David Beckingsale
David Beckingsale
I think you want `CMAKE_CUDA_FLAGS`. I don't believe `CMAKE_CUDA_ARCH` does anything, neither does `CUDA_ARCH`. `CUDA_NVCC_FLAGS` is the old "FindCUDA" variable.
Would they be "extra" compile flags, or the only compile flags?
Are we going to use sphinx to document blt itself?
Sounds good - I was going to volunteer to help out with a user guide since I have been a user-turned-developer 😄
@rrsettgast - this is a problem with GCC and CUDA 10, fixed in CUDA 11. The correct workaround is to add the flag `-Xcompiler -mno-float128` to your CUDA compile flags....
Hope so! I've thought about this before. I think the challenge would be ensuring that we have a robust way of defining valid values for `openmp-targets`.
Hey @evaleev - thanks for reporting this. I believe it's actually an error in what we implemented. The log and 'replay' output streams are correctly set to a specific file,...
That's where `umpire-config.cmake` is installed. After running `find_package`, you should be able to just depend on the `umpire` target within your own CMake project.
@adrienbernede looks like latest commit broke, was that expected?
@msimberg - can you please confirm that #781 fixes this issue for you?