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

ufs_public_release: update version number for CCPP, disable dynamic CCPP, cleanup work and bug fixes for regression tests #25

Conversation

climbfuji
Copy link
Collaborator

@climbfuji climbfuji commented Jan 7, 2020

  • cleanup regression tests: remove unused regression test configurations, run scripts and namelist templates
  • add CCPP and IPD regression tests for comparison in REPRO mode (GFSv15p2 and GFSv16beta)
  • disable inline post on all supported platforms
  • fix regression tests: correct run scripts, check for output from tests in debug mode
  • convert namelist templates to unix line endings
  • disable dynamic CCPP build
  • update date tag for regression test baseline
  • update submodule pointer for the stochastic_physics submodule
  • update of gaea modulefiles and regression test baseline directories (from develop)
  • update of cheyenne default queue (from develop)

With these changes,

  • results are bit-for-bit identical between the IPD and CCPP tests in REPRO mode on Hera and Cheyenne using the Intel compiler,
  • results are bit-for-bit identical between the IPD and CCPP tests in REPRO mode on macOS using the GNU compiler (1hr fcst only instead of the usual 24h fcst)
  • all tests complete in DEBUG mode instead of crashing, silent failures are avoided by explicitly checking for the output files (see issue Regression tests that are not looking for specific output files can fail silently #23)

Known issue: The results differ between the IPD and CCPP tests in REPRO mode on Cheyenne using the GNU compiler for 24hr fcst runs. I have not tested if the results are still identical after 1hr (since they are on macOS). For this reason, the IPD verification against the CCPP results in REPRO mode are turned off for Cheyenne+GNU.

…ons, run scripts and namelist templates

- add CCPP and IPD regression tests for comparison in REPRO mode
- disable inline post on all supported platforms
- fix regression tests: correct run scripts, check for output from tests in debug mode
- convert namelist templates to unix line endings
- disable dynamic CCPP build
- update date tag for regression test baseline
@climbfuji
Copy link
Collaborator Author

@climbfuji
Copy link
Collaborator Author

New regression test baselines have been created on Cheyenne and are currently located in:

/glade/p/ral/jntp/GMTB/NEMSfv3gfs/RT/TMP_ufs-public-release-20200107

On hera, they are in:

/scratch1/NCEPDEV/stmp4/Dom.Heinzeller/FV3_RT/REGRESSION_TEST_UFS_PUBLIC_RELEASE_INTEL

Once the code review is complete, I can rename the Cheyenne baseline and EMC will need to move the Hera baseline into the right place before running the regression tests once more and updating the logs in the ufs-weather-model PR.

… for cheyenne.gnu, update regression baseline date tag
@climbfuji
Copy link
Collaborator Author

All regression tests have passed and the logs have been updated in the PR.

@climbfuji
Copy link
Collaborator Author

@DusanJovic-NOAA this is ready from my side. Took me three attempts to revert the .gitmodules change correctly.

@DusanJovic-NOAA DusanJovic-NOAA merged commit bde62f9 into ufs-community:ufs_public_release Jan 13, 2020
climbfuji pushed a commit to climbfuji/ufs-weather-model that referenced this pull request Jun 5, 2020
…_other_updates

Add drag suite and other updates
LarissaReames-NOAA pushed a commit to LarissaReames-NOAA/ufs-weather-model that referenced this pull request Oct 22, 2021
Bugfix for regional runs when dycore is compiled in double precision
MinsukJi-NOAA pushed a commit to MinsukJi-NOAA/ufs-weather-model that referenced this pull request May 26, 2022
epic-cicd-jenkins pushed a commit that referenced this pull request Apr 17, 2023
…for which to build FV3 from valid_vals_CCPP_PHYS_SUITES (#25)

This PR must be merged at the same time as PR #[301](https://github.com/NOAA-EMC/regional_workflow/pull/301) into NOAA-EMC/regional_workflow.

## DESCRIPTION OF CHANGES:
* For stability, freeze version of UFS_UTILS in Externals.cfg to a specific hash.
* In build_forecast.sh, get the list of CCPP physics suites for which to build the FV3 executable from the array valid_vals_CCPP_PHYS_SUITES in the file valid_param_vals.sh in regional_worfklow.
* Delete build_global_equiv_resol.sh and build_mosaic_file.sh since those scripts are no longer needed.  The mosaic_file code has been removed, and the global_equiv_resol code is now in UFS_UTILS and is built in/by that repo.
* For consistency with the name of the code, change the name of the executable for building a regional grid that gets copied to ufs-srweather-app/exec from regional_grid to regional_esg_grid.
* Remove unneeded lines from regional_build.cfg.

## TESTS CONDUCTED: 
See PR #[301](https://github.com/NOAA-EMC/regional_workflow/pull/301) into NOAA-EMC/regional_workflow for tests conducted.
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.

2 participants