7b2601b | Mark Taylor | 12 January 2022, 16:43:37 UTC | Merge branch 'mt5555/remove-deta' (PR #4724) Remove two unused variables. Causes roundoff level changes in some standalone HOMME tests due to change in loop structure and size of hyvcoord struct. [BFB] except for some standalone HOMME tests | 12 January 2022, 16:43:37 UTC |
c657b6f | Jon Wolfe | 11 January 2022, 19:44:24 UTC | Merge branch 'alicebarthel/ocn/fix-ml-heat-budget' (PR #4715) Fix calculation of the ML-averaged Brunt-Vaisala frequency This PR is correcting isuues in the calculation of the mixed-layer-averaged Brunt-Vaisala frequency Changes included: * corrected last layer thickness to partial in bruntVaisalaFreqML calculation * fetch index_temp as best practice (instead of assuming index_temp=1) [BFB] | 11 January 2022, 19:44:24 UTC |
5ec00ed | Mark Taylor | 09 January 2022, 22:09:55 UTC | Merge branch 'mt5555/remap-options2' (PR# 4722) Remove obsolete/unused remap algorithms add q_alg=11 ( from @ambrad - more accurate, doesn't impose monotonicity at boundaries) added option to specify dynamics algorithm independently from tracer algorithm (from Jason Torchinsky summer intern project) q_alg=11 fixes long standing issue with noise in HOMME's internal gravity wave test cases [BFB] except for one HOMME test, which was modified to test the new code. | 09 January 2022, 22:09:55 UTC |
d97bfd2 | James Foucar | 06 January 2022, 21:45:47 UTC | Update CIME submodule ... to bc80870bfa0b08d154e89b1fd46493cad1b39a02 Changes: 1) Split CESM configs out of repo 2) Adds enforcing code formatting to pre-commit 3) Refactor all CIME internal testing, split into smaller files and move to a more-logical dir 4) Cleanup Scorpio I/O performance data in the run directory 5) Remove documentation sections for pieces no longer included in CIME 6) Improve Macro.make generation Fixes: *) Loads of small fixes related to recent code churn. 1) Cleanup Scorpio I/O performance data in the run directory [BFB] | 07 January 2022, 22:43:12 UTC |
e7beed5 | Mark Taylor | 07 January 2022, 19:55:27 UTC | update unit test after q_alg=1,2 removal | 07 January 2022, 21:00:39 UTC |
b3e8364 | Alice Barthel | 07 January 2022, 17:36:11 UTC | modified comment on partial thickness to clarify | 07 January 2022, 17:36:11 UTC |
ccea41d | Mark Taylor | 07 January 2022, 16:41:32 UTC | remove unneeded reamp_alg argument from compute_ppm() | 07 January 2022, 16:41:32 UTC |
61214eb | Peter Schwartz | 06 January 2022, 22:28:44 UTC | Merge branch 'bishtgautam/lnd/mkmapdata' (PR #4602) The script that lists all the maps needed for creating a new ELM surface dataset is fixed to account for the name change of the land model to ELM ELM tools directory is reorganized by moving all tools from clm4_5 and shared directory to one-level up. The old directory structure was inherited from CESM in which tools supported two models (clm4_0 and clm4_5) with a common shared directory. [BFB] | 06 January 2022, 22:28:44 UTC |
e7789f5 | Mark Taylor | 06 January 2022, 21:11:26 UTC | Merge remote-tracking branch 'mt5555/cleanup' (PR #4721) Remove a lot of obsolete code three minor bug fixes: lap_p_wk array dimension, Nthreads=-1 logic: only enable threads if HORIZ_OPENMP or COLUMN_OPENMP enabled parseE3.py typo [BFB] | 06 January 2022, 21:11:26 UTC |
72e8ca9 | Mark Taylor | 06 January 2022, 01:24:38 UTC | remove unused variables | 06 January 2022, 01:24:38 UTC |
ddebdcb | Mark Taylor | 04 January 2022, 23:44:32 UTC | improve 10/11 logic | 04 January 2022, 23:44:32 UTC |
29bb14f | Mark Taylor | 03 January 2022, 00:28:40 UTC | update unit tests for interface changes | 03 January 2022, 00:28:40 UTC |
745a96e | Mark Taylor | 02 January 2022, 23:48:57 UTC | check for ncl in runall script | 02 January 2022, 23:48:57 UTC |
3053da2 | Mark Taylor | 02 January 2022, 23:42:55 UTC | remove hypervis_power (more) | 02 January 2022, 23:42:55 UTC |
de9d2d0 | Mark Taylor | 02 January 2022, 19:17:46 UTC | add vert_remap_u_alg option | 02 January 2022, 23:33:15 UTC |
eba902c | Mark Taylor | 29 December 2021, 06:08:08 UTC | update test case to test q_alg=11 | 02 January 2022, 19:17:09 UTC |
4c6b9f9 | Mark Taylor | 28 December 2021, 00:38:58 UTC | update comments | 02 January 2022, 19:15:12 UTC |
790bc2c | Mark Taylor | 26 December 2021, 23:40:57 UTC | add remap alg option to SL code | 26 December 2021, 23:54:25 UTC |
9a8d580 | Mark Taylor | 21 February 2021, 16:22:59 UTC | reorg vertical remap code and add new options vert_remap_q_alg=0 unchanged vert_remap_q_alg=1 unchanged vert_remap_q_alg=2 removed vert_remap_q_alg=3 removed (it was identical to 1) vert_remap_q_alg=10 unchanged vert_remap_q_alg=11 new: q_alg=10 with out limiters, for research [BFB] | 26 December 2021, 23:35:10 UTC |
4849b2e | Mark Taylor | 24 December 2021, 23:37:06 UTC | typo in python parsing script | 25 December 2021, 14:23:57 UTC |
3ed833c | Mark Taylor | 24 December 2021, 23:36:46 UTC | fix wrong dimensions | 25 December 2021, 14:23:56 UTC |
f56c25d | Mark Taylor | 24 December 2021, 22:50:48 UTC | remove extraneous lines from some CMakeLists.txt files These cmake files define macros that are used by other executables. The should only contain the setup macro and a call to createTestExec() Any other commands wont be run by the other executables | 25 December 2021, 14:23:56 UTC |
d680cb8 | Mark Taylor | 24 December 2021, 22:49:45 UTC | add comments about EOS needing to return dpnh_i=1 at surface | 25 December 2021, 14:23:56 UTC |
e8d6a19 | Mark Taylor | 24 December 2021, 22:49:13 UTC | remove obsolete HOMME_IMEX_MKLSOLVE code | 25 December 2021, 14:23:56 UTC |
3db07e8 | Mark Taylor | 20 April 2021, 16:34:15 UTC | dont run a state check when scale3=0 | 25 December 2021, 14:23:56 UTC |
7a1ab5f | Mark Taylor | 20 April 2021, 16:29:40 UTC | remove unused qn0 argument | 25 December 2021, 14:23:56 UTC |
07546bd | Mark Taylor | 20 April 2021, 14:02:56 UTC | remove hypervis_power restore correct var_coef behavoir remove hypervis power2 | 25 December 2021, 14:23:48 UTC |
c7a8708 | Mark Taylor | 20 April 2021, 13:45:28 UTC | dont try to turn on horiz/vert threads if no compiler support remove unecessary if statement in nu_div/nu logic remove unused namelist variable | 24 December 2021, 21:47:02 UTC |
679cebf | Oksana Guba | 23 December 2021, 18:39:55 UTC | Merge branch 'origin/oksanaguba/homme/xx-ttype9' into master (PR #4682) Adding ttype9 and dcmip2012 tests 2.1, 2.2, and 3 to the bfb suite. Converting some of the tests to ttype9. Not yet added is a limiter call between ttype9 stages. There will be another PR to move limiter from CaarFunctorImpl and to verify it is bfb with F code. [nonbfb] for homme bfb suite for theta-fhs3 (now with ttype9) and new baselines for theta-fdc tests. | 23 December 2021, 18:39:55 UTC |
7e954ba | Robert Jacob | 23 December 2021, 18:36:19 UTC | Merge branch 'wlin/atm/update_ftest_ena_twp_rrm' (PR #4655) F2010 is being used in extra coverage test suite for enax4v1 and twpx4v1, and the tests failed to proceed. Change to use compset F2010-CICE as no mpas mesh has been configured to work with these two atm RRM meshes. [BFB] | 23 December 2021, 18:36:19 UTC |
80a67f0 | Jon Wolfe | 23 December 2021, 15:59:37 UTC | Merge branch 'mark-petersen/lanl-scratch3' (PR #4716) Update LANL scratch3 to scratch4 LANL /lustre/scratch3 will be decommissioned January 4, 2022. All data should now be written to /lustre/scratch4. This updates the machine files for LANL grizzly and badger. [BFB] | 23 December 2021, 15:59:37 UTC |
a3b22b9 | mark-petersen | 22 December 2021, 22:29:30 UTC | Update LANL scratch3 to scratch4 LANL /lustre/scratch3 will be decommisioned January 4, 2022. All data should now be written to /lustre/scratch4. | 22 December 2021, 22:29:30 UTC |
acc9bbe | James Foucar | 22 December 2021, 01:09:06 UTC | Merge branch 'jgfouca/use_lmod_modules_for_mappy' into master (PR #4714) Use lmod modules for mappy [BFB] | 22 December 2021, 01:09:06 UTC |
5fd19d6 | James Foucar | 22 December 2021, 01:05:17 UTC | Use lmod modules for mappy | 22 December 2021, 01:05:17 UTC |
6528b78 | Azamat Mametjanov | 21 December 2021, 21:52:10 UTC | Merge remote-tracking branch 'mt5555/slurm-bindings' (PR #4712) Update some srun settings for HOMME tests Recent upgrades on Anvil/Chrysalis require setting the -c option to avoid poor performance when running with threads. Now that scripts set -c, remove hardcoded -c options on Cori. Also - fixed CPRNC path for Anvil/Chrysalis (they were swapped) - updated CPRNC path on Cori (it was using an ancient version that couldn't parse newer output files) - fixed CMAKE to pass MPI options to the limiter_ut test. srun hangs on Compy without --mpi=pmi2 On compy, HOMME machine file compy.cmake changed to a link, linked to compy-intel.cmake [BFB] | 21 December 2021, 21:52:10 UTC |
46b5c6b | Wuyin Lin | 21 December 2021, 15:03:54 UTC | Merge branch 'wlin/atm/diags_tropopause_cpp' (PR #4713) New algorithm for determining tropopause and additional diagnostic output Add a new algorithm for determining tropopause: cold point parabolic. Output at additional levels in the stratosphere are also added. While the new algorithm is enabled during runtime, no additional variables are saved by default. [BFB] | 21 December 2021, 15:03:54 UTC |
9ffdb63 | Wuyin Lin | 21 December 2021, 00:12:04 UTC | New algorithm for determining tropopause and additional diagnostic output Add a new algorithm for determining tropopause: cold point parabolic. Output at additional levels in the stratosphere are also added. While the new algorithm is enabled during runtime, no additional variables are saved by default. [BFB] | 21 December 2021, 00:12:04 UTC |
fa890d1 | Mark Taylor | 19 December 2021, 20:23:18 UTC | tweaks to pass MPI arguments to unit test | 19 December 2021, 20:23:18 UTC |
1d7b57c | Balwinder Singh | 19 December 2021, 15:21:59 UTC | Merge branch 'singhbalwinder/machines/sooty-mach-update' (PR #4668) Updates PNNL- Sooty's machine files to use PIO2 and other module changes Some recent machine files updates by Phil Rasch to run E3SM v2 on PNNL cluster Sooty. * singhbalwinder/machines/sooty-mach-update: Updates machine files to use PIO2 and other module changes | 19 December 2021, 15:21:59 UTC |
89ef6be | Mark Taylor | 18 December 2021, 20:34:11 UTC | thread bindings for knl and haswell, update cprnc path | 18 December 2021, 20:34:11 UTC |
004d15e | noel | 17 December 2021, 19:52:52 UTC | Merge branch E3SM-Project/ndk/machinefiles/correct-small-pelayout-knl-ne120pg2 (PR #4705) After recent PR, correct PE layout for "S" size of ne102pg2 F cases using CICE | 17 December 2021, 19:52:52 UTC |
b5525f3 | noel | 17 December 2021, 17:46:39 UTC | Merge branch E3SM-Project/ndk/machinefiles/perlmutter-env-update (PR #4710) Update module environment for Perlmutter | 17 December 2021, 17:46:39 UTC |
a445407 | Mark Taylor | 17 December 2021, 17:31:44 UTC | anvil/chrysalis threading and cprnc fix fixes to standalone HOMME: correct cprnc paths for anvil, chrysalis add cores-per-task srun option (otherwise HOMME threading tests have terrible performance) | 17 December 2021, 17:31:44 UTC |
60ec507 | Jon Wolfe | 16 December 2021, 19:51:01 UTC | Merge branch 'akturner/seaice/kstrength_constant_bug' (PR #4702) Removed Cf from sea ice constants The sea ice parameter Cf (config_ratio_ridging_work_to_PE) was incorrectly set in both constants and shared parameters in the column package, so that changing config_ratio_ridging_work_to_PE had no effect. Removed Cf from constants. Fixes #4701 [BFB] (Unless config_ratio_ridging_work_to_PE changed) | 16 December 2021, 19:51:01 UTC |
d02413b | jayeshkrishna | 16 December 2021, 17:47:46 UTC | Merge branch 'dqwu/anlgce/downgrade-netcdf-lib' (PR #4706) Some E3SM cases run on the ANL GCE nodes with the netcdf4p type and NetCDF 4.8.1 might return NC_EHDFERR from nc_enddef(). So switching back to NetCDF 4.8.0, an older but more stable release. For consistency on GCE nodes, we downgrade the serial version of the NetCDF lib to 4.8.0 as well. [BFB] | 16 December 2021, 17:47:46 UTC |
4af5914 | wadeburgess | 16 December 2021, 16:10:07 UTC | Merge pull request #4708 from E3SM-Project/wadeburgess-patch-1 Edit sandiatoss3 to change 'gpfs1' to 'gpfs' | 16 December 2021, 16:10:07 UTC |
e778eaf | wadeburgess | 16 December 2021, 15:20:18 UTC | Edit sandiatoss3 to change 'gpfs1' to 'gpfs' | 16 December 2021, 15:20:18 UTC |
cc97cda | dqwu | 15 December 2021, 21:41:45 UTC | Downgrading NetCDF lib from 4.8.1 to 4.8.0 for ANL GCE nodes It has been confirmed that some E3SM cases run on ANL GCE nodes with netcdf4p type and latest NetCDF 4.8.1 might return NC_EHDFERR from nc_enddef(). An issue has been created for NetCDF developers to fix. Before this issue is fixed in newer releases of NetCDF, we have to switch back to NetCDF 4.8.0, an older but more stable release. For consistency on GCE nodes, we downgrade the serial version of the NetCDF lib to 4.8.0 as well. | 15 December 2021, 21:41:45 UTC |
c8d23d2 | jayeshkrishna | 15 December 2021, 21:20:03 UTC | Merge branch 'dqwu/anlgce/add-mpilib-mpi-serial' (PR #4665) Define NETCDF_PATH for E3SM to use a serial NetCDF lib installed on anlgce when mpilib is set to mpi-serial. This fix is required to run SMS_D_Ln9_Mmpi-serial.f19_g16_rx1.A, a cime_developer test that uses mpi-serial. [BFB] | 15 December 2021, 21:20:03 UTC |
2e14457 | noel | 15 December 2021, 18:37:45 UTC | After recent PR, correct PE layout for "S" size of ne102pg2 F cases using CICE [bfb] | 15 December 2021, 18:37:45 UTC |
de04350 | noel | 15 December 2021, 18:29:02 UTC | Merge branch E3SM-Project/ndk/machinefiles/cori-knl-pe-layouts-ne120pg2 (PR #4696) Update PE layouts on cori-knl for ne120pg2 F compsets | 15 December 2021, 18:29:02 UTC |
33b4841 | Jon Wolfe | 15 December 2021, 16:17:28 UTC | Merge branch 'njeffery/seaice/correct_don_coupling' (PR #4699) Corrects coupling of ocean surface DON with sea ice DON Replaces index_x2i_So_don1 with index_x2i_So_doc3 in ice_comp_mct. Fixes issue #4698 [non-BFB] only with ocean and sea ice BGC active | 15 December 2021, 16:17:28 UTC |
30a7f2b | Jon Wolfe | 15 December 2021, 16:12:25 UTC | Merge branch 'akturner/seaice/testcase_grid_ccw_error' (PR #4691) Updated seaice spherical test cases for current MPAS mesh spec Update sea ice test cases to use spherical unit meshes that comply with current MPAS mesh spec. Also update PWL variational method to use the current MPAS mesh spec Fixes #4690 [BFB] | 15 December 2021, 16:12:25 UTC |
35076fe | Adrian K. Turner | 14 December 2021, 00:02:29 UTC | Removed Cf from sea ice constants | 14 December 2021, 00:02:29 UTC |
d64cdc1 | Peter Schwartz | 13 December 2021, 18:20:19 UTC | Merge remote-tracking branch 'pnlfangfork/pnlfang/fix_restart_bfb_error'(PR #4688) This bug fix is related to the restart bfb error when plant hydraulics is activated. [BFB] | 13 December 2021, 18:20:19 UTC |
db532d1 | noel | 13 December 2021, 17:17:49 UTC | whitespace cleanup | 13 December 2021, 17:17:49 UTC |
98af778 | Nicole Jeffery | 13 December 2021, 17:12:43 UTC | Corrects coupling of ocean surface DON with sea ice DON. Replaces index_x2i_So_don1 with index_x2i_So_doc3 in ice_comp_mct. non-BFB with ocean and sea ice BGC active. BFB otherwise. Fixes issue #4698 | 13 December 2021, 17:12:43 UTC |
0c857cc | Alice Barthel | 13 December 2021, 16:51:29 UTC | 1. corrected last layer thickness to partial, 2. fetch index_temp as best practice | 13 December 2021, 16:51:29 UTC |
c40a28c | noel | 11 December 2021, 00:40:48 UTC | Add allow_error="true" to try getting around module "error" messages. | 11 December 2021, 00:40:48 UTC |
379fbcc | noel | 10 December 2021, 23:10:59 UTC | For Perlmutter, Update modules for cuda after maint. | 10 December 2021, 23:10:59 UTC |
170450b | noel | 10 December 2021, 21:46:21 UTC | Update PE layouts on cori-knl for ne120pg2 F compsets Add S,M,L,X,T for MPASSI and update S,M,L,X,T for CICE [bfb] | 10 December 2021, 21:46:21 UTC |
89d4726 | Jon Wolfe | 08 December 2021, 17:16:49 UTC | Merge branch 'xylar/ocean/fix-two-analysis-member-restart-filename-intervals' (PR #4687) Fix filename interval for 2 analysis member restarts The Eliassen-Palm and time-filters analysis members both had hard-coded restart file intervals of 1 year. This causes restarts with these analysis members to fail in situations where the restart interval does not match the file interval. This merge changes the filename interval to be the same as the output interval to match restart streams in all other analysis members. [BFB] | 08 December 2021, 17:16:49 UTC |
9233213 | Adrian K. Turner | 08 December 2021, 15:14:57 UTC | Updated seaice spherical test cases for current MPAS mesh spec | 08 December 2021, 15:14:57 UTC |
c2cd51a | Gautam Bisht | 07 December 2021, 23:22:27 UTC | Merge branch 'singhbalwinder/elm/ozone-dry-dep-bug-fix' (PR #4678) Currently, stomatal resistances for shaded and sunlit leaves are added in series but they should be added in parallel Fixes #4677 [BFB] | 07 December 2021, 23:22:27 UTC |
37260f0 | Gautam Bisht | 07 December 2021, 23:20:52 UTC | Merge branch 'donghuix/mosart/unstructured-grid' (PR #4661) Fixes a bug in MOSART for reading the latitude from an unstructured input file. [non-BFB] | 07 December 2021, 23:20:52 UTC |
5af86a4 | Oksana Guba | 07 December 2021, 19:46:05 UTC | fix random fail | 07 December 2021, 19:46:05 UTC |
2a3b779 | Oksana Guba | 07 December 2021, 18:24:20 UTC | Merge 'origin/bartgok/hommexx-fixes' into next (PR #4681) Two mods: -- fix driver mod for theta-l_kokkos, to compile test forcing stuff only in standalone mode (fixes link error). -- change Kokkos.cmake script, so that it does not add assumed include/link folders/libs when Kokkos is built externally (fixes hundreds of non-existent include dir in SCREAM). [BFB] | 07 December 2021, 18:24:20 UTC |
1b1f762 | Oksana Guba | 07 December 2021, 16:39:34 UTC | avoid crash when on gpu | 07 December 2021, 16:39:34 UTC |
588dcea | Oksana Guba | 07 December 2021, 16:20:01 UTC | fixing random fails | 07 December 2021, 16:20:01 UTC |
c764017 | Jon Wolfe | 07 December 2021, 15:32:07 UTC | Merge branch 'brian-oneill/mpas-ocean/openacc-vmix' (PR #4680) GPU port of ocean vmix routines OpenACC added to vmix velocity and tracer routines. Routines for vmix coefs untouched. Changes include: * replaced mesh pool with ocn_mesh module * minimized copies of tridiag coefs in solution loops * inlined tridiag and kpp_rhs routines * added OpenACC directives for loop kernels and data updates [BFB] | 07 December 2021, 15:32:07 UTC |
99b39fc | Oksana Guba | 07 December 2021, 03:37:25 UTC | implement AMB suggestions | 07 December 2021, 03:37:25 UTC |
def827f | Oksana Guba | 07 December 2021, 02:29:09 UTC | redo namelists | 07 December 2021, 02:29:09 UTC |
6719607 | Oksana Guba | 07 December 2021, 00:02:22 UTC | delete old nl | 07 December 2021, 00:02:22 UTC |
38243bf | Yilin Fang | 04 December 2021, 04:36:08 UTC | Correction to resolve restart bfb error. | 04 December 2021, 04:36:08 UTC |
524443b | Oksana Guba | 03 December 2021, 21:37:30 UTC | make bfb for test22 | 03 December 2021, 21:37:30 UTC |
02e25cf | Oksana Guba | 03 December 2021, 20:41:05 UTC | ttype9 in xx, bfb for dcmip12-test3 | 03 December 2021, 20:41:05 UTC |
89b5192 | Xylar Asay-Davis | 03 December 2021, 08:10:35 UTC | Fix filename interval for 2 analysis member restarts The Eliassen-Palm and time-filters analysis members both had hard-coded restart file intervals of 1 year. This causes restarts with these analysis members to fail in situations where the restart interval does not match the file interval. This merge changes the filename interval to be the same as the output interval to match restart streams in all other analysis members. | 03 December 2021, 08:18:55 UTC |
bc5a2cc | Gautam Bisht | 02 December 2021, 21:41:11 UTC | Merge branch 'tizhou/water_feature_test' (PR #4676) A new test is added to e3sm_land_developer testing suite with the following land features turned on: two-way irrigation, lake storage, bgc mode, and transient mode. It also turns on two river features: inundation and water management. These features were developed for V2 BGC simulations. [BFB] | 02 December 2021, 21:41:11 UTC |
91d66b5 | Oksana Guba | 02 December 2021, 19:17:53 UTC | Merge branch 'AMB/ambrad/homme/fix-gllfvremap-ftype0' into next (PR #4636) ftype 0 is not used in practice in EAM, but it is supported. @bartgol pointed out to me that dp_coupling and gfr_fv_phys_to_dyn do not agree on the inputs and outputs for q and FQ when ftype is 0. This PR fixes that. It also adds two CIME tests for ftype 0 and pg2. Fixes #4632. [BFB] except for two new CIME tests (no baseline) and Homme standalone test thetah-sl-dcmip16_test1pg2 (diff) | 02 December 2021, 19:17:53 UTC |
363eed3 | Jon Wolfe | 02 December 2021, 17:38:45 UTC | Merge branch 'njeffery/cpl/ice_marbl_coupling' (PR #4641) Corrected coupling of ice biogeochemistry for MARBL Updates seaice-ocean coupling for BGC: * Adds coupling of refractory DOC with sea ice humics * Adds option to couple sea ice DOC with the macromolecules or the MARBL DOC pool * Corrects coupling of DON to conserve carbon New carbon conservation analysis members are still needed in the ocean and updates in the sea ice before we can confirm that the coupling conserves carbon. Fixes #4637 [NML] [non-BFB] only with the ocean/ice BGC components active | 02 December 2021, 17:38:45 UTC |
69cd15a | Oksana Guba | 01 December 2021, 20:58:20 UTC | more tests, bfb | 01 December 2021, 20:58:20 UTC |
b0c33b0 | Oksana Guba | 01 December 2021, 02:16:41 UTC | formatting | 01 December 2021, 02:16:41 UTC |
7543b61 | Oksana Guba | 01 December 2021, 02:14:48 UTC | Merge branch 'origin/master' into oksanaguba/homme/xx-ttype9 | 01 December 2021, 02:14:48 UTC |
eef5664 | Oksana Guba | 01 December 2021, 02:07:58 UTC | fix ut tests, compose. all is bfb, ut tests pass | 01 December 2021, 02:07:58 UTC |
a02a913 | Oksana Guba | 01 December 2021, 00:41:02 UTC | Merge 'AMB/ambrad/homme/fix-some-compose-warnings' into next (PR #4622) Cuda 10.1.105 with gcc 8.5.0 erroneously reports an error for a certain line. Details are in SCREAM issue 1277. This PR works around the issue. I studied the issue in part using -Wall -pedantic with various compiler versions and thus have also fixed some miscellaneous warnings in SL and physgrid remap code. [BFB] | 01 December 2021, 00:41:02 UTC |
f88fa0a | Oksana Guba | 30 November 2021, 23:23:25 UTC | changes to more rearth out of physicsconstants: move rearth to simulationParameters, which will add it to geometry, and then to sphereOperators. had to touch both preqxx and thetaxx. change for Compose is temporary, not sure about proper change. preqxx and theta tests (incl. dcmip2012 test 2.1) are bfb. unit tests (builds) are broken. | 30 November 2021, 23:23:25 UTC |
5fce6ec | Benjamin R. Hillman | 30 November 2021, 19:51:41 UTC | Merge branch 'whannah/mmf/refactor-esmt-pgf' (#4664) Refactor the scalar momentum transport parameterization (ESMT) for the C++ CRM code. Previously, the pressure gradient force calculation was done on the CPU for GPU runs, which greatly reduced the throughput. This PR fixes that and maximizes the parallelism for better performance. [BFB] except for MMF+ESMT test | 30 November 2021, 19:51:41 UTC |
d87b55e | Luca Bertagna | 30 November 2021, 17:51:03 UTC | Homme: in theta-l_kokkos, compile forcing stuff only in standalone mode | 30 November 2021, 17:51:05 UTC |
1c76f58 | Luca Bertagna | 19 November 2021, 00:37:33 UTC | Homme: improved Kokkos.cmake script The improvement allows to use this script, avoid building kokkos, and avoid assuming location of the include/lib paths. Before this mods, the script automatically addes some include paths that were not necessarily the correct ones. | 30 November 2021, 17:48:29 UTC |
cb98833 | Oksana Guba | 30 November 2021, 00:37:19 UTC | remove phys const rearth from sp ops | 30 November 2021, 00:37:19 UTC |
a5cb66a | Brian O'Neill | 29 November 2021, 23:40:27 UTC | remove deleted subroutines from module member function declarations declarations for nonexistant member functions appears to have caused ci build error | 29 November 2021, 23:40:27 UTC |
f49c36a | Brian O'Neill | 29 November 2021, 21:26:28 UTC | GPU port of ocean vmix routines - replaced mesh pool with ocn_mesh module - minimized copies of tridiag coefs in solution loops - inlined tridiag and kpp_rhs routines - added OpenACC directives for loop kernals and data updates | 29 November 2021, 21:26:28 UTC |
2b0394b | James Foucar | 29 November 2021, 17:57:48 UTC | Merge branch 'jgfouca/enhance_macro_linter' (PR #4651) Develop a tool to clean up CMake macro files and apply it to current files. Clean up a lot of problems. [BFB] | 29 November 2021, 17:57:48 UTC |
d92d90d | hydrotian | 25 November 2021, 03:51:50 UTC | daily MOSART output | 25 November 2021, 03:51:50 UTC |
62749be | Balwinder Singh | 24 November 2021, 23:23:58 UTC | Fixes a bug in O3 dry dep following Emmons et al. 2020 Currently, stomatal resistances for shaded and sunlit leaves are addd in series but they should be added in parallel Fixes #4677 [BFB] | 24 November 2021, 23:28:12 UTC |
d06463b | noel | 23 November 2021, 23:54:29 UTC | Merge branch E3SM-Project/ndk/machinefiles/add-compiler-flag-for-one-mpas-source-with-gnu (PR #4672) For gnu compiler version 9 and above, adjust compiler flag for one mpas source | 23 November 2021, 23:54:29 UTC |
34db72d | Nicole Jeffery | 23 November 2021, 18:20:06 UTC | changed variable avgOceanSurfaceDOCSum to avgOceanSurfaceDOCSemiLabile Verified with a short 10 day GCASE run. | 23 November 2021, 18:20:06 UTC |
4f29ad3 | Jon Wolfe | 23 November 2021, 17:49:58 UTC | Merge branch 'jonbob/scripts/fix-erp-test' (PR #4673) Change ERP_Ld3.ne4_oQU240.F2010 test to ERP_Ln18.ne4_oQU240.F2010 The ERP_Ld3.ne4_oQU240.F2010 test in the suite currently fails on mappy due to the memory leak test and is apparently related to the writing of restart files and not an actual memory leak. Changing the test to run 18 timesteps instead of 3 days avoids the restart issue and allows the test to pass. Fixes #4546 [BFB] | 23 November 2021, 17:49:58 UTC |
488ac4c | Jon Wolfe | 23 November 2021, 17:42:40 UTC | Merge branch 'mark-petersen/mpas/update-Makefile-gnu' (PR #4643) Add -ffpe-summary=none to stand-alone MPAS Makefile for gnu Currently, gnu-compiled executables produce a summary table of exceptions at the end and prints one line per processor on finalize for both debug and optimized runs. When running on hundreds of cores, this is a headache. Other modeling groups see this with gnu 5+ as well, and say to ignore it or add -ffpe-summary=none. See gnu compile flags. This affects MPAS-seaice and MALI as well. This Makefile is not used for E3SM compiling. [BFB] | 23 November 2021, 17:42:40 UTC |
2330a08 | James Foucar | 23 November 2021, 17:20:06 UTC | Now that the LINUX macro is actually being loaded it appears not to work, so remove | 23 November 2021, 17:20:06 UTC |