Changes for CICECORE load balance

[aoloso]

Changes were made to CICECORE in GEOS_CICE4ColumnPhysGridComp.F90 for load balancing.

[zhaobin74]

This has been tested with zero-diff results.

[sanAkel]

@aoloso would you please add some documentation (preferably inline, in GEOSagcm_GridComp/GEOSphysics_GridComp/GEOSsurface_GridComp/GEOSsaltwater_GridComp/BufferPacking.h and in GEOSagcm_GridComp/GEOSphysics_GridComp/GEOSsurface_GridComp/GEOSsaltwater_GridComp/BufferUnpacking.h) to describe what is:

• purpose
• how it is being accomplished

If you do not like to add that in the code, please have it in the PR "comments" section.

Thank you!

[mathomp4]

I believe this requires GEOS-ESM/MAPL#1650, right? If so, we should block this until a new MAPL is released.

[zhaobin74]

I believe this requires GEOS-ESM/MAPL#1650, right? If so, we should block this until a new MAPL is released.

@mathomp4, yes, it depends on that one. DNA label added.

[aoloso]

@sanAkel
I have added comments to GEOSagcm_GridComp/GEOSphysics_GridComp/GEOSsurface_GridComp/GEOSsaltwater_GridComp/[BufferPacking.h, BufferUnpacking.h] as you suggested.

[zhaobin74]

@aoloso, since LB_ON is zero-diff with LB_OFF but runs faster, how about making ON as default in the code? Could you do that?

[sanAkel]

Thank you @aoloso

[sanAkel]

@aoloso, since LB_ON is zero-diff with LB_OFF but runs faster, how about making ON as default in the code? Could you do that?

I support this request!

[aoloso]

@sanAkel I made it "TRUE" by default in this pull request. It was "FALSE" by default only for development when working off of "aogcm" branch.

[zhaobin74]

@sanAkel I made it "TRUE" by default in this pull request. It was "FALSE" by default only for development when working off of "aogcm" branch.

Thanks you, @aoloso.

[mathomp4]

Okay. MAPL 2.25 has been released:

https://github.com/GEOS-ESM/MAPL/releases/tag/v2.25.0

I suppose what is next is to have someone try this code with that tag. I'll make a PR to GEOSgcm to get MAPL 2.25 in there.