guillaumetousignant / NDG_cuda

MIT License
1 stars 0 forks source link

NDG_cuda

GitHub license GitHub release Documentation Status build-ubuntu Actions Status build-windows Actions Status

Installation:

mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=.. -DCMAKE_CUDA_ARCHITECTURES="60;61;70"
cmake --build .
cmake --install .

sm_61 is GTX 10X0, sm_60 is Tesla P100, sm_70 is Tesla V100

Dependencies:

To build without cgns and HDF5, add -DUSE_CGNS=OFF to cmake call.

To build tests, add -DBUILD_TESTING=ON to cmake call. Then:

ctest

To generate documentation, add -DBUILD_DOC=ON to cmake call. Additional dependencies:

To include in a CMake project:

find_package(SEM-CUDA 1.0.0 REQUIRED)
target_link_libraries(example SEM-CUDA::SEM-CUDA)