FTN95 or C++ AND CUDA

NVIDIA have a compiler NVCC which compiles CUDA parallel code. This code can then be linked to the users code using Microsoft C++. Both compilers are needed with complicated command line calls. I have attempted some of the simpler examples of CUDA code using SILVERFROST C and have got nowhere.

I would like to run CUDA ideally from a Silverfrost FTN95 program but will settle for a Silverfrost C++ program as a start.

Has anyone got CUDA working using the Silverfrost Compiler rather than MS C++. I would like a nice simple example case to get started.

Without knowing anything about CUDA my advice would be...

It depends what language features are being used in CUDA - if it is recent code then perhaps it uses a lot stl etc.? If this is the case SCC may not support it. SCC is not really maintained any longer (save for bug fixes with the existing feature set) and is included with FTN95 as a freebie.

If the features you need are not supported then you will be better off using MSVC and calling from FTN95.