-
Notifications
You must be signed in to change notification settings - Fork 64
CCPP Framework Meeting Minutes 2024 01 22
Attendees: Mike Kavulich,
CCPP Framework (issues, PRs, discussions)
-
PR 519: ccpp_prebuild.py: add support for chunked arrays
- Ready to merge after Courtney reviews?
- Associated issue; Issue 520: ccpp_prebuild.py: add capability to use chunks of contiguous arrays in the run phase
-
Issue 521: Add check to disallow usage of horizontal_dimension in the run phase
- Needed further discussion within capgen group, updates?
- PR 523: Add support to use mpi_f08 MPI module
- Need PR for https://github.com/climbfuji/ccpp-framework/tree/feature/dummy_allocate_inactive
- Part of the conditionally-allocated array PRs for -check all NCO requirement (https://github.com/NOAA-EMC/fv3atm/pull/767)
Standard names (issues, PRs, discussions)
-
PR 56: Add new standard names from NCAR idealized physics schemes
- Needs approval from Mike, Grant
-
Issue 57: Feedback on the meaning of: sea_ice_area_fraction_of_sea_area_fraction
- Some proposal for standard name change
- Failing CI tests
- Appears to be a capgen test that is failing on the main branch: https://github.com/NCAR/ccpp-framework/actions/runs/7494120637/job/20401325581
- Proposal for adding new metadata header property for getting constituent number at runtime [Courtney]
- Long-term plan for prebuild testing in capgen (including SCM testing)? [Michael W]
- Timelines for PR reviews
- 1 week from initial review, code manager will send out reminder
- Ways to track review requests centrally?
- SRW uses a GitHub project: https://github.com/orgs/ufs-community/projects/38/views/1
Attendees: Mike Kavulich, Dustin Swales, Cheryl Craig, Dom Heinzeller, Ligia Bernardet, Michael Waxmonsky, Courtney Peverly, Grant Firl, Matt Dawson
Framework PRs/Issues/Discussion: #519 Support for chunked arrays. Ready to merge? Status in the UFS? All RTs are not passing. #521 Disallow horizontal_dimension in run phase. #523 MPI_F08 What is best solution?
Allocate to zero PR? We’ve discussed this, but it would be good for the Capgen/CESM folk take a look. Ligia and Grant, RRFS code freeze is imminent. This needs to get in sooner than later. This alloc(0) change is a stopgap fix, not a permanent solution. Dom will setup a meeting for us to discuss this later this week.
Standard Names PRs/Issues/Discussion: #56 Mike will finish up reviewing this week. #57 Change to standard name for clarity
New items for discussion: Failing CI tests? Courtney, could be related to py3.7 vs. py3.8. New metadata header property. Dynamic determination for number of tracers.
Chat:
https://github.com/NCAR/ccpp-framework/wiki/CCPP-Framework-Meeting-Minutes-2024-01-22 Dom Heinzeller3:09 PM "use mpi" is f90 "use mpi_f08" is f08 "include mpi.h" is f77 Dom Heinzeller3:29 PM they are not allowed to query the allocation status Dustin Swales - NOAA Federal3:30 PM says who? Dom Heinzeller3:30 PM CCPP requirements doc Dustin Swales - NOAA Federal3:31 PM I can't find this anywhere? Grant Firl - NOAA Affiliate3:33 PM Isn't it also necessary for automatic unit conversions? Grant Firl - NOAA Affiliate3:39 PM Yes. That is a drawback! Grant Firl - NOAA Affiliate3:41 PM Now, it's checking for size > 0 instead of allocated(). I suppose it's still against the CCPP ethos. Dom Heinzeller3:50 PM Is that attribute optional, i.e. when not present assumed to be empty (so that models that don't use the constituent logic don't need to have it)?
And is it a host-model attribute only? Dom Heinzeller3:55 PM More generaL?
type=scheme set_parameters_subroutines=sub_A,sub_B, ...
? Ligia Bernardet - NOAA Federal3:56 PM I need to head out. Thanks, folks