support running landice/greenland/mesh_gen test case on workstations

[cwsmith]

This PR passes the parallel_executable to a few functions that were hardcoded to use srun. With the right input files, this allows testing on workstations with limited memory and cores.

Testing was done with subsampled GIS inputs (to 4km, thanks to @matthewhoffman) for the landice/greenland/mesh_gen test.

Checklist

• User's Guide has been updated - no change needed
• Developer's Guide has been updated - no change needed
• API documentation in the Developer's Guide (api.rst) has any new or modified class, method and/or functions listed - no changes needed - the doc string for interp_gridded2mali was updated in compass/landice/mesh.py
• Documentation has been built locally and changes look as expected - the new parallel_executable arg is rendered in the html
• not applicable The E3SM-Project submodule has been updated with relevant E3SM changes
• not applicable The MALI-Dev submodule has been updated with relevant MALI changes
• Document (in a comment titled Testing in this PR) any testing that was used to verify the changes
• not applicable? New tests have been added to a test suite

[cwsmith]

@matthewhoffman Can this be reviewed/merged? It looks like the failing CI tests are unrelated to the changes and caused by package install failures during creation of the python environment.

[matthewhoffman]

@cwsmith , yes, I will look at it closer today. Sorry for letting it drop!

[cwsmith]

No worries. Thank you.

[matthewhoffman]

@cwsmith , we'll have to add corresponding changes to compass/landice/tests/antarctica/mesh.py (which also calls interp_gridded2mali). I'm happy to do that if you'd prefer not to.

[cwsmith]

@matthewhoffman Whoops. I didn't consider the other landice meshgen test. Sorry about that. I'll make those changes.

[cwsmith]

Sorry for the multiple small commits. The pre-commit hook setup/build phase is failing in my environment for some reason...

[cwsmith]

It looks like the github actions are all passing. Does it or some other CI run the meshgen cases?

[xylar]

No, CI just runs some very low level tests to make sure the conda environment isn't broken. It doesn't run any compass tests. Most tests are too big for CI.

[matthewhoffman]

Thanks for the additional changes, @cwsmith !