Features for EarthWorks v2.1.001 Patch Release

[gdicker1]

This release will address some issues from the v2.1 release and include some new features.

• Addresses CPU builds of the FullyCoupled compset that were broken in v2.1 due to MPAS infrastructure and framework changes for MPAS-A OpenACC. (Fixes fully coupled model does not compile or run due to mpas-framework issues. #31)
• Addresses issues with dependence on CICE by adding a "prescribed mode" for MPAS-seaice (Fixes ice_boundary.F90 internal compiler error under nvhpc 23.11 compiler  #23)
• Incorporates a GPU version of RRTMGP by incorporating upstream changes into EarthWorksOrg/CAM

Externals Affected and PRs:

• MPAS-framework: mpasfrwk-ew2.1.000 -> mpasfrwk-ew2.1.001
  • Modify mpas framework to match mpasa in EW v2.1 mpas-framework#6
• MPAS-seaice: mpassi-ew2.1.000 -> mpassi-ew2.1.001
  • Prescribed seaice mode mpas-seaice#8
• CAM: cam-ew2.1.000 -> cam-ew2.1.001 -> cam-ew2.1.002
  • Incorporate upstream changes for RRTMGP CAM#14
  • Fix intel-oneapi builds of CAM that include changes from cam6_3_147 or newer CAM#15

EarthWorksOrg/EarthWorks PR for this change: #32

Tests Needed (once External PRs are ready):

• Demonstrate FullyCoupled runs are fixed on CPUs
• Demonstrate compset runs with MPASSI%PRES
• Demonstrate ability to run (smoke-test) CPU version of RRTMGP that is brought in
  - [ ] Demonstrate ability to run CPU FHS94 at 120km on Perlmutter
  • Copy data from Derecho to Perlmutter into m4180 project space

Documentation/Writing Needed:

• Document steps to run compsets with MPASSI%PRES
  - [ ] Update documentation to run on Perlmutter
• Release notes for v2.1.001 Release

---

(NOTE: this top level comment is subject to edits based on the discussion. To respond to a particular point, please highlight some text in a comment and use the "Quote reply" feature.)

[gdicker1]

• Copy data from Derecho to Perlmutter into m4180 project space

@supreethms1809 solved this, input data on Perlmutter is currently located within "/global/cfs/cdirs/m4180/inputdata"

[gdicker1]

• Demonstrate FullyCoupled runs are fixed on CPUs

Based on runs @dazlich has done, I think this is complete

[gdicker1]

Changed references to Perlmutter from top comment based on changes in #32

[gdicker1]

• Demonstrate ability to run (smoke-test) CPU version of RRTMGP that is brought in

See this comment in EWOrg/CAM #14. Marking complete

[gdicker1]

Finalization of this release is in #37