Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Release/public-v3: Add WoFS suite definition file for SRW v3 release #535

Merged
merged 9 commits into from
May 25, 2022

Conversation

MicroTed
Copy link
Contributor

Description

Adds SDF file suite_FV3_WoFS_v0.xml for SRW public-v3

Testing

Dependencies

Newer CCPP has an update to the NSSL microphysics that fixes a restart and MPI decomposition issues. We can point to a newer commit for that (which needs a couple other updates to work, however).

@junwang-noaa junwang-noaa changed the title Add WoFS suite definition file for SRW v3 release Release/public-v3: Add WoFS suite definition file for SRW v3 release May 12, 2022
@junwang-noaa
Copy link
Collaborator

@MicroTed It looks to me that the ccpp physics in your fv3atm branch pub-v3-wofssdf is pointing to an old revision of ccpp physics public-v6 branch. You may need to update the fv3 submodule to use the latest ccpp physics publib-v6 branch (revision 7e35351) and run the Wofs test. Also just let you know that your ccpp fixes for the WoFS testsin ccpp physics PR#904 is not in the public-v6 release branch yet. @grantfirl FYI

https://github.com/NCAR/ccpp-physics/commits/release/public-v6

@grantfirl
Copy link
Collaborator

@MicroTed It looks to me that the ccpp physics in your fv3atm branch pub-v3-wofssdf is pointing to an old revision of ccpp physics public-v6 branch. You may need to update the fv3 submodule to use the latest ccpp physics publib-v6 branch (revision 7e35351) and run the Wofs test. Also just let you know that your ccpp fixes for the WoFS testsin ccpp physics PR#904 is not in the public-v6 release branch yet. @grantfirl FYI

https://github.com/NCAR/ccpp-physics/commits/release/public-v6

@junwang-noaa @MicroTed OK, sorry for the delay. Let me update the release/public-v6 branch of ccpp-physics so that this can point to the latest... Stand by.

@grantfirl
Copy link
Collaborator

@MicroTed @junwang-noaa I've updated the release/public-v6 branch with commits that belong there to the best of my knowledge, including Ted's PR#904 bugfixes in ccpp-physics. The commit at the top of that branch is now: NCAR/ccpp-physics@b3e0944

I've also noticed that the .gitmodules file in this repo is still pointing to the main branch of ccpp-physics rather than `release/public-v6'. I'm guessing this should be updated too?

@MicroTed
Copy link
Contributor Author

I've also noticed that the .gitmodules file in this repo is still pointing to the main branch of ccpp-physics rather than `release/public-v6'. I'm guessing this should be updated too?

Thanks, @grantfirl I saw that, too. I think that commit 7e35351 (one scheme one file) has companion changes in framework. If you want me to add all those into this PR, I can do that, or it could be a separate PR and then merge in this one (which only adds the SDF).

@grantfirl
Copy link
Collaborator

@MicroTed The "one scheme to one file" changes were already in the ccpp-physics release branch before I just updated it. But, you're right that there are changes required for that code. However, the changes are in the fv3atm repo and not the ccpp-framework. This PR (#519) should be put into the fv3atm release branch. I'm not sure why it wasn't already there. The PR branch for this change (https://github.com/grantfirl/fv3atm/tree/one_scheme_to_one_file) still exists in my fork, so it should be easy just to merge that branch into MicroTed:pub-v3-wofssdf and push back to solve this issue.

Regarding ccpp-framework, it is my understanding that the release is just going to take whatever is in the top of main when the code is really frozen. Nothing that effects how the code generation, etc. works will be changed between now and the release. So, for the purpose of this PR, I would just checkout the latest ccpp-framework main and add it your fv3atm PR branch so that it is at least pointing to the latest.

@MicroTed
Copy link
Contributor Author

@grantfirl Correct that "one scheme" was there already in public-v6, but the v3 release was pointing to an earlier commit.
I'll update to the current public-v6 CCPP (and .gitmodules) and ccpp-framework (main). I cherry-picked your corresponding commit that updated ccpp_prebuild_config.py, and it all compiles. I am now running RT for rrfs_v1nssl on hera, then will do the regional_3km_wofs

@junwang-noaa
Copy link
Collaborator

@ligiabernardet @ChunxiZhang-NOAA Please see the updates in fv3atm/ccpp public_v3 branch. FYI,

@grantfirl
Copy link
Collaborator

The updates all look correct to me. Thanks, @MicroTed .

@MicroTed
Copy link
Contributor Author

Updates done, and RT/ORT tests run. The dcp test for regional_3km_wofs is successful with nrows_blend=0. (It should fail with nrows_blend=10). (Also updated PR ufs-community/ufs-weather-model#1220 )

@jkbk2004
Copy link
Collaborator

@MicroTed can you sync up your branch with NOAA-EMC:develop? 8 commits behind NOAA-EMC:develop. so that @daniel-hildebrandt can take next steps for merging.

@MicroTed
Copy link
Contributor Author

can you sync up your branch with NOAA-EMC:develop? 8 commits behind NOAA-EMC:develop. so that @daniel-hildebrandt can take next steps for merging.

@jkbk2004 That should be fine. Just to be clear, you are wanting the public-v3 to catch up to the current develop branch?

@jkbk2004
Copy link
Collaborator

can you sync up your branch with NOAA-EMC:develop? 8 commits behind NOAA-EMC:develop. so that @daniel-hildebrandt can take next steps for merging.

@jkbk2004 That should be fine. Just to be clear, you are wanting the public-v3 to catch up to the current develop branch?

@MicroTed I think it depends on the SRW release plan. @daniel-hildebrandt can you make sure release/public-v3 is the one created for srw release and should be synced with develop?

@jkbk2004
Copy link
Collaborator

can you sync up your branch with NOAA-EMC:develop? 8 commits behind NOAA-EMC:develop. so that @daniel-hildebrandt can take next steps for merging.

@jkbk2004 That should be fine. Just to be clear, you are wanting the public-v3 to catch up to the current develop branch?

@MicroTed I think it depends on the SRW release plan. @daniel-hildebrandt can you make sure release/public-v3 is the one created for srw release and should be synced with develop?

@junwang-noaa If we don't need to maintain RT tests with the release/public-v3 branch (SRW tests will be maintained by the SRW group), @daniel-hildebrandt can merge in this PR to the release branch, right? Ted has PR#514 to develop branch. So I think we can follow up separately later.

@junwang-noaa
Copy link
Collaborator

@jkbk2004 I am not sure who will be testing/maintaining the RT test for CCPP v6 and SRW public release. @grantfirl @ligiabernardet @JeffBeck-NOAA @arunchawla-NOAA FYI.

@jkbk2004
Copy link
Collaborator

@jkbk2004 I am not sure who will be testing/maintaining the RT test for CCPP v6 and SRW public release. @grantfirl @ligiabernardet @JeffBeck-NOAA @arunchawla-NOAA FYI.

@junwang-noaa @daniel-hildebrandt For CCPPv6/SRW-release cases, EPIC should support the RT tests. Then, sync release/public-v3 to develop will be much easier way.

@MicroTed
Copy link
Contributor Author

Note that there is a separate PR to add the suite for the release branch:
#514
But it has been held up by unrelated ORT issues for the regional.
So I was requested to open a separte PR for v3 to make sure it had the suite.

@daniel-hildebrandt
Copy link

Note that there is a separate PR to add the suite for the release branch: #514 But it has been held up by unrelated ORT issues for the regional. So I was requested to open a separte PR for v3 to make sure it had the suite.

@MicroTed Would you be able to verify that this is up-to-date with the release branch? The SRW release team wanted to ensure that the release/branch was up-to-date with develop.

@jkbk2004
Copy link
Collaborator

Note that there is a separate PR to add the suite for the release branch: #514 But it has been held up by unrelated ORT issues for the regional. So I was requested to open a separte PR for v3 to make sure it had the suite.

@MicroTed Would you be able to verify that this is up-to-date with the release branch? The SRW release team wanted to ensure that the release/branch was up-to-date with develop.

@MicroTed Once pub-v3-wofssdf is synced up to date, we can merge in this PR to release/public-v3.

@MicroTed
Copy link
Contributor Author

I tried merging with develop, but then there are cmake errors (e.g., looks for physics/smoke), so I think CCPP will also need to be updated?

@jkbk2004
Copy link
Collaborator

I tried merging with develop, but then there are cmake errors (e.g., looks for physics/smoke), so I think CCPP will also need to be updated?

@MicroTed Your branch is using ccpp/release/public-v6. Can you switch to ccpp/develop? On the other hands, srw/release/public-v2 still points to ccpp/develop (not v6 yet). So, I think it will be ok to merge in your PR with ccpp/develop. @daniel-hildebrandt We need to work on the branch pointer issue of srw/release/public-v2.

@MicroTed
Copy link
Contributor Author

OK, I saw that release/public-v3 had been updated, so I merged to that. So I think (hope) it is up to date now.

@jkbk2004 jkbk2004 merged commit fbc3c9a into NOAA-EMC:release/public-v3 May 25, 2022
grantfirl added a commit to grantfirl/fv3atm that referenced this pull request Jun 3, 2022
…OAA-EMC#535)

* Add WoFS suite definition file for SRW v3 release

* Set ccpp/physics submodule

* Set pointer for ccpp/physics

* Update ccpp_prebuild_config.py for one-to-one CCPP physics schemes/files (NOAA-EMC#519)

* Move up to current CCPP public-v6

* Minor bugfixes for CCPP v6 (NOAA-EMC#541)

* Point .gitmodules ccpp to 'main' branch

Co-authored-by: Grant Firl <grant.firl@noaa.gov>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants