCtxCreateFcnTy *)getAPIHandle(HandleCuda, "cuCtxCreate_v2"); CuModuleLoadDataExFcnPtr = - (CuModuleLoadDataExFcnTy *) getAPIHandle(HandleCuda, "cuModuleLoadDataEx"); + (CuModuleLoadDataExFcnTy *)getAPIHandle(HandleCuda, "cuModuleLoadDataEx"); - CuModuleGetFunctionFcnPtr = - (CuModuleGetFunctionFcnTy *)getAPIHandle(HandleCuda, "cuModuleGetFunction"); + CuModuleGetFunctionFcnPtr = (CuModuleGetFunctionFcnTy *)getAPIHandle( + HandleCuda, "cuModuleGetFunction"); CuDeviceComputeCapabilityFcnPtr = - (CuDeviceComputeCapabilityFcnTy *)getAPIHandle(HandleCuda, - "cuDeviceComputeCapability"); + (CuDeviceComputeCapabilityFcnTy *)getAPIHandle( + HandleCuda, "cuDeviceComputeCapability"); CuDeviceGetNameFcnPtr = - (CuDeviceGetNameFcnTy *) getAPIHandle(HandleCuda, "cuDeviceGetName"); + (CuDeviceGetNameFcnTy *)getAPIHandle(HandleCuda, "cuDeviceGetName"); /* Get function pointer to CUDA Runtime APIs. */ CudaEventCreateFcnPtr = - (CudaEventCreateFcnTy *) getAPIHandle(HandleCudaRT, "cudaEventCreate"); + (CudaEventCreateFcnTy *)getAPIHandle(HandleCudaRT, "cudaEventCreate"); CudaEventRecordFcnPtr = - (CudaEventRecordFcnTy *) getAPIHandle(HandleCudaRT, "cudaEventRecord"); + (CudaEventRecordFcnTy *)getAPIHandle(HandleCudaRT, "cudaEventRecord"); - CudaEventSynchronizeFcnPtr = - (CudaEventSynchronizeFcnTy *) getAPIHandle(HandleCudaRT, - "cudaEventSynchronize"); + CudaEventSynchronizeFcnPtr = (CudaEventSynchronizeFcnTy *)getAPIHandle( + HandleCudaRT, "cudaEventSynchronize"); - CudaEventElapsedTimeFcnPtr = - (CudaEventElapsedTimeFcnTy *) getAPIHandle(HandleCudaRT, - "cudaEventElapsedTime"); + CudaEventElapsedTimeFcnPtr = (CudaEventElapsedTimeFcnTy *)getAPIHandle( + HandleCudaRT, "cudaEventElapsedTime"); CudaEventDestroyFcnPtr = - (CudaEventDestroyFcnTy *) getAPIHandle(HandleCudaRT, "cudaEventDestroy"); + (CudaEventDestroyFcnTy *)getAPIHandle(HandleCudaRT, "cudaEventDestroy"); - CudaThreadSynchronizeFcnPtr = - (CudaThreadSynchronizeFcnTy *) getAPIHandle(HandleCudaRT, - "cudaThreadSynchronize"); + CudaThreadSynchronizeFcnPtr = (CudaThreadSynchronizeFcnTy *)getAPIHandle( + HandleCudaRT, "cudaThreadSynchronize"); return 1; } @@ -269,7 +264,7 @@ void polly_initDevice(PollyGPUContext **Context, PollyGPUDevice **Device) { } /* We select the 1st device as default. */ - *Device = malloc(sizeof(PollyGPUDevice)); + *Device = malloc(sizeof(PollyGPUDevice)); if (*Device == 0) { fprintf(stdout, "Allocate memory for Polly GPU device failed.\n"); exit(-1); @@ -297,8 +292,8 @@ void polly_getPTXModule(void *PTXBuffer, PollyGPUModule **Module) { exit(-1); } - if (CuModuleLoadDataExFcnPtr(&((*Module)->Cuda), PTXBuffer, 0, 0, 0) - != CUDA_SUCCESS) { + if (CuModuleLoadDataExFcnPtr(&((*Module)->Cuda), PTXBuffer, 0, 0, 0) != + CUDA_SUCCESS) { fprintf(stdout, "Loading ptx assembly text failed.\n"); exit(-1); } @@ -313,8 +308,8 @@ void polly_getPTXKernelEntry(const char *KernelName, PollyGPUModule *Module, } /* Locate the kernel entry point. */ - if(CuModuleGetFunctionFcnPtr(&((*Kernel)->Cuda), Module->Cuda, KernelName) - != CUDA_SUCCESS) { + if (CuModuleGetFunctionFcnPtr(&((*Kernel)->Cuda), Module->Cuda, KernelName) != + CUDA_SUCCESS) { fprintf(stdout, "Loading kernel function failed.\n"); exit(-1); } @@ -354,9 +349,8 @@ void polly_stopTimerByCudaEvent(PollyGPUEvent *Start, PollyGPUEvent *Stop, free(Stop); } -void polly_allocateMemoryForHostAndDevice(void **HostData, - PollyGPUDevicePtr **DevData, - int MemSize) { +void polly_allocateMemoryForHostAndDevice( + void **HostData, PollyGPUDevicePtr **DevData, int MemSize) { if ((*HostData = (int *)malloc(MemSize)) == 0) { fprintf(stdout, "Could not allocate host memory.\n"); exit(-1); @@ -378,7 +372,7 @@ void polly_copyFromHostToDevice(PollyGPUDevicePtr *DevData, void *HostData, void polly_copyFromDeviceToHost(void *HostData, PollyGPUDevicePtr *DevData, int MemSize) { - if(CuMemcpyDtoHFcnPtr(HostData, DevData->Cuda, MemSize) != CUDA_SUCCESS) { + if (CuMemcpyDtoHFcnPtr(HostData, DevData->Cuda, MemSize) != CUDA_SUCCESS) { fprintf(stdout, "Copying results from device to host memory failed.\n"); exit(-1); } @@ -405,10 +399,9 @@ void polly_launchKernel(PollyGPUFunction *Kernel, int GridWidth, fprintf(stdout, "CUDA kernel launched.\n"); } -void polly_cleanupGPGPUResources(void *HostData, PollyGPUDevicePtr *DevData, - PollyGPUModule *Module, - PollyGPUContext *Context, - PollyGPUFunction *Kernel) { +void polly_cleanupGPGPUResources( + void *HostData, PollyGPUDevicePtr *DevData, PollyGPUModule *Module, + PollyGPUContext *Context, PollyGPUFunction *Kernel) { if (HostData) { free(HostData); HostData = 0; |
