-
- Downloads
There was an error fetching the commit references. Please try again later.
Added wrapper function for CUDA kernel launch
Advantages: - pure C++ for client code, no "<<< ... >>>" - hence, clang-format can be used for source code formatting - the HAVE_CUDA guards are not necessary anymore - type of the kernel function can be added to exception messages - easier debugging of kernel launch configurations
parent
9c99291f
No related branches found
No related tags found
Showing
- src/TNL/Algorithms/ParallelFor.h 168 additions, 158 deletionssrc/TNL/Algorithms/ParallelFor.h
- src/TNL/Algorithms/detail/CudaMultireductionKernel.h 49 additions, 58 deletionssrc/TNL/Algorithms/detail/CudaMultireductionKernel.h
- src/TNL/Algorithms/detail/CudaReductionKernel.h 77 additions, 100 deletionssrc/TNL/Algorithms/detail/CudaReductionKernel.h
- src/TNL/Algorithms/detail/CudaScanKernel.h 127 additions, 100 deletionssrc/TNL/Algorithms/detail/CudaScanKernel.h
- src/TNL/Cuda/DummyDefs.h 34 additions, 0 deletionssrc/TNL/Cuda/DummyDefs.h
- src/TNL/Cuda/KernelLaunch.h 114 additions, 0 deletionssrc/TNL/Cuda/KernelLaunch.h
Loading
Please register or sign in to comment