Skip to content
Sign in
Builds

Changes

#49 (Jul 29, 2026, 3:32:06 AM)

  1. The old mixing length calculation seems to be making dycoms2_rf02 fail when using the bicgstab solver. This case was always unstable with bicgstab, so it seemed more like luck before that it was passing. This change just removes the failed output data to prevent pyplotgen from crashing — Gunther Huebler / githubweb

#48 (Jul 28, 2026, 12:14:34 PM)

  1. Recently a commit removed the l_tke_aniso flag, which is BFB unless that flag is enabled. That flag was used in the clubb_BFB_varying_flags_gfortran_test, causing it to fail even though we don't care since it only failed for a removed flag, so RESOLVED:8ab3902e812c2883976b4f1f57c085d682666d92 — Gunther Huebler / githubweb

#47 (Jul 28, 2026, 3:32:09 AM)

  1. Archiving experimental files that were used to explore some crazy PDF diagnosis methods. Much of this is sporatic LLM generated code and reports. These two methods were tested and did work well when using LES input to model the PDF, but were extemely noisy and unstable in free running clubb. The 'parcel tracking' one performs a mixing-length-like calculation to evolve a persistent 3rd gaussian, then computes a residual ADG1 pair to balance everything out and adhere to clubbs predicted moments. The second one (dubbed trivariate even though they're technically all trivariate) is a 2 Gaussian method that allows for many more degrees of freedom in how we place Gaussian centers and allocate covariance - it does this in part by taking into account rt and thl skewness, which require calculating rtp3 and thlp3. This trivate actually does really well when modelled using LES data, but it's heavy reliance on xp3 terms makes it extremely noisy and bad in clubb - this is why this code also includes a much more sophisticated (and barely tested) advance_xp3_module.F90, as an attempt to make this less worse for the trivariate pdf. — Gunther Huebler / githubweb
  2. Fixing bug in advance_xp3 code. This is BFB because advance_xp3 is hardcoded to false. — Gunther Huebler / githubweb

#46 (Jul 26, 2026, 3:32:07 AM)

  1. Setting l_diag_Lscale_from_tau to false by default, and changing it to true in the flag set list the run_bindiff_w_flags test uses so that we keep testing the tau code nightly. This flag changes answers, so this is BIT_CHANGING. — Gunther Huebler / githubweb
  2. Setting l_diag_Lscale_from_tau to false caused dycoms2_rf01_fixed_sst to start failing the python vs fortran test. The differences were small and occured at a late timestep, so to fix the test I put a limiter on the number of timesteps the test runs. — Gunther Huebler / githubweb
  3. Removing the l_tke_aniso flag — Gunther Huebler / githubweb
  4. Making les paths in _model.in files correct — Gunther Huebler / githubweb
  5. Adding prompt to automate updating host models — Gunther Huebler / githubweb

#45 (Jul 21, 2026, 3:32:11 AM)

  1. Clarify that the SILHS preprocessor flag in CLUBB is connected to the same flag in sam_clubb. — noreply / githubweb

#44 (Jul 15, 2026, 3:32:06 AM)

  1. Fixing SILHS usage so clubb/sam_clubb are compatible — Trevor Hancock / githubweb
  2. Fixing error/diff with morrison microphysics — Trevor Hancock / githubweb

#43 (Jul 14, 2026, 3:32:06 AM)

  1. Removed option to disable silhs — Trevor Hancock / githubweb
  2. Revert changes to src/Microphys/Morrison_microphys/microphysics.F90 — Trevor Hancock / githubweb

#42 (Jul 11, 2026, 3:32:10 AM)

  1. Removed quad precision option (#1330) — noreply / githubweb
  2. Removed option to disable netcdf (#1331) — noreply / githubweb
  3. Removed option to disable silhs (#1332) — noreply / githubweb

#41 (Jul 8, 2026, 3:32:08 AM)

  1. Fixing little path bug in the old tuner script and the thread test script — Gunther Huebler / githubweb
  2. Update dash app README with alternative run command — noreply / githubweb
  3. Local python venv for dash app (#1329) — noreply / githubweb

#40 (Jul 7, 2026, 3:32:10 AM)

  1. Remove old SVN version control identifier ($Id$) from tunable_parameters.in — noreply / githubweb
  2. Remove old SVN version control identifier from compatible_r8029/tunable_parameters.in — noreply / githubweb
  3. Remove blank line to test git update scripts — noreply / githubweb
  4. Clubb timer upgrade (#1328) — noreply / githubweb

#38 (Jul 5, 2026, 3:32:09 AM)

  1. Tiny dash updates (#1327) — noreply / githubweb

#36 (Jul 3, 2026, 3:35:10 AM)

  1. Tuner additions (#1324) — noreply / githubweb

#35 (Jul 2, 2026, 3:32:06 AM)

  1. Dash compile tab (#1323) — noreply / githubweb

#33 (Jun 26, 2026, 12:15:59 PM)

  1. Updating python contract audit to better detect API arg list discrepancies from the underlying fortran arg lists. Also adding a temporary handoff file and improving the python api fix shortcut — Gunther Huebler / githubweb
  2. Our argument_contract_audit.py script checks if the python_api matches the underlying fortran - this makes it more strict, and fixes many of the inconsistencies it found. — Gunther Huebler / githubweb
  3. advance_clubb_core.py cleanup and removing strange compat_kwargs usages that seem like artifacts of laziness — Gunther Huebler / githubweb
  4. Improving audit script even more, fixing issues it found. Also making advance_clubb_core.py nice and making an LLM shortcut for how to repeat the process — Gunther Huebler / githubweb
  5. Building varying flags bindiff into run_bindiff_all.py, this just makes sense and helps fix the varying flags BFB test — Gunther Huebler / githubweb

#30 (Jun 26, 2026, 3:32:09 AM)

  1. Cleaning up run_scripts (#1320) — noreply / githubweb
  2. Removing jacobian code - it has not worked in a long time — Gunther Huebler / githubweb
  3. Removing recl.inc, just dead code for an outdated issue — Gunther Huebler / githubweb
  4. Removing ifdefs that are only for E3SM and only on a specific machine — Gunther Huebler / githubweb
  5. Removing the CLUBBND_CAM option, not sure what it was for, it's not defined in this repo or any of our host model repos — Gunther Huebler / githubweb
  6. Removing unused variables — Gunther Huebler / githubweb

#29 (Jun 23, 2026, 4:47:34 PM)

  1. Recently clubb_BFB_varying_flags_gfortran_test failed due to a script bug, dispite it being BFB. RESOLVED:2a334f87922dfaceb7d3a37b1924d374f1b986d5 — Gunther Huebler / githubweb

#27 (Jun 18, 2026, 3:32:10 AM)

  1. Clubb loss driver (#1303) — noreply / githubweb

#26 (Jun 17, 2026, 3:32:08 AM)

  1. 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. — Gunther Huebler / githubweb

#25 (Jun 12, 2026, 3:32:07 AM)

  1. Update Jenkinsfile to match moved file — noreply / githubweb

#24 (Jun 11, 2026, 3:32:08 AM)

  1. Deleted the old compile directory and docs directory (#1315) — noreply / githubweb
  2. Update README, BUILD.md, and pyplotgen README.md (#1309) — noreply / githubweb

#23 (Jun 10, 2026, 3:32:06 AM)

  1. Removed/deleted outdated Matlab files (#1313) — noreply / githubweb
  2. Removed/Deleted GrADS (.ctl, .dat) file format and reference to it (#1310) — noreply / githubweb

#22 (Jun 5, 2026, 4:06:27 PM)

  1. Removed perl and references to it (#1311) — noreply / githubweb

#21 (Jun 5, 2026, 3:32:06 AM)

  1. Create README for python_sam_budgets_plotter folder — noreply / githubweb
  2. Removed plotgen (#1312) — noreply / githubweb

#20 (Jun 4, 2026, 3:32:09 AM)

  1. Clubb organization part3 (#1306) — noreply / githubweb
  2. GPU async script (#1307) — noreply / githubweb

#19 (Jun 3, 2026, 3:32:09 AM)

  1. Update JAX conversion plan for clarity and detail — noreply / githubweb

#18 (May 28, 2026, 1:18:58 PM)

  1. The last commit should've been marked as non-BFB since it corrected a crashing bug - so comparing output to no output caused the output to be considered different. This was missed because the BFB checking tests didn't fail, and that was because the commit that caused the crashing was marked non-BFB, so an unusual combo. In general though any commit that fixes crashing behavior should be labelled as bit changing. RESOLVED:73dc365c77ab4bf2405d753f167307ae9be330a0 — Gunther Huebler / githubweb

#17 (May 28, 2026, 3:32:08 AM)

  1. Removing unneccesary wp2_zt conversion that should've been removed already, but was accidentally added back in the last PR when trying to split BFB from non-BFB changes. This was an unused local variable so this does not affect anything. — Gunther Huebler / githubweb

#16 (May 27, 2026, 3:32:09 AM)

  1. Clubb organization part2 (#1304) — noreply / githubweb

#15 (May 23, 2026, 3:32:06 AM)

  1. Improvements to get clubb working on mac (#1305) — noreply / githubweb

#14 (May 20, 2026, 3:32:06 AM)

  1. Removing a couple variables from openacc statements to fix the clubb_driver_test - these were causing stats errors on the GPU for chi and chip2. This is technically BIT_CHANGING. — Gunther Huebler / githubweb

#13 (May 19, 2026, 3:32:10 AM)

  1. Clubb organization part1 (#1302) — noreply / githubweb

#12 (May 14, 2026, 1:30:45 PM)

  1. This fixes the GPU BFB test that failed due to stats-only differences in rtm_ed and thlm_ed, and as a result is technically BIT_CHANGING. Also, the BFB with varying flags failed because flag set 1 changed to use the lapack solvers since the custom ones it set before are now default. RESOLVED:34c6bb07afb6e445f60ea87dcc632969086f9fa1 — Gunther Huebler / githubweb

#11 (May 14, 2026, 3:32:07 AM)

  1. advance_clubb_core cleanup (#1301) — noreply / githubweb

#10 (May 8, 2026, 3:32:08 AM)

  1. Switching default call to advance_clubb_core in python and jax drivers to use the python api version, rather than the python ported version. — Gunther Huebler / githubweb

#8 (May 5, 2026, 5:16:49 PM)

  1. Fixing python api and python/jax driver. Also fixing bug that caused non-BFBness in stats outputs (only upwp_cl/vpwp_cl) when using l_predict_upwp_vpwp=.false., and adding ulimit -s 8388608 to fix GPU test (though it's unclear why exactly we need this all of a sudden for this test). RESOLVED:643d128a5613019595aecd643f1028e937ad4b76 RESOLVED:f146255f50732c445ccbd012ca45888a9ce44d0b — Gunther Huebler / githubweb

#7 (May 5, 2026, 3:32:06 AM)

  1. Clip covar simplifications (#1297) — noreply / githubweb
  2. Cmake netcdf fallback (#1294) — noreply / githubweb

#6 (Apr 9, 2026, 3:32:08 AM)

  1. Small fix to python api string handling. There was an issue that crept up with this when running on mitch - this should make the string/char array handling more robust — Gunther Huebler / githubweb
  2. Python update (#1291) — noreply / githubweb
  3. Remove line to test python scripts — noreply / githubweb
  4. Add blank line to test python scripts — noreply / githubweb

#5 (Apr 8, 2026, 3:32:08 AM)

  1. Adding err_info back to stats_end_timestep_api — Gunther Huebler / githubweb
  2. Fixing monoflux limiter test - there was a comment that was split across lines, which broke the test — Gunther Huebler / githubweb
  3. Removing unused variable — Gunther Huebler / githubweb
  4. Adding -fresh flag to compile.py to remove the previous build before compiling — Gunther Huebler / githubweb

#4 (Apr 7, 2026, 3:32:08 AM)

  1. Clubb f2py (#1274) — noreply / githubweb
  2. Little fixes for python code — Gunther Huebler / githubweb
  3. Remove blank line to trigger git update script — noreply / githubweb
  4. Added blank line to trigger git update scripts — noreply / githubweb
  5. Removed blank line to trigger git update scripts — noreply / githubweb
  6. Add blank line to trigger git update scripts — noreply / githubweb
  7. Add blank line to trigger git update scripts — noreply / githubweb
  8. Add blank line to trigger git update scripts — noreply / githubweb

#3 (Apr 6, 2026, 2:02:57 PM)

  1. Standards fix (#1288) — noreply / githubweb