CCPP Framework Meeting Minutes 2023 11 27

Agenda

• GitHub issues/PRs
• New items
• Updates
• Other business
  • notes

---

Attendees: Mike Kavulich,

Main discussion topic: NCO compliance issues

AQM operational implementation of CCPP was rejected due to runtime failures when the gfortran -check all compilation flag is used. This is required for compiling in DEBUG mode as specified on page 19 of the latest NCO standards document. Grant provided an example of this failure based on his tests:

242: forrtl: severe (408): fort: (8): Attempt to fetch from allocatable variable JINDX1_AER_LOCAL when it is not allocated
242:
242: Image              PC                Routine            Line        Source
242: fv3.exe            000000000BD11D1F  Unknown               Unknown  Unknown
242: fv3.exe            0000000008DF5EAC  ccpp_fv3_gfs_v15p        3008  ccpp_FV3_GFS_v15p2_time_vary_cap.F90
242: fv3.exe            00000000084AEA76  ccpp_fv3_gfs_v15p         160  ccpp_FV3_GFS_v15p2_cap.F90
242: fv3.exe            0000000008109201  ccpp_static_api_m         233  ccpp_static_api.F90
242: fv3.exe            00000000081010D1  ccpp_driver_mp_cc         108  CCPP_driver.F90
242: fv3.exe            00000000042B0F3E  atmos_model_mod_m         762  atmos_model.F90
242: fv3.exe            0000000003C8ECBC  module_fcst_grid_         764  module_fcst_grid_comp.F90

• A fix that is likely not CCPP-compliant, so probably will be specific to this release branch, can be found here: https://github.com/ufs-community/ccpp-physics/pull/132

• Possibly related to https://github.com/NCAR/ccpp-framework/issues/513

• How are we only finding out about this now? Is debug mode not tested in the regression tests? Is it specific to some operational settings? How can we avoid problems like this in the future?

Secondary discussion topic: failing CI tests

• Grant is testing a fix: looks like it was due to outdated/deprecated python features in CI test script?

GitHub issues/PRs

CCPP Framework

• PR 495: Constituents
  • Looks almost ready! Just waiting on some final requested changes from Dom.
• PR 504: Capgen Transforms
  • Looks like CI tests are failing?
• Issue 511: Implement CI for ccpp_track_variables.py
  • Assigned myself, I would like to postpone until after capgen is merged to main branch.
• PR 512: Add debug switch to capgen (perform variable allocation checks etc) - OPEN FOR DISCUSSION
• Issue 513: Should capgen support pointer metadata?

Standard names

• Issue 42: Want Github Action to ensure standard names are following character rules
  • PR 52: Fix invalid standard names, add standard name checking script and github action
    • Still needs one more review before merge
• Issue 50: Add suffix to indicate a 3D field has an additional point above the model top
  • Related: PR 51 (draft): Add including_point_above_model_top suffix
  • Still waiting on more discussion?

New items for discussion

Updates from last time

Previous notes

Other business

Meeting notes