I currently have a Visual Studio 2019 solution with about 12 Linux console app projects, each one a part of a larger distributed processing queue based on Docker. We found recently one of these apps can benefit from GPU processing; I tried to find out a way to include a CUDA Linux project to the solution, to no avail. How would one go about doing that?
Any given answer has to comply with the following:
- The project must be able to compile CUDA kernels and regular C++ code at the same time;
- The project must compile and run in Linux machines and/or Docker images;
- The project can be made part of a larger Visual Studio solution.