-
Notifications
You must be signed in to change notification settings - Fork 64
CCPP Framework Meeting Minutes 2024 06 13
Attendees: Mike Kavulich,
JESSE is taking notes.
CCPP Framework (issues, PRs, discussions)
-
- Updates? Has it been assigned to the merge queue?
- How does one check on the status of a PR in the merge queue anyway?
- Grant: PR is scheduled for testing on Monday, June 17th. If it passes then it should get merged in then.
- Updates? Has it been assigned to the merge queue?
-
Remove requirement that suite definition file names begin with "suite_" PR#569
- PRs are open for SCM and UFS weather model/FV3, but discussion about the specifics of CCPP suite naming continues
- If anyone would like to be involved in the discussion, we are meeting with UFS code developers who have some concerns about new naming guidelines at 11am Mountain Time, I can send you an invite.
- Michael W.: What were the actual concerns people had with this PR?
- Ligia: There were concerns that using bird names would hide information on what the suite was representing from a configuration standpoint. There was also some concerns on who would do the downstream work to update the suite names in the UFS Apps themselves.
- Dom: The actual framework PR itself is uncontroversial, as it is just removing the
suite_naming requirement from the framework (but a user could still usesuite_in their SDF filename if they wanted to). - Michael K.: This might cause a slight problem downstream becasue the prebuild version automatically strips out the
suite__text in the generated code, while capgen does not.
- PRs are open for SCM and UFS weather model/FV3, but discussion about the specifics of CCPP suite naming continues
Standard names (issues, PRs, discussions)
- (NCAR SEs) In dictionary, "turbulent_kinetic_energy" has units of "J" - should this be "J kg-1"?
- Group decided to rename the variable
specific_turbulent_kinetic_energywith units ofJ kg-1.
- Group decided to rename the variable
- (NCAR SEs) Unit conversion from "m2 s-2" to "J kg-1" ok?
- Group agreed that allowing the framework to recognize that
m2 s-2andJ kg-1were equivalent was a good idea. Long term it might help to have a dictionary of equivalent units, but in the short term just implementing a "convertor" that muliplies by one would be fine.
- Group agreed that allowing the framework to recognize that
Mike K is in the process of updating the "Development Workflow" wiki page. It's not done yet, but you're welcome to spy on my progress and raise objections to anything you see there.
We should schedule a backlog cleaning for a future meeting. Or we can start today if we have time. Several dozen ccpp-framework issues are more than a year old, and likely can either be closed as completed/won'tfix or have a more up-to-date label applied.
Ligia: EPIC has received funding to formalize the governance of CCPP Standard Names (and the associated repo), which will be done in collaboration with NCAR CGD.
Cheryl: Should we start trying to formalize group names amongst the various suites, so that one host model could use the SDF from a different host model?
- Dustin: That is an ambitious idea (but I like it).
- Dom: In the UFS the group calling structure is hardcoded in the host model itself.
- Dustin: SCM just runs through the groups in sequential order.
- Grant: I am not sure SDFs could ever be shared because the dynamics-physics coupling could be drastically different (and thus the types of groups and their ordering would also be drastically different).
- Cheryl: Sounds good, will go ahead and drop this concern then.
Dom: Blocked data structure PR should be tested soon, but there is a concern with optional arguments and the active attribute, specifically that some variables are being labeled "optional" even though they are technically required. Thus it would be good to discuss possible solutions with the core CCPP framework developers.
- Group decided to try and meet in person when Michael W. was in town.