Support aocc and nvfortran compilers#44
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #44 +/- ##
==========================================
+ Coverage 55.13% 62.45% +7.31%
==========================================
Files 42 42
Lines 6170 7519 +1349
Branches 832 835 +3
==========================================
+ Hits 3402 4696 +1294
- Misses 2414 2574 +160
+ Partials 354 249 -105 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
facf17f to
b9c85e4
Compare
34adb30 to
e190db9
Compare
14d3369 to
54fabf2
Compare
|
So the CI is finally green, including the aocc and nvfortran jobs. (I still need to do a bit of manual testing since not all the codepaths modified here are covered by tests, but otherwise this is good to go). Notably, earlier versions of nvfortran (e.g. 26.1) fails to compile the unit tests, but the latest version (26.5 works) so I won't bother implementing the workaround. |
|
Restricting myself to the Fortran. A couple of extra bits of commented-out code have appeared. May want to try to remove. Otherwise, ok! |
|
Thanks for taking a look. I'll keep those bits around since we might need them soon, and if not I am going to do a larger cleanup of commented out code separately in one swoop. |
(PR opened on behalf of @kevinstratford, as discussed in #39)
This PR makes OpenFMS compatible with nvfortran and aocc compilers. This required some workarounds due to compiler bugs, but also fixing issues in the RestartModule, where we were writing unallocated arrays! (I am very surprised the other compilers were fined with this).
Closes #39.