| 44d5343 | Robert Jacob | 15 September 2021, 21:03:46 UTC | Merge branch 'rljacob/fix-extra-read' (PR #4523) Make drv_resume a logical and set it to FALSE to prevent read of restart immediately after it is written. Introduce drv_resume_file to use where a filename is really needed. Also add a log message for the real restart read and make the seq_io_read_avs log messages more informative. Fixes #4513 | 15 September 2021, 21:03:46 UTC |
| 5341618 | Jon Wolfe | 15 September 2021, 14:55:46 UTC | Merge branch 'mark-petersen/ocn/fix-diagnostics-landIceFraction' (PR #4518) Add land ice logical to avoid allocation errors Use of land ice variables, which are in packages and may not be allocated, were not protected by if statements. Fixes #4517 [BFB] | 15 September 2021, 14:55:46 UTC |
| 24bb118 | Robert Jacob | 14 September 2021, 23:32:12 UTC | Merge branch 'rljacob/circleci/fix-compset' (PR #4539) Remove the "A_" from the compset used in the build test to match new v2 standards. [BFB] | 14 September 2021, 23:32:12 UTC |
| 6e163cf | Robert Jacob | 14 September 2021, 23:00:10 UTC | Change compset name to build Remove the "A_" from the compset name to match new v2 standards. | 14 September 2021, 23:00:10 UTC |
| c85d9f7 | Wuyin Lin | 14 September 2021, 12:29:47 UTC | Merge branch 'wlin/atm/fix_conusx4v1_f2010' (PR #4535) Reset grid and compset for F2010 tests with RRM The existing bi-grid config for RRM test cannot work with MPASSI. For conusx4v1, continue to use F2010 but changed to use tri-grid conusx4v1_r05_oECv3. For enax4v1 and twpx4v1, continue to use bi-grid but changed compset to F2010-CICE. The current arrangement is based on availability of grid for using MPASSI and the inputdata. [BFB] except for new test with conusx4v1_r05_oECv3 | 14 September 2021, 12:29:47 UTC |
| 4e12683 | Wuyin Lin | 14 September 2021, 00:32:21 UTC | Reset grid and compset for F2010 tests with RRM Continue to use F2010 for conusx4v1 but changed to use tri-grid conusx4v1_r05_oECv3. For enax4v1 and twpx4v1, continue to use bi-grid but changed compset to F2010-CICE. The current arrangement is based on availability of grid for using MPASSI and the inputdata. [BFB] except for new test with conusx4v1_r05_oECv3 | 14 September 2021, 00:32:21 UTC |
| 1a02264 | Wuyin Lin | 13 September 2021, 16:41:43 UTC | Merge branch 'wlin/atm/cleanup_v2compsets_for_tests' (PR #4488) Update and clean-up compsets for use in test suites Remove obsolete eam and allactive compsets and the associated eam use_case files, and update the compsets used in nightly test suites. All compsets referring to AV1C or v0atm are removed. F20TRC5-CMIP6 and F2010SC5-CMIP6 are no more needed (previously retained for ongoing testings). Remaining use case files referring to av1c are removed, including those used by MMF tests. Also Remove obsolete use_case files referring to cam3 and cam4 . This PR also makes all standard F-cases use prescribed data ice mode of MPASSI. This includes F1850, F1950, F2010, F20TR. The corresponding compsets that use CICE are renamed as F1850-CICE, F1950-CICE, F2010-CICE, and F20TR-CICE, respecticely. The F####-CICE compsets set to go with SROF while the MPASSI counterparts with MOSART as river component. The following compsets are replaced in nightly test suites. . FC5AV1C-L with F2010 . A_WCYCL1850 and A_WCYCL1850S_CMIP6 with WCYCL1850 . FC5AV1C-H01A with F2010 as there is no v2 compset just for hires . A_WCYCL1950S_CMIP6_HR and A_WCYCL2000_H01AS with WCYCL1950 as a placeholder With compset changes and the use of MPASSI data ice mode, F2010 tests using ne4_ne4 change to use ne4_oQU240, and ne4pg2_ne4pg2 to ne4pg2_oQU480; F2010 tests with ERS_Ln# and ERP_Ln# change to use ERS_Ld3. [non-BFB] Virtually for all tests involving the compsets above. Tests using the old compsets would appear missing while new tests would appear correspondingly. | 13 September 2021, 16:41:43 UTC |
| 7d6fa20 | Wuyin Lin | 12 September 2021, 02:21:26 UTC | Replaced F2010 ERP_Ln# tests with ERP_Ld3 | 12 September 2021, 02:21:26 UTC |
| c06d4ce | Benjamin R. Hillman | 10 September 2021, 19:33:30 UTC | Remove cam4 use cases files | 10 September 2021, 20:52:33 UTC |
| 66318c3 | Benjamin R. Hillman | 10 September 2021, 19:12:13 UTC | Remove some unused use cases | 10 September 2021, 19:12:13 UTC |
| 6dba9f4 | Benjamin R. Hillman | 10 September 2021, 18:57:05 UTC | Update MMF F-cases to F2010 | 10 September 2021, 18:57:05 UTC |
| 0f5c969 | Benjamin R. Hillman | 10 September 2021, 18:56:38 UTC | Update MMF use case files for 2010 forcings | 10 September 2021, 18:56:38 UTC |
| 31c32e4 | Robert Jacob | 10 September 2021, 03:04:11 UTC | Fix whitespace/alignment Fix some whitespace/alignment on a couple of lines. | 10 September 2021, 03:04:11 UTC |
| 2ee5845 | Benjamin R. Hillman | 09 September 2021, 18:31:34 UTC | Merge branch 'whannah/mmf/push-chunk-loop-crm-physics' (#4476) This refactors the way the CRM is called so that all CRM instances in a single task can be run at once on the GPU, rather than all CRMs in a given "chunk". On properly configured GPU cases the speedup from this change is significant, and there is no notable change to CPU configurations tested on Summit and Cori-KNL. The changes required to push the chunk loop down are quite substantial, but there were additional complications from the crm_history() subroutine and the crm_rad and crm_ecpp_output data structures that required more extensive changes. A new "ncrms" variable is defined in crm_physics to define the dimension of variables that hold all CRMs in a task (i.e. combined across all chunks in a task). Note about testing - the changes are non-BFB for MMF tests only because the grouping of CRM columns has changed, which changes the the time step and sedimentation subcycling. Otherwise the changes are BFB, which can be demonstrated using the mmf_fixed_subcycle test mod. [BFB] - except for MMF tests without mmf_fixed_subcycle test mod | 09 September 2021, 18:31:34 UTC |
| f2f1eba | Robert Jacob | 09 September 2021, 17:52:25 UTC | Merge branch lukaszlacinski/lukaszlacinski/circleci (PR #4528) This PR removes a bug that was introduced with the last PR that added a backup HTTP server with the Singularity source code tarball. | 09 September 2021, 17:52:25 UTC |
| 9a5dab5 | James Foucar | 09 September 2021, 17:08:37 UTC | Merge branch 'jgfouca/fix_cimeroots' into master (PR #4526) Fix CIMEROOTs for buildnml scripts It looks like these did not get updated when the scripts moved to E3SM. Also, remove old python module load from mappy env. [BFB] * jgfouca/fix_cimeroots: Undo cime subm update Fix some stale CMake paths Fix CIMEROOTs for buildnml scripts | 09 September 2021, 17:08:37 UTC |
| e0d6219 | Wuyin Lin | 09 September 2021, 16:17:42 UTC | Removed gust_export testmod dir | 09 September 2021, 16:17:42 UTC |
| a120bae | Wuyin Lin | 09 September 2021, 16:12:58 UTC | Fixed F2010 ERS tests on ne4 and ne4pg2 with _Ln# replaced with _Ld3 | 09 September 2021, 16:12:58 UTC |
| 52167a2 | lukaszlacinski | 09 September 2021, 16:11:05 UTC | Fix the CircleCI script that installs Singularity | 09 September 2021, 16:11:05 UTC |
| f47bb19 | Jon Wolfe | 09 September 2021, 15:49:38 UTC | Merge branch 'proteanplanet/seaice/v2cryoalignment' (PR #4524) Align sea ice settings of SOwISC12to60E2 and WCAtl12to45E2 with WC14to60 This aligns the E3SM V2 Cryospheric campaign regionally-refined Southern Ocean mesh with the V2 Water Cycle WC14 mesh settings for sea ice. Thus the Antarctic regionally-refined settings in E3SM would become consistent with the Arctic settings, and vice versa. The two changes made are that the number of iterations in the atmospheric flux calculations use 10 not 5 iterations, and high-frequency coupling is switched on, which means that a transient Ekman solution is permitted in the sea ice model. [BFB] for all tested configurations | 09 September 2021, 15:49:38 UTC |
| bc45dc5 | Robert Jacob | 09 September 2021, 15:38:20 UTC | Merge branch 'lukaszlacinski/circleci' (PR #4508) Add the NERSC HTTP as a primary server with the Singularity source code tarball and use the ALCF Petrel HTTP server as a backup server. If the CircleCI test script fails when downloading the tarball from the server, the script falls back to the ALCF Petrel HTTP server. | 09 September 2021, 15:38:20 UTC |
| a7c0047 | Jon Wolfe | 09 September 2021, 15:40:46 UTC | Merge branch 'njeffery/mpas-seaice/zero-aerosol-coupling' (PR #4515) Zeroed out dummy ocean concentrations of aerosols This is a fix to issue #4514 : Sea ice receives negative ocean concentrations of aerosols when zaerosols are active in MPAS-seaice. Without this fix, the code crashes with a negative concentration error in the sea ice when zaerosols are active. Fixes #4514 [BFB] in all tested configurations | 09 September 2021, 15:40:46 UTC |
| d0d2eb4 | Wuyin Lin | 09 September 2021, 00:38:11 UTC | Replaced ne4_ne4 with ne4_oQU240 and ne4pg2_ne4pg2 with ne4pg2_OQU480 for F2010 tests. Added compset WCYCL1850NS for testing on small grids without spun-up ICs for ocean/ice. Removed eam-gust_param and eam-gust_export that are not needed for EAMv2. | 09 September 2021, 00:38:11 UTC |
| e1aa52b | James Foucar | 08 September 2021, 23:01:35 UTC | Undo cime subm update | 08 September 2021, 23:01:35 UTC |
| 5dcf1b6 | James Foucar | 08 September 2021, 22:57:53 UTC | Fix some stale CMake paths | 08 September 2021, 22:57:53 UTC |
| 7cc7a61 | Robert Jacob | 08 September 2021, 21:32:52 UTC | Fix comments and spacing Fix comments and spacing in response to review. | 08 September 2021, 21:32:52 UTC |
| 90fd936 | James Foucar | 08 September 2021, 19:39:49 UTC | Fix CIMEROOTs for buildnml scripts It looks like these did not get updated when the scripts moved to E3SM. Also, remove old python module load from mappy env. [BFB] | 08 September 2021, 19:39:49 UTC |
| 012a712 | Walter Hannah | 07 September 2021, 20:42:45 UTC | clean up crm_history modified: components/eam/src/physics/crm/crm_history.F90 | 08 September 2021, 16:36:27 UTC |
| 9bf83e2 | Walter Hannah | 07 September 2021, 20:42:09 UTC | revert changes to crm interface rad variables modified: components/eam/src/physics/crm/crm_physics.F90 modified: components/eam/src/physics/crm/sam/crm_module.F90 modified: components/eam/src/physics/crm/samomp/crm_module.F90 | 08 September 2021, 16:36:27 UTC |
| 68972f6 | Walter Hannah | 03 September 2021, 23:17:34 UTC | another bug fix | 08 September 2021, 16:36:27 UTC |
| 35392b0 | Walter Hannah | 03 September 2021, 23:05:05 UTC | bug fix | 08 September 2021, 16:36:27 UTC |
| 1924365 | Walter Hannah | 03 September 2021, 21:13:03 UTC | Add check for icol bounds | 08 September 2021, 16:36:27 UTC |
| b80ab9c | Walter Hannah | 03 September 2021, 20:59:48 UTC | be more careful about ncol vs pcol | 08 September 2021, 16:36:27 UTC |
| 5ea43a9 | Walter Hannah | 03 September 2021, 20:57:05 UTC | change icrm to icol in crm_history | 08 September 2021, 16:36:27 UTC |
| 4d28b8e | Walter Hannah | 03 September 2021, 20:54:30 UTC | Add trim to be safe | 08 September 2021, 16:36:27 UTC |
| d670d02 | Walter Hannah | 03 September 2021, 20:53:14 UTC | Cosmetic update from PR review | 08 September 2021, 16:36:27 UTC |
| 3c58ed1 | Walter Hannah | 31 August 2021, 15:46:29 UTC | Minor bug fixes and clean up modified: components/eam/src/physics/crm/crm_physics.F90 modified: components/eam/src/physics/crm/physpkg.F90 | 08 September 2021, 16:36:18 UTC |
| e46c12c | Walter Hannah | 29 July 2021, 22:01:50 UTC | switch outfld calls from pcols to ncol modified: components/eam/src/physics/crm/crm_history.F90 | 08 September 2021, 16:35:22 UTC |
| 9a191bf | Walter Hannah | 14 July 2021, 16:56:40 UTC | more bug fixes and clean up modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:35:22 UTC |
| 08fd70c | Walter Hannah | 14 July 2021, 16:56:21 UTC | Refactor crm_history modified: components/eam/src/physics/crm/crm_history.F90 | 08 September 2021, 16:35:22 UTC |
| 0b2ecd0 | Walter Hannah | 14 July 2021, 16:45:12 UTC | bug fix modified: components/eam/src/physics/crm/physpkg.F90 | 08 September 2021, 16:35:21 UTC |
| c556c11 | Walter Hannah | 12 July 2021, 23:12:53 UTC | rework handling of ECPP output modified: components/eam/src/physics/crm/crm_ecpp_output_module.F90 modified: components/eam/src/physics/crm/crm_physics.F90 modified: components/eam/src/physics/crm/physpkg.F90 | 08 September 2021, 16:34:50 UTC |
| 7f6f3ad | Walter Hannah | 12 July 2021, 22:13:58 UTC | bug fix | 08 September 2021, 16:24:35 UTC |
| 1eb27e5 | Walter Hannah | 12 July 2021, 22:09:44 UTC | update SAM/SAMOMP interface modified: components/eam/src/physics/crm/crm_physics.F90 modified: components/eam/src/physics/crm/sam/crm_module.F90 modified: components/eam/src/physics/crm/samomp/crm_module.F90 | 08 September 2021, 16:24:35 UTC |
| d0e6989 | Walter Hannah | 12 July 2021, 20:53:08 UTC | Refactor post-CRM-run section modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:35 UTC |
| 77aead4 | Walter Hannah | 12 July 2021, 19:40:53 UTC | clean up CRM run section | 08 September 2021, 16:24:35 UTC |
| b71ef38 | Walter Hannah | 12 July 2021, 19:33:04 UTC | refactor non-first-step init section modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:35 UTC |
| c098839 | Walter Hannah | 12 July 2021, 18:58:23 UTC | clean up some first step stuff modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:35 UTC |
| 2aacd96 | Walter Hannah | 12 July 2021, 18:41:07 UTC | refactor init and first_step sections of crm_physics_tend modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:35 UTC |
| debc67e | Walter Hannah | 12 July 2021, 17:29:10 UTC | update crm_physics_tend local variables modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:35 UTC |
| 8419c4d | Walter Hannah | 12 July 2021, 17:01:06 UTC | Clean up crm_physics_tend interface modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:34 UTC |
| e44051a | Walter Hannah | 12 July 2021, 16:46:40 UTC | white space change | 08 September 2021, 16:24:34 UTC |
| 8328332 | Walter Hannah | 12 July 2021, 16:43:10 UTC | single use statement for ppgrid modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:34 UTC |
| 0dce59f | Walter Hannah | 12 July 2021, 16:25:08 UTC | Apply new pbuf indices modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:34 UTC |
| ed8d790 | Walter Hannah | 12 July 2021, 16:23:23 UTC | minor change to CRM register and init routines modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:34 UTC |
| 8dc4944 | Walter Hannah | 12 July 2021, 16:12:33 UTC | Add global variance transport index variables modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:34 UTC |
| 421d83b | Walter Hannah | 12 July 2021, 16:07:41 UTC | refactor crm_physics_tend call in phys_run1 modified: components/eam/src/physics/crm/physpkg.F90 | 08 September 2021, 16:24:34 UTC |
| ade3d31 | Walter Hannah | 12 July 2021, 16:04:17 UTC | Add public ncrms variable modified: components/eam/src/physics/crm/crm_physics.F90 | 08 September 2021, 16:24:33 UTC |
| 4afd400 | Andrew Roberts | 08 September 2021, 16:23:59 UTC | This commit adds WCAtl12to45E2r4 to the regionally-refined MPAS-SI configurations | 08 September 2021, 16:23:59 UTC |
| 2cc4e6f | Andrew Roberts | 08 September 2021, 03:45:57 UTC | Align sea ice settings of SOwISC12to60E2r4 with WC14to60 | 08 September 2021, 03:45:57 UTC |
| 5cb4bce | Robert Jacob | 08 September 2021, 00:09:55 UTC | Add more info to log message for seq_io_read_avs When a restart file is read, include info about which avs are being read. Also fix the subname for seq_io_read_avscomp. | 08 September 2021, 00:09:55 UTC |
| 58c5fbf | Robert Jacob | 08 September 2021, 00:06:29 UTC | Separate drv_resume file and logical Make drv_resume a logical and use it to determine if a restart file should be read. Introduce drv_resume_file and use it where drv_resume was assumed to be a filename. Also add a log message for the real restart read at the start of a run. Fixes #4513 | 08 September 2021, 00:06:29 UTC |
| f5528f0 | Mark Petersen | 07 September 2021, 19:41:26 UTC | Change land ice logic flags from Off to On | 07 September 2021, 19:59:21 UTC |
| c0eec93 | James Foucar | 07 September 2021, 18:23:05 UTC | Merge branch 'jgfouca/cime_update_2021_09_03' into master (PR #4516) Update CIME submodule to e3274c2ffcd1ce5821499d33484d6b9384ee73b6 Changes: 1) Support the creation of multiple user_nl files for a component 2) Allow multiple methods for running CIME unit tests 3) handle neon input data better by first checking the neon server 4) changes to have new user_nl_stream_xxx for cdeps 5) moab driver in src/drivers/moab is no longer needed, so it's been removed. Fixes: 1) pbs doesnt allow some compset characters in job name 2) fix an issue with file locks when not run from caseroot 3) Fix namelist duplications 4) Fixes provenance in git worktree 5) make sure CLM_USRDAT_NAME is not UNSET if set [BFB] | 07 September 2021, 18:23:05 UTC |
| 081942d | Mark Petersen | 04 September 2021, 22:58:46 UTC | add land ice logical to avoid errors | 04 September 2021, 22:58:46 UTC |
| 596b667 | Azamat Mametjanov | 03 September 2021, 19:39:01 UTC | Merge branch 'azamat/ascent/update-builds' (PR #4473) Update builds on Ascent: * machine-specific updates to modules, compiler flags, Depends files * run ibmgpu builds with 42 tasks/node in SMT1 mode * run F-MMFOMP OpenMP-offload case without multithreading * add single-node pelayout for ne4np4.pg2 grid Addresses E3SM-Project/E3SM#4458 [BFB] | 03 September 2021, 19:39:01 UTC |
| ca59343 | James Foucar | 03 September 2021, 19:13:09 UTC | Update CIME submodule to e3274c2ffcd1ce5821499d33484d6b9384ee73b6 Changes: 1) Support the creation of multiple user_nl files for a component 2) Allow multiple methods for running CIME unit tests 3) handle neon input data better by first checking the neon server 4) changes to have new user_nl_stream_xxx for cdeps 5) moab driver in src/drivers/moab is no longer needed, so it's been removed. Fixes: 1) pbs doesnt allow some compset characters in job name 2) fix an issue with file locks when not run from caseroot 3) Fix namelist duplications 4) Fixes provenance in git worktree 5) make sure CLM_USRDAT_NAME is not UNSET if set [BFB] | 03 September 2021, 19:13:09 UTC |
| cf9781e | Jon Wolfe | 03 September 2021, 14:40:46 UTC | Merge branch 'jonbob/scripts/remove-deprecated-mpas-grids' (PR #4467) Remove deprecated MPAS grids from E3SM scripts Removes grids no longer supported by the MPAS ocn/ice components, either because they have been determined to be unnecessary or have been replaced. The deprecated grids are: * mpasgx1 * mpas120 * oEC60to30 * oEC60to30wLI * oRRS30to10 * oRRS30to10wLI * oRRS18to6 Please note we would like to drop the oRRS15to5 as well, but it is being used by SCREAM/high-res atm [BFB] for all current tests | 03 September 2021, 14:40:46 UTC |
| 3dc3f32 | Wuyin Lin | 03 September 2021, 03:49:00 UTC | Restored compset BGWCYCL1850 and replaced AV1C-L phys | 03 September 2021, 03:49:00 UTC |
| 26b2646 | Nicole Jeffery | 02 September 2021, 22:11:36 UTC | Zeroed out dummy ocean concentrations of aerosols This fix prevents erroneous aerosol fluxes between the ocean and ice which causes the code to crash with active z-aerosols in sea ice. BFB in all other configurations. | 02 September 2021, 22:11:36 UTC |
| 4b10660 | Robert Jacob | 02 September 2021, 19:35:36 UTC | Merge branch 'brhillman/remove-landm' (PR #4384) Remove the LANDM variable from the COMSRF module and replace all uses with cam_in%landfrac. LANDM was previously read in from the LANDM_COSLAT variable in the topography dataset. This was a smoothed version of the input topography, but was actually inconsistent with the land fraction data specified by the surface models. Instead of carrying around two versions of land fraction, we opt to remove the LANDM variable and instead use a land fraction consistent with the surface models. This will be BFB for all currently supported configurations except for MMF, where we were still using the old cloud effective radius calculations with LANDM instead of LANDFRAC. Since these are switched over to use LANDFRAC now, the results will be non-BFB but should still be reasonable. Also, any cases using RK microphysics will DIFF because stratiform_tend was also using LANDM, buried down in the findmcnew routine, and also in the effective radii computation. The code was changed here but not tested for this configuration because we do not have a supported configuration for that microphysics package at the moment. Fixes #4276 [non-BFB] only for MMF cases. | 02 September 2021, 19:35:36 UTC |
| 6be1207 | Robert Jacob | 02 September 2021, 19:29:02 UTC | Merge branch `forsyth2/run_e3sm/replace-template` (PR #4455) Replace run_e3sm template script with most recent version for v2. [BFB] | 02 September 2021, 19:29:02 UTC |
| 475fe28 | Wuyin Lin | 02 September 2021, 16:51:43 UTC | Renamed A_BG1850CN as BG1850CN | 02 September 2021, 16:51:43 UTC |
| 97c5292 | Wuyin Lin | 02 September 2021, 16:19:34 UTC | Remove compsets A_WCYCL1850-DIB_CMIP6 and A_WCYCL1850-DIB-ISMF_CMIP6 A_WCYCL1850-DIB-ISMF_CMIP6 has been replaced by CYRO1850. A_WCYCL1850-DIB_CMIP6 is not needed in v2. | 02 September 2021, 16:19:34 UTC |
| 0361c87 | lukaszlacinski | 02 September 2021, 13:06:47 UTC | Add backup HTTP server with Singularity source tarball | 02 September 2021, 13:06:47 UTC |
| d0af8de | Azamat Mametjanov | 02 September 2021, 05:48:04 UTC | Run ne4-grid cases on single-node | 02 September 2021, 05:48:04 UTC |
| 9cd985d | âRyan Forsyth | 10 August 2021, 17:26:09 UTC | Replace run_e3sm template Replace run_e3sm template script with most recent version for v2. [BFB] | 01 September 2021, 23:15:48 UTC |
| d0d80b9 | Azamat Mametjanov | 01 September 2021, 23:06:08 UTC | Run F-MMFOMP OpenMP-offload case without multithreading Add a testmod to set nthreads=1 | 01 September 2021, 23:06:08 UTC |
| a2f8311 | noel | 01 September 2021, 20:38:29 UTC | Merge branch E3SM-Project/ndk/homme/add-machinefiles-perlmutter (PR #4503) Add initial HOMME cmake machinefiles for perlmutter and cori-gpu. | 01 September 2021, 20:38:29 UTC |
| aa70c75 | Jon Wolfe | 01 September 2021, 16:05:49 UTC | Merge branch 'brobey/mpas-ocean/openacc-diagnostics-new' (PR #4486) Expanding OpenACC to encompass all of diagnostics routine Expanding the OpenACC region in diagnostics to encompass the whole routine. The OpenACC region has also been expanded for the equation-of-state routine since it is called from diagnostics. Two versions of the equation-of-state interface now exist -- one to enable the OpenACC calls from diagnostics and one for backward compatibility for other calls. Should not impact CPU code - results should be bit-for-bit identical [BFB] | 01 September 2021, 16:05:49 UTC |
| e6375e2 | Jon Wolfe | 01 September 2021, 14:55:07 UTC | Merge branch 'darincomeau/scripts/fix-SORRM-path' (PR #4506) Fix path for SOwISC12to60E2r4 grid domain file Fixes the path for the ocn domain file for the SOwISC12to60E2r1 grid that was mistakenly brought in with #4204. [BFB] | 01 September 2021, 14:55:07 UTC |
| c248331 | James Foucar | 31 August 2021, 19:01:12 UTC | Merge branch 'worleyph/driver-mct/eliminate-recursive-timer-call' into master (PR# 4501) Eliminate recursve call to the CPL:RUN timer The following logic in cime_run_atmocn_setup (in cime_comp_mod.F90): call t_drvstartf ('CPL:ATMOCNP',cplrun=.true.,barrier=mpicom_CPLID,hashint=hashint(7)) ... if (do_budgets) then call cime_run_calc_budgets3() endif ... call t_drvstopf ('CPL:ATMOCNP',cplrun=.true.,hashint=hashint(7)) and in cime_run_calc_budgets (also in cime_comp_mod.F90): call t_drvstartf ('CPL:BUDGET0',cplrun=.true.,budget=.true.,barrier=mpicom_CPLID) ... call t_drvstopf ('CPL:BUDGET0',cplrun=.true.,budget=.true.) nests calls to the CPL:RUN timer, leading to recursion being recorded in the timing library. This is not a bug, but it is also not necessary. Everything in the CPL:ATMOCNP timer is marked as being in the CPL:RUN timer. It adds nothing to redundantly mark the CPL:BUDGET0 timer as being in the CPL:RUN timer. Here an optional logical parameter 'in_cplrun' is added to cime_run_calc_budgets3, and to cime_run_calc_budgets1 and cime_run_calc_budgets2 for consistency, such that if set to 'true' it indicates that the call to the routine is already inside of the CPL:RUN timer and that this should not be set again within the routine. Fixes #4496 BFB | 31 August 2021, 19:01:12 UTC |
| 2ce900a | Benjamin R. Hillman | 31 August 2021, 15:37:36 UTC | Merge branch 'whannah/mmf/fix-aqua-config' (#4477) This fixes the wildcards used to specify compset options for certain MMF compsets. There are also some additional changes to the MMF physpkg.F90 initialization code to avoid using the microp_aero module. This ensures MMF cases will avoid an error that gets throw if liqcf_fix=true. The microp_aero code and the liqcf_fix option are irrelevant for all MMF configurations. fixes issue #4474 [BFB] except for MMF AQP and RCE compsets | 31 August 2021, 15:37:36 UTC |
| 0744426 | Darin Comeau | 31 August 2021, 14:44:20 UTC | Fixing path for ocn grid domain file | 31 August 2021, 14:44:20 UTC |
| bb8fab6 | Jon Wolfe | 31 August 2021, 14:37:55 UTC | Merge branch 'maltrud/e3sm/data_river_nutrients' (PR #4439) Read data river nutrients in MOSART, pass to MPASO This PR allows river outflow nutrient loads calculated by GNEWS to be read into MOSART, then passed through the coupler to MPAS-Ocean to be used as a surface nutrient flux when ocean BGC is turned on. The nutrient inputs to MOSART are concentrations (kg-Nutrient/kg-water) that vary with location, but are constant in time. These concentrations are multiplied by the calculated model runoff to create a nutrient flux to the ocean. Note that this currently only works with MOSART, not with data runoff. Also note that this is not enabled by any currently defined compset so one needs to be an expert to use this capability for now. [NML] due to new cpl flag [BFB] | 31 August 2021, 14:37:55 UTC |
| 574c2cc | Jon Wolfe | 30 August 2021, 19:37:17 UTC | Replace or remove settings for deprecated MPAS grids | 30 August 2021, 19:37:17 UTC |
| 864ad0c | Wuyin Lin | 30 August 2021, 18:20:45 UTC | Make MPASSI the default prescribed data ice model for F cases (F1850, F1950, F2010, and F20TR). The corresponding compsets that use CICE are renamed as F1850-CICE, F1950-CICE, F2010-CICE, and F20TR-CICE, respecticely. The F####-CICE compsets also set to go with SROF while the MPASSI counterparts with MOSART as river component. | 30 August 2021, 18:20:45 UTC |
| e79f647 | noel | 30 August 2021, 15:38:45 UTC | Add initial cmake machinefiles for perlmutter and cori-gpu. No impact on other machines. [bfb] | 30 August 2021, 15:38:45 UTC |
| 7794aa0 | Jon Wolfe | 30 August 2021, 15:25:30 UTC | Merge branch 'philipwjones/ocn/gputadvect' (PR #4457) GPU port and reorganization of ocean tracer advection Initial GPU port for the ocean tracer advection routines that also includes some significant reorganization of these routines. In particular, this includes: * OpenACC port for all advection options, including loop directives and data transfers * move of several advection-specific variables out of Registry and into local module allocatables to reduce size of mesh struct * move of some advection coefficient calculations out of framework and into local coefficient module * extraction of some common vertical flux calculation into a vertical module shared by both advection options * removal of some leftover pool retrievals and other general cleanup Fixes #4386 [BFB] | 30 August 2021, 15:25:30 UTC |
| bbf418f | Patrick Worley | 29 August 2021, 18:27:49 UTC | Eliminate recursve call to the CPL:RUN timer The following logic in cime_run_atmocn_setup (in cime_comp_mod.F90): call t_drvstartf ('CPL:ATMOCNP',cplrun=.true.,barrier=mpicom_CPLID,hashint=hashint(7)) ... if (do_budgets) then call cime_run_calc_budgets3() endif ... call t_drvstopf ('CPL:ATMOCNP',cplrun=.true.,hashint=hashint(7)) and in cime_run_calc_budgets (also in cime_comp_mod.F90): call t_drvstartf ('CPL:BUDGET0',cplrun=.true.,budget=.true.,barrier=mpicom_CPLID) ... call t_drvstopf ('CPL:BUDGET0',cplrun=.true.,budget=.true.) nests calls to the CPL:RUN timer, leading to recursion being recorded in the timing library. This is not a bug, but it is also not necessary. Everything in the CPL:ATMOCNP timer is marked as being in the CPL:RUN timer. It adds nothing to redundantly mark the CPL:BUDGET0 timer as being in the CPL:RUN timer. Here an optional logical parameter 'in_cplrun' is added to cime_run_calc_budgets3, and to cime_run_calc_budgets1 and cime_run_calc_budgets2 for consistency, such that if set to 'true' it indicates that the call to the routine is already inside of the CPL:RUN timer and that this should not be set again within the routine. BFB | 29 August 2021, 18:27:49 UTC |
| 4651ca0 | Azamat Mametjanov | 28 August 2021, 06:03:00 UTC | Run ibmgpu builds with 42 tasks/node in SMT1 mode | 28 August 2021, 06:03:00 UTC |
| 6d1a472 | Bob Robey | 27 August 2021, 20:42:25 UTC | Merge branch 'master' of https://github.com/E3SM-Project/E3SM into brobey/mpas-ocean/openacc-diagnostics-new | 27 August 2021, 20:42:25 UTC |
| fbc468b | Jon Wolfe | 27 August 2021, 15:03:27 UTC | Merge branch 'darincomeau/scripts/add-cryo1950-compset' (PR #4489) Adding CRYO1950 compset Adds a new Cryosphere compset that is similar to WCYCL1950, except uses Cryosphere settings for Antarctic runoff (disables Antarctic runoff from the coupler to the ocn, includes ice-shelf melt fluxes and data icebergs). This compset points to spun-up ocn/ice initial conditions (restarts from a one month G-case). [BFB] for all currently tested configurations. | 27 August 2021, 15:03:27 UTC |
| 3b1602c | Benjamin R. Hillman | 26 August 2021, 21:10:22 UTC | Merge branch 'brhillman/atm/update-rrtmgp' (#4464) Update RRTMGP to point to our (E3SM) fork, which includes the C++/YAKL port we need for SCREAM and MMF. This will be non-BFB for the RRTMGP test in e3sm_atm_integration and for the MMF tests (that all use RRTMGP) only because of a floating point literal change in RRTMGP that was made to make the Fortran and C++ codes BFB with one another. The code is otherwise BFB (i.e., a test in which this single change was implemented in master is BFB with this branch). Also brings YAKL up to latest version. [non-BFB] (only for RRTMGP and MMF tests) | 26 August 2021, 21:10:22 UTC |
| 430719f | Robert Jacob | 25 August 2021, 23:49:31 UTC | Merge branch 'mkstratos/machines/add-chry-html-dir' (PR #4462) Add CIME_HTML root and CIME_URL_ROOT for Chrysalis, allowing for e3sm_atmnbfb tests to have public web output. [BFB] | 25 August 2021, 23:49:31 UTC |
| 6fc344b | Benjamin Hillman | 25 August 2021, 22:20:01 UTC | Fix samxx CMake for new YAKL | 25 August 2021, 22:20:01 UTC |
| 36c09af | Bob Robey | 25 August 2021, 21:23:31 UTC | Fixing stray present clause on OMP directive | 25 August 2021, 21:23:31 UTC |
| 7d87136 | Benjamin Hillman | 25 August 2021, 20:10:56 UTC | Update YAKL to fix chrysalis build issue with intel | 25 August 2021, 20:10:56 UTC |
| 4683974 | Wuyin Lin | 25 August 2021, 18:34:04 UTC | Merge branch '/wlin/atm/F20TR-MPASSI' (PR #4483) Update F compsets with mpas-seaice for prescribed ice mode F20TR-MPASSI is added for 20th century transient AMIP with prescribed ice mode from mpas-seaice. Existing F2010SC5-CMIP6-MPASSI is renamed as F2010-MPASSI. Note that F2010SC5-CMIP6-MPASSI was using use_case file 2010S_cam5_CMIP6.xml, which would override the final v2 default values for clubb_tk1 and gw_convect_hcf. [BFB] for existing compsets | 25 August 2021, 18:34:04 UTC |
| d84a5e1 | Jon Wolfe | 25 August 2021, 15:15:42 UTC | Add rof2ocn_nutrients to seq_cplflds_inparm namelist, use for new fields | 25 August 2021, 15:15:42 UTC |
