#23 (Jun 17, 2026, 7:27:40 PM)

Started by user Gunther Huebler

• refs/remotes/origin/clubb_loss_driver

1. Deleted the old compile directory and docs directory (#1315) (details / githubweb)
2. Update README, BUILD.md, and pyplotgen README.md (#1309) (details / githubweb)
3. Update Jenkinsfile to match moved file (details / githubweb)
4. Fixing bug that caused incorrect times to be written to netcdf in cases where stats_tout (output interval) > dt_main (timestep). This is BIT_CHANGING because of that, but only for the time variable (no field differences), and only for cases where we don't output each timestep. (details / githubweb)
5. First set of commits for the new loss driver. (details / githubweb)
6. Improvements (details / githubweb)
7. Batching capabilities (details / githubweb)
8. Lot's of cleanup, simplifications, and consolidation (details / githubweb)
9. More testing (details / githubweb)
10. Python interfaces for the loss driver (details / githubweb)
11. Interface for tuner and tuner with modes. (details / githubweb)
12. Nice updates (details / githubweb)
13. Missed file (details / githubweb)
14. Little updates to fix inconsistencies in the plots and averaging ranges (details / githubweb)
15. Adding a script that converts sam and coamps netcdf files to a clubb named version, including things that need to be calculated. This fixes a bug where fields were being averaged before being calculated, see https://github.com/larson-group/clubb/issues/1284, but is currently only wired up in the new tuner and the dash app plotter (details / githubweb)
16. Making some things public for nvhpc (details / githubweb)
17. Organization and consistency test fix (details / githubweb)
18. Adding taylor diagram to tune tab, this needs taylor metrics to be computed so the loss driver needed to output this info. (details / githubweb)
19. Adding more fancy loss functions based on Taylor metrics and a way to switch how the loss is computed in the dash tuner (details / githubweb)
20. Making NaN in code not an issue (details / githubweb)
21. Mostly dash app improvements to plotting, also made subwindows for tuning (details / githubweb)
22. Tuner tracking by case (details / githubweb)
23. Removing stats time-variance code (details / githubweb)
24. Simplifying loss driver output - now loss metrics are output explicitly rather than as packaged taylor_metrics and having different versions for subwindows and such (details / githubweb)
25. Mainly ui updates and small simplifications (details / githubweb)
26. Improving comments and cleaning things up a bit (details / githubweb)
27. Removing strictness on stats window vs output interval, and adding a stats consistency test to tests for issues with batch size and output frequency and the windowing mode (details / githubweb)
28. Fixing openmp bug and making incomplete stats intervals not output, as it was before (details / githubweb)
29. Making jenkins test use standard stats, and fixing tiny bug that caused error running with debug+no rcp2 stats (details / githubweb)
30. Adding gate around bindiff, there's an edge case where a cosmetic diff can trigger a full-diff even in a failed case, which should otherwise be considered BFB and skipped. (details / githubweb)
31. Adding some README entries for the dash app and a python requirements file (details / githubweb)
32. Adding comments and little bug fix to dash app (details / githubweb)
33. Update to loss consistency test, making it simpler since stats consistency test covers most things. Also adding it to the stats consistency jenkins test since it is still similar (details / githubweb)
34. Cleanup and code consolidation, new jenkins test for tuner (details / githubweb)
35. Adding comments (details / githubweb)
36. Fixing little bugs (details / githubweb)