A small test project to explore CUDA's standard parallelism feature shipped with nvc++ using WSL2.
This project has been created and successfully tested using following setup:
- Windows 10 21H2 (which is the minimum required version to use CUDA in WSL as stated in the MS documentation)
- Ubuntu 22.02
- NVIDIA HPC SDK 22.5
- Visual Studio 17.2.4
The project possibly may also be built using previous HPC SDK versions.
After cloning this repository the PATH variable set in the CMakePresets.json file may have to be updated to include the correct path to the HPC SDK:
"environment": {
"PATH": "/opt/nvidia/hpc_sdk/Linux_x86_64/22.5/compilers/bin/:$penv{PATH}"
}
The project can not be built with MSVC because the parallelized std::transform overloads require the iterators to fulfill LegacyForwardIterator which is not the case for the iterators of std::views::iota.