search

ufs-community / UFS_UTILS

Utilities for the NCEP models.
Other
21 stars 104 forks source link

Add regression test for GNU compiler #151

Closed GeorgeGayno-NOAA closed 3 years ago

GeorgeGayno-NOAA commented 3 years ago

@DusanJovic-NOAA recommends that UFS_UTILS should have regression tests for GNU compilers. Or at a minimum, the repo should be compiled successfully with GNU before merges.

GeorgeGayno-NOAA commented 3 years ago

GNU-based NCEPLIBS are available on Hera, by using this module: https://github.com/ufs-community/ufs-weather-model/blob/develop/modulefiles/hera.gnu/fv3

GeorgeGayno-NOAA commented 3 years ago

The branch at 6301855 was successfully compiled on Hera using Gnu.

The chgres_cube and grid_gen regression tests ran to completion. Results did not match the baseline, but that was expected.

The global_cycle and snow2mdl tests failed - the executables broke with segmentation faults. The ice2_blend test failed because the copygb2 step, which runs before the executable, failed because it could not find the libiopm5.so library. I will likely need a Gnu version of copygb2.

chgres_cube and the grid generation programs are part of the public release. So it is important they work with Gnu.

The global_cycle, ice_blend and snow2mdl programs are not supported as part of the public release. So, it is not as critical they work with Gnu. I will think about whether to fix them under this issue or a future issue.

GeorgeGayno-NOAA commented 3 years ago

@DusanJovic-NOAA have I satisfied your request?

GeorgeGayno-NOAA commented 3 years ago

@JeffBeck-NOAA Under this issue I am renaming the build files - and some of the files are for your machines. Who is the best person in your group to review these changes?

DusanJovic-NOAA commented 3 years ago

@DusanJovic-NOAA have I satisfied your request?

Yes. Thanks.

JeffBeck-NOAA commented 3 years ago

@JeffBeck-NOAA Under this issue I am renaming the build files - and some of the files are for your machines. Who is the best person in your group to review these changes?

@GeorgeGayno-NOAA, I don't see a related PR attached to this issue, so I'm assuming you're still working on this locally? If you're referring to Odin/Stampede2, the contact people for those machines would be @LarissaReames-NOAA and @ywangwof. For Cheyenne, it would be @JulieSchramm or @mkavulich.

GeorgeGayno-NOAA commented 3 years ago

Brach at 6840076 was successfully tested on Orion, Hera, Jet, WCOSS-Cray and WCOSS-Dell. Will merge.

GeorgeGayno-NOAA commented 3 years ago

Merged to 'develop' at 1be59bd. Closing issue.