AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

SWDEV-368881 - Implement hipModuleLaunchCooperative* APIs

Browse Source 
- Add implementation for hipModuleLaunchCooperativeKernel, hipModuleLaunchCooperativeKernelMultiDevice APIs

Change-Id: I2a80264e458549211c34b413397a08b57d021147


[ROCm/hipother commit: 9808909320]
This commit is contained in:
Rakesh Roy
2022-12-06 00:04:55 +05:30
committed by Rakesh Roy
parent bb291a85fd
commit d0cb83e7a8
1 changed files with 19 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/hipother/hipnv/include/hip/nvidia_detail/nvidia_hip_runtime_api.h
+19
View File
@@ -437,6 +437,7 @@ typedef struct cudaArray* hipArray_t;
typedef struct cudaArray* hipArray_const_t;
typedef struct cudaFuncAttributes hipFuncAttributes;
typedef struct cudaLaunchParams hipLaunchParams;
typedef CUDA_LAUNCH_PARAMS hipFunctionLaunchParams;
#define hipFunction_attribute CUfunction_attribute
#define hipPointer_attribute CUpointer_attribute
#define hip_Memcpy2D CUDA_MEMCPY2D
@@ -2807,11 +2808,29 @@ inline static hipError_t hipLaunchCooperativeKernel(const void* f, dim3 gridDim,
cudaLaunchCooperativeKernel(f, gridDim, blockDim, kernelParams, sharedMemBytes, stream));
}
inline static hipError_t hipModuleLaunchCooperativeKernel(hipFunction_t f, unsigned int gridDimX,
unsigned int gridDimY, unsigned int gridDimZ,
unsigned int blockDimX, unsigned int blockDimY,
unsigned int blockDimZ, unsigned int sharedMemBytes,
hipStream_t stream, void** kernelParams) {
return hipCUResultTohipError(cuLaunchCooperativeKernel(f, gridDimX, gridDimY, gridDimZ,
blockDimX, blockDimY, blockDimZ,
sharedMemBytes, stream,kernelParams));
}
inline static hipError_t hipLaunchCooperativeKernelMultiDevice(hipLaunchParams* launchParamsList,
int numDevices, unsigned int flags) {
return hipCUDAErrorTohipError(cudaLaunchCooperativeKernelMultiDevice(launchParamsList, numDevices, flags));
}
inline static hipError_t hipModuleLaunchCooperativeKernelMultiDevice(
hipFunctionLaunchParams* launchParamsList,
unsigned int numDevices,
unsigned int flags) {
return hipCUResultTohipError(cuLaunchCooperativeKernelMultiDevice(launchParamsList,
numDevices, flags));
}
inline static hipError_t hipImportExternalSemaphore(hipExternalSemaphore_t* extSem_out,
const hipExternalSemaphoreHandleDesc* semHandleDesc) {
return hipCUDAErrorTohipError(cudaImportExternalSemaphore(extSem_out,(const struct cudaExternalSemaphoreHandleDesc*)semHandleDesc));