search

amrvac / AGILE-experimental

MPI-AMRVAC: A Parallel Adaptive Mesh Refinement Framework
https://amrvac.org/
GNU General Public License v3.0
2 stars 1 forks source link

manage to compile with nvhpc #2

Closed oporth closed 5 months ago

oporth commented 7 months ago

There will be issues with mod-random and multigrid. We don't really need mod-random and I have a fix for multigrid which someone (Jannis?) should review though. Let me know if I should share that!

jannisteunissen commented 6 months ago

I did update the octree-mg library a while ago for compatibility with the nvidia compiler, and just now also the 'single module' files that are used inside MPI-AMRVAC: https://github.com/jannisteunissen/octree-mg/tree/master/single_module

oporth commented 6 months ago

Hi Jannis, should we then include those updated 'single module' files here? Could you push them also to AGILE-experimental in that case?

Compilation of our test case with NVHPC seems to work fine. Thanks @beevageeva .

jannisteunissen commented 6 months ago

@oporth I have now pushed a new version that includes a fix similar to the one you did (and I also backported some changes from MPI-AMRVAC to the octree-mg repository)

oporth commented 6 months ago

Hi, ok great, thanks!
I believe @beevageeva is still investigating some weird issues with the auto-tests and a failing 1D compilation.

oporth commented 5 months ago

As discussed in monthly call 4th June, issue should be closed as it does not affect the test case of interest.