-
- Downloads
Improve support for GPU Thread Indexing
- Introduce GpuThreadRange and GpuKernelFunction to export the launch space expected for the kernel - Refactor GPU platforms to create and export a thread range
Showing
- src/pystencils/backend/kernelcreation/iteration_space.py 16 additions, 2 deletionssrc/pystencils/backend/kernelcreation/iteration_space.py
- src/pystencils/backend/kernelfunction.py 101 additions, 1 deletionsrc/pystencils/backend/kernelfunction.py
- src/pystencils/backend/platforms/__init__.py 4 additions, 1 deletionsrc/pystencils/backend/platforms/__init__.py
- src/pystencils/backend/platforms/cuda.py 88 additions, 0 deletionssrc/pystencils/backend/platforms/cuda.py
- src/pystencils/backend/platforms/generic_gpu.py 48 additions, 70 deletionssrc/pystencils/backend/platforms/generic_gpu.py
- src/pystencils/backend/platforms/platform.py 2 additions, 1 deletionsrc/pystencils/backend/platforms/platform.py
- src/pystencils/backend/platforms/sycl.py 31 additions, 27 deletionssrc/pystencils/backend/platforms/sycl.py
- src/pystencils/kernelcreation.py 27 additions, 52 deletionssrc/pystencils/kernelcreation.py
- tests/nbackend/kernelcreation/platform/test_basic_gpu.py 2 additions, 2 deletionstests/nbackend/kernelcreation/platform/test_basic_gpu.py
Loading
Please register or sign in to comment