b325c71 | jedwards4b | 05 May 2021, 22:36:52 UTC | Merge pull request #3952 from mvertens/features/add_cice6 Addition of cice6 as a new CESM component Note that cice6 will only work with the NUOPC cap Test suite: hand-testing of various nuopc tests with cice6 Test baseline: no baselines yet since cice6 was not added before Test namelist changes: No Test status: bit for bit Fixes: None User interface changes?: None Update gh-pages html (Y/N)?: No | 05 May 2021, 22:36:52 UTC |
7c99238 | jedwards4b | 05 May 2021, 21:03:47 UTC | Merge pull request #3925 from briandobbins/container_inputdata This PR seeks to change the behavior of the check in case_setup on the DIN_LOC_ROOT directory so that when it doesn't exist an attempt is made to create it. This is primarily useful for containers on laptops/desktops, since the non-shared environment means there isn't usually a pre-existing directory. Any supported machines will already have a valid entry for inputdata, so it shouldn't effect them. Test suite: None yet; would like input on which to run, if needed. Test baseline: None; not answer changing Test namelist changes: No; doesn't affect namelists Test status: Doesn't impact model Fixes #3892 User interface changes?: In a very minor way; previous fails where a DIN_LOC_ROOT directory didn't exist will now succeed if it can be created. Update gh-pages html (Y/N)?: N Code review: | 05 May 2021, 21:03:47 UTC |
12acf49 | Mariana Vertenstein | 04 May 2021, 23:30:14 UTC | put in fixes for short term archiving of cice6 restart/history - thanks for Jim Edwards | 04 May 2021, 23:30:14 UTC |
b95a28b | Bill Sacks | 04 May 2021, 20:39:55 UTC | Merge pull request #3939 from billsacks/fix_retry Reset a test that is rerun after a run failure When running create_test with the --retry option set to 1 or more: If a test (ERS, ERI and possibly others) failed in the last run, some xml values would be set incorrectly in the retried run. This is because --retry sets use_existing, which in turn changes a FAILed RUN phase into a PEND. Then, in the call to _resetup_case(RUN_PHASE) from system_tests_common:run, the test reset wasn't done, since PENDing phases are not reset. This commit solves this problem by adding a comment that the now-PENDing phase is actually being RERUN, and checking for that in _resetup_case. There may be other situations where we should add this RERUN comment, but this change fixes my original issue, and it wasn't obvious to me what, if anything, else should be changed from looking through the code (especially test_scheduler.py). Test suite: - scripts_regression_tests on cheyenne - Tested with ERS_Ln3 with the diff in https://github.com/ESMCI/cime/issues/3933#issuecomment-824387449, on both my Mac (no batch system) and cheyenne (batch system): in both cases, the retried run failed with `stop_n value 1 too short` before this fix, but with this fix got all the way to the end Test baseline: N/A Test namelist changes: none Test status: bit for bit Fixes #3933 User interface changes?: N Update gh-pages html (Y/N)?: N | 04 May 2021, 20:39:55 UTC |
56a2ca1 | James Foucar | 04 May 2021, 19:13:00 UTC | Merge pull request #3949 from jasonb5/update_git_provenance Adds git submodule info to provenance Provenance now captures git status, diff and log for the repo at srcroot as well as all submodules recursively. The output of the three commands are stored in their respective output files; GIT_STATUS, GIT_DIFF and GIT_LOG. Test suite: scripts_regression_tests.py Test baseline: n/a Test namelist changes: n/a Test status: bfb Fixes #3290 User interface changes?: n/a Update gh-pages html (Y/N)?: n/a | 04 May 2021, 19:13:00 UTC |
205c5dd | James Foucar | 04 May 2021, 19:09:13 UTC | Merge pull request #3924 from jasonb5/fix_pet_multisubmit Fixes phase 1 of multisubmit not executing during rerun On a batch system with multisubmit, the first phase would be skipped if rerunning after a failed phase. This was due to case 1's "RESUBMIT" variable being decremented before phase 2. During a rerun if the specific condition is met, the test will call _setup_case which triggers the default test settings to be loaded. Test suite: scripts_regression_tests.py Test baseline: Test namelist changes: Test status: Fixes #3805 User interface changes?: n/a Update gh-pages html (Y/N)?: N Code review: | 04 May 2021, 19:09:13 UTC |
96396ae | Brian Dobbins | 04 May 2021, 15:35:02 UTC | Removed 'exist_ok' in makedirs call to be compatible with Python2 | 04 May 2021, 15:35:02 UTC |
6af7a49 | Mariana Vertenstein | 03 May 2021, 21:54:23 UTC | udpated ChangeLog to master | 03 May 2021, 21:54:23 UTC |
c2015d1 | Mariana Vertenstein | 27 April 2021, 17:38:06 UTC | removed added DEBUG statements | 03 May 2021, 21:41:00 UTC |
9f909f8 | Mariana Vertenstein | 27 April 2021, 17:34:07 UTC | changes needed to add cice6 as an external component | 03 May 2021, 21:41:00 UTC |
74aa55d | Chris Fischer | 08 April 2021, 14:55:10 UTC | Update for cime5.8.42 | 03 May 2021, 21:41:00 UTC |
c613e6a | Chris Fischer | 03 May 2021, 19:45:22 UTC | Update for cime5.8.44. | 03 May 2021, 19:45:22 UTC |
6cbfdd6 | jedwards4b | 03 May 2021, 19:22:19 UTC | Merge pull request #3951 from ESMCI/fischer/mpt_cheyenne Revert cheyenne intel mpt to 2.22. Having issues with mpt/2.23 for the intel compiler on cheyenne. Some WACCM and WACCM-X tests were failing. Test suite: ERS_Ld3.f19_f19_mg17.FXHIST.cheyenne_intel.cam-waccmx_weimer Test baseline: Test namelist changes: Test status: bit for bit Fixes [CIME Github issue #] User interface changes?: Update gh-pages html (Y/N)?: | 03 May 2021, 19:22:19 UTC |
1e3fd5e | Chris Fischer | 03 May 2021, 16:25:32 UTC | revert cheyenne intel mpt to 2.22. | 03 May 2021, 16:25:32 UTC |
454c8fb | Jason Boutte | 30 April 2021, 20:18:00 UTC | Changes phase and forces reset | 30 April 2021, 20:18:00 UTC |
296aeaf | jedwards4b | 30 April 2021, 19:29:43 UTC | Merge pull request #3948 from jedwards4b/move_to_py3 | 30 April 2021, 19:29:43 UTC |
d0627fb | Jason Boutte | 29 April 2021, 19:40:50 UTC | Adds GIT_CONFIG and GIT_REMOTE to provenance | 29 April 2021, 19:40:50 UTC |
c6d31ec | Jason Boutte | 29 April 2021, 16:52:51 UTC | Copies compatibility fix to other tests | 29 April 2021, 16:52:51 UTC |
677aa17 | Jason Boutte | 29 April 2021, 16:14:55 UTC | Fixes py2 compatibility | 29 April 2021, 16:14:55 UTC |
b1e2019 | Jason Boutte | 29 April 2021, 02:01:07 UTC | Adds recursive git output to provenance | 29 April 2021, 02:01:07 UTC |
5ce0e25 | Jim Edwards | 28 April 2021, 23:01:28 UTC | remove deprecated isAlive | 28 April 2021, 23:01:28 UTC |
cf09afe | Jim Edwards | 28 April 2021, 22:13:13 UTC | change workflow remove py 3.10 | 28 April 2021, 22:13:13 UTC |
8d8d687 | Jim Edwards | 28 April 2021, 22:04:13 UTC | change workflow to remove python 2.7 test and add 3.7 and 3.10 | 28 April 2021, 22:04:13 UTC |
e0d0275 | Jim Edwards | 28 April 2021, 21:47:37 UTC | change shebang to explict python3 | 28 April 2021, 21:47:37 UTC |
3bd687d | Chris Fischer | 26 April 2021, 22:13:30 UTC | Update for cime5.8.43 | 26 April 2021, 22:13:30 UTC |
40be61f | fischer-ncar | 26 April 2021, 21:24:18 UTC | Merge pull request #3941 from ESMCI/fischer/cheyenne_intel Update Cheyenne intel compiler to 19.1.1, also update mpt to 2.23 and pnetcdf to 1.12.2. Update ESMF libraries for intel to 8.1.1. Test suite: aux_cime_baselines Test baseline: Test namelist changes: Test status: bit for bit Fixes [CIME Github issue #] User interface changes?: Update gh-pages html (Y/N)?: | 26 April 2021, 21:24:18 UTC |
2b1b610 | Chris Fischer | 26 April 2021, 21:07:47 UTC | Remove old failure comment, and update netcdf module | 26 April 2021, 21:07:47 UTC |
2b2a18d | jedwards4b | 26 April 2021, 20:56:24 UTC | Merge pull request #3942 from sjsprecious/master Fix issue to build a testcase on Casper through an interactive node In order to compile CAM on Casper through an interactive node with the recently added pgi-gpu and nvhpc-gpu compilers, we need to make the following changes (thanks to @jedwards4b 's help): Change the PIO_FILESYSTEM_HINTS from glade to gpfs. glade is not a supported option but there is just no error check about it. Load the cmake module. Code review: @jedwards4b @johnmauff | 26 April 2021, 20:56:24 UTC |
83ddb3e | Jim Edwards | 26 April 2021, 20:04:37 UTC | fix py3 error in srun syntax | 26 April 2021, 20:04:37 UTC |
f7f4114 | Jian Sun | 26 April 2021, 19:56:18 UTC | Follow Jim's suggestion to change PIO_FILESYSTEM and load cmake in order to compile CAM on Casper with pgi-gpu and nvhpc-gpu in an interactive node modified: config_compilers.xml modified: config_machines.xml | 26 April 2021, 19:56:18 UTC |
eea98c2 | Chris Fischer | 26 April 2021, 19:09:04 UTC | Update intel compiler and esmf for intel on cheyenne. | 26 April 2021, 19:09:04 UTC |
c15311a | Bill Sacks | 23 April 2021, 19:23:23 UTC | Fix for python2 | 23 April 2021, 19:23:23 UTC |
1faa36f | Bill Sacks | 23 April 2021, 17:30:31 UTC | Add a TESTRUNFAILRETRY to pick up possible errors on retry This is like TESTRUNFAIL but with some extra code to confirm that the test's values are reset when it is rerun. | 23 April 2021, 19:13:30 UTC |
51a18b9 | Bill Sacks | 22 April 2021, 22:22:17 UTC | Reset a test that is rerun after a run failure When running create_test with the --retry option set to 1 or more: If a test (ERS, ERI and possibly others) failed in the last run, some xml values would be set incorrectly in the retried run. This is because --retry sets use_existing, which in turn changes a FAILed RUN phase into a PEND. Then, in the call to _resetup_case(RUN_PHASE) from system_tests_common:run, the test reset wasn't done, since PENDing phases are not reset. This commit solves this problem by adding a comment that the now-PENDing phase is actually being RERUN, and checking for that in _resetup_case. There may be other situations where we should add this RERUN comment, but this change fixes my original issue, and it wasn't obvious to me what, if anything, else should be changed from looking through the code (especially test_scheduler.py). Resolves ESMCI/cime#3933 | 22 April 2021, 22:22:17 UTC |
8b48a2e | Bill Sacks | 22 April 2021, 18:58:59 UTC | Merge pull request #3938 from billsacks/pr_template_cleanup Remove "Code review" line from PR template This is redundant with GitHub's PR reviewers list Test suite: none Test baseline: n/a Test namelist changes: none Test status: bit for bit Fixes none User interface changes?: no Update gh-pages html (Y/N)?: N Code review: How very meta | 22 April 2021, 18:58:59 UTC |
ad5f1d2 | Bill Sacks | 22 April 2021, 17:35:35 UTC | Remove "Code review" line from PR template This is redundant with GitHub's PR reviewers list | 22 April 2021, 17:35:35 UTC |
6407bc2 | jedwards4b | 21 April 2021, 22:24:27 UTC | Merge pull request #3935 from sbrus89/wav_primary_fix This PR allows WAV to be identified as the primary component for compsets with WAVEWATCH III, such as: 2000_DATM%IAF_SLND_SICE_SOCN_SROF_SGLC_WW3_SIAC_SESP Test suite: Test baseline: Test namelist changes: Test status: bit for bit Fixes #3934 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 21 April 2021, 22:24:27 UTC |
b4af9de | Steven Brus | 21 April 2021, 21:56:54 UTC | Add support for WAV as primary component | 21 April 2021, 22:00:32 UTC |
44b241d | James Foucar | 20 April 2021, 19:09:31 UTC | Merge pull request #3926 from jasonb5/update_case_status Update case status to include jobid on batch system Adds jobid to CaseStatus lines when running on a batch system. Test suite: scripts_regression_tests.py Test baseline: n/a Test namelist changes: n/a Test status: Fixes #3255 User interface changes?: n Update gh-pages html (Y/N)?: n Code review: @jedwards4b @jgfouca | 20 April 2021, 19:09:31 UTC |
8d0c287 | Jason Boutte | 20 April 2021, 18:11:53 UTC | Reverts changes to build.py | 20 April 2021, 18:11:53 UTC |
ebdb297 | James Foucar | 19 April 2021, 22:48:24 UTC | Merge pull request #3930 from ESMCI/rljacob/move-for-e3sm Allow E3SM to use its versions of coupler and data,stub,x models. Modify config_files.xml in config/e3sm to point to new locations of the driver (MCT and MOAB) and MCT data,stub,x models within E3SM Slightly modify buildlib.csm_share to use COMP_ROOT_DIR_CPL to find shr dir. Still using cime/src/externals and cime/src/share Test with e3sm_integration (build only) and an E3SM F-case (to test BFB) [BFB] | 19 April 2021, 22:48:24 UTC |
e256ea7 | jedwards4b | 16 April 2021, 21:07:33 UTC | Merge pull request #3927 from jedwards4b/add_neon_to_inputdata add neon server to inputdata | 16 April 2021, 21:07:33 UTC |
1cc7d71 | Jim Edwards | 16 April 2021, 17:39:46 UTC | add neon server to inputdata | 16 April 2021, 17:39:46 UTC |
de7f886 | Brian Dobbins | 16 April 2021, 03:00:01 UTC | Fixed a typo (missing ')'..) | 16 April 2021, 03:00:01 UTC |
fa9bc7a | Brian Dobbins | 16 April 2021, 02:55:32 UTC | Now imports errno and checks for permission error on directory creation. It does NOT check that a user can write to that directory, since that's not necessary in many circumstances (eg, existing data). Also uses Python3's os.makedirs w/ exist_ok, so this isn't compatible with Python2 now. | 16 April 2021, 02:55:32 UTC |
3611a72 | Robert Jacob | 16 April 2021, 01:11:21 UTC | Move all cime component models Move points to all the data, stub and x models to their new locations in SRCROOT/components | 16 April 2021, 01:11:21 UTC |
3c01b0f | Jason Boutte | 16 April 2021, 00:29:23 UTC | Fixes linting | 16 April 2021, 00:29:23 UTC |
346168d | Jason Boutte | 15 April 2021, 22:25:11 UTC | Merge branch 'master' into update_case_status | 15 April 2021, 22:25:11 UTC |
7135fba | Jason Boutte | 15 April 2021, 22:21:51 UTC | Updates CaseStatus to include jobid on batch systems | 15 April 2021, 22:21:51 UTC |
b289939 | Brian Dobbins | 13 April 2021, 20:06:20 UTC | Fix for indentation issue | 13 April 2021, 20:06:20 UTC |
4abbc92 | Brian Dobbins | 13 April 2021, 19:27:06 UTC | Changed behavior of DIN_LOC_ROOT directory so that when it doesn't exist an attempt is made to create it. This is primarily useful for containers on laptops/desktops, since the non-shared environment means there isn't usually a pre-existing directory | 13 April 2021, 19:27:06 UTC |
edf89b1 | James Foucar | 13 April 2021, 17:38:37 UTC | Fix for buildlib.csm_share It was looking in the old location for mct/shr | 13 April 2021, 17:38:37 UTC |
e580706 | Robert Jacob | 12 April 2021, 23:42:43 UTC | Move driver for E3SM Move the location of the driver for E3SM config Now in E3SM/driver-mct | 12 April 2021, 23:42:43 UTC |
d5506bf | Jason Boutte | 12 April 2021, 17:59:23 UTC | Fixes linting error | 12 April 2021, 17:59:23 UTC |
f543392 | Bill Sacks | 10 April 2021, 13:37:30 UTC | Merge pull request #3914 from billsacks/retry_baseline_overwrite Fix create_test --retry when generating baselines Test suite: - Manually ran the reproducer in #3912 - scripts_regression_tests A & B (pass except for the pre-existing issue in #3913) - full scripts_regression_tests on cheyenne: pass Test baseline: n/a Test namelist changes: none Test status: bit for bit Fixes #3912 User interface changes?: N Update gh-pages html (Y/N)?: N Code review: @jgfouca @jedwards4b | 10 April 2021, 13:37:30 UTC |
255bb04 | Jason Boutte | 10 April 2021, 07:21:10 UTC | Fixes executing phase1 of SystemTestsCompareTwo during rerun | 10 April 2021, 07:21:10 UTC |
599e0bd | fischer-ncar | 09 April 2021, 22:34:37 UTC | Merge pull request #3917 from ESMCI/fischer/cime_sep Generalize paths to source code in build.py and buildlib.py, to use config_files.xml. Update Makefile to be able to handle different locations of of mct and pio. Also add an if statement place holder for furture CESM share repository. Update case_clone.py to be able to handle the future CESM share repository. Add get_src_root to utils.py, similar to get_cime_root. Needed this to correct an assumption in utilities.py that srcroot=cimeroot/... Added place holder if statement in scripts_regression_tests for the new CESM share repository. Did some minor cleanup of build_scripts. There's a couple of place holder if statements that I added that will be updated when we bring in the new CESM share repository. Build_scripts and CMake files will still need to be updated when the CESM share repository is ready. I didn't include any changes since the structure of the CESM share repository will be changing. Test suite: scripts_regression_tests Test baseline: Test namelist changes: Test status: bit for bit Fixes [CIME Github issue #] User interface changes?: Update gh-pages html (Y/N)?: Code review: | 09 April 2021, 22:34:37 UTC |
14bbf27 | Chris Fischer | 09 April 2021, 21:57:02 UTC | Fix pylint issue. | 09 April 2021, 21:57:02 UTC |
25d85e5 | Chris Fischer | 09 April 2021, 21:52:36 UTC | Update for cime5.8.41 and cime5.8.42 | 09 April 2021, 21:52:36 UTC |
5d79468 | Chris Fischer | 09 April 2021, 21:29:17 UTC | Remove unneeded code from get_src_root. | 09 April 2021, 21:29:17 UTC |
86f0fc8 | jedwards4b | 08 April 2021, 21:53:50 UTC | Merge pull request #3922 from sjsprecious/master In the current configuration files, there are settings related to Casper but they have some drawbacks: It uses the Slurm scheduler while Casper recently has switched to PBS. Some modules like compiler and openmpi are out of date. The way to use the NVIDIA V100 GPUs also changes due to the switch of job scheduler. Therefore, we have made a few updates to the configuration files so that we could run CESM on Casper with PBS scheduler. We have added three compilers for CPU simulation: pgi intel nvhpc and two compilers for GPU simulation in the future: pgi-gpu nvhpc-gpu Some additional files and OpenACC directives are also added here to enable the PUMAS/MG3 codes running on GPU. These changes, together with the two compilers for GPU, have no effect for the pure CPU simulation on Casper. To validate our changes, we use the ensemble consistency test (ECT) since we are using a different machine and compiler than Cheyenne. Some details related to ECT are listed below: CAM codes: https://github.com/PUMASDevelopment/CAM/tree/sjsp_cam_gpus_p3 PUMAS codes: https://github.com/ESCOMP/PUMAS/tree/sjsp_cam_gpus_pumas_p3 Compset: F2000climo Resolution: f19_f19_mg17 With these information, we then perform two sets of ECT: Generate the baseline by using the intel compiler on Cheyenne. Then we do the ECT for intel, pgi and nvhpc compilers on Casper. The test with intel compiler on Casper passes the ECT, while the tests with pgi and nvhpc compilers on Casper fail the ECT. Generate the baseline by using the pgi compiler on Casper. Then we do the ECT for intel, pgi and nvhpc compilers on Casper. All the three tests on Casper pass the ECT. One additional thought: Currently, the GPU simulation will only use 1 V100 GPU on Casper by default. A new argument for create_newcase or create_test may be added to let users specify the number of GPUs for their simulations Code review: @Katetc @johnmauff @jedwards4b | 08 April 2021, 21:53:50 UTC |
80c3cae | Jian Sun | 08 April 2021, 21:39:43 UTC | simplify the mpilib specification for Casper modified: config_machines.xml | 08 April 2021, 21:39:43 UTC |
d2ecb26 | Jian Sun | 08 April 2021, 21:24:47 UTC | Remove mpi-serial for Casper based on Jim's suggestion modified: config_machines.xml | 08 April 2021, 21:24:47 UTC |
3b9b63e | Chris Fischer | 08 April 2021, 18:08:44 UTC | Fix pes path. | 08 April 2021, 18:08:44 UTC |
1c68bb0 | Chris Fischer | 08 April 2021, 17:15:35 UTC | Fix typo | 08 April 2021, 17:15:35 UTC |
68f1873 | Chris Fischer | 08 April 2021, 16:55:38 UTC | Address Jim E. requests. | 08 April 2021, 16:55:38 UTC |
7aa6b0c | jedwards4b | 08 April 2021, 16:37:58 UTC | Merge pull request #3921 from jedwards4b/query_config_py3_update add UTF-8 decoding for xml option in query_config Test suite: hand testing query_config with py2 and py3 Test baseline: Test namelist changes: Test status: bit for bit Fixes #3910 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 08 April 2021, 16:37:58 UTC |
8a7186a | Jim Edwards | 08 April 2021, 15:31:44 UTC | fix query_config for py3 output format | 08 April 2021, 15:31:44 UTC |
b655c91 | jedwards4b | 08 April 2021, 13:16:52 UTC | Merge pull request #3918 from jedwards4b/ninst_fix no need to test ESMF_AWARE_THREADING here Test suite: hand tested as in #3916 and scripts_regression_tests.py Test baseline: Test namelist changes: Test status: [bit for bit, roundoff, climate changing] Fixes #3916 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 08 April 2021, 13:16:52 UTC |
c3694cb | Chris Fischer | 08 April 2021, 00:41:02 UTC | Fix what I broke to fix pylint issue | 08 April 2021, 00:41:02 UTC |
24f84d5 | Chris Fischer | 08 April 2021, 00:19:53 UTC | Fix more pylint issues | 08 April 2021, 00:19:53 UTC |
b58ea81 | Chris Fischer | 07 April 2021, 23:56:56 UTC | Fix pylint issues | 07 April 2021, 23:56:56 UTC |
0d278e7 | Chris Fischer | 07 April 2021, 23:38:43 UTC | Reverting create_test | 07 April 2021, 23:38:43 UTC |
aa72101 | Jim Edwards | 07 April 2021, 23:07:10 UTC | no need to test ESMF_AWARE_THREADING here | 07 April 2021, 23:07:10 UTC |
483295c | Chris Fischer | 07 April 2021, 21:54:25 UTC | Need to replace case_cimeroot with case_srcroot | 07 April 2021, 21:54:25 UTC |
dd8a91b | Bill Sacks | 07 April 2021, 21:05:16 UTC | Fix create_test --retry when generating baselines Resolves ESMCI/cime#3912 | 07 April 2021, 21:05:16 UTC |
f79d313 | Chris Fischer | 07 April 2021, 17:43:21 UTC | Generalize builds, and start seperation of cesm build | 07 April 2021, 17:44:53 UTC |
bc64962 | Chris Fischer | 07 April 2021, 17:42:15 UTC | Generalize builds, and start seperation of cesm build | 07 April 2021, 17:44:53 UTC |
871a361 | Jian Sun | 07 April 2021, 16:41:16 UTC | Remove -lpmi flag, which will cause PMIx version error after Casper switches from Slurm to PBS modified: config_compilers.xml | 07 April 2021, 16:41:16 UTC |
81d4f20 | James Foucar | 05 April 2021, 21:43:54 UTC | Merge pull request #3904 from ESMCI/jayeshkrishna/spio_stats_provenance Save Scorpio I/O statistics Save Scorpio I/O performance stats for E3SM runs. All stats are saved in spio_stats*.tar.gz Test suite: scripts_regression_tests Test baseline: None Test namelist changes: None Test status: bit for bit User interface changes?: N Update gh-pages html (Y/N)?: N Code review: @jgfouca | 05 April 2021, 21:43:54 UTC |
ada43e1 | jedwards4b | 05 April 2021, 21:23:42 UTC | Merge pull request #3899 from jedwards4b/esmf_threading support for threading in two different ways for esmf Adds threading support for nuopc in two ways with the introduction of a new xml variable in cmeps ESMF_AWARE_THREADING. When this variable is False or None the threading implementation is the same as that in mct, this is the default. When ESMF_AWARE_THREADING is True the driver uses all available mpi tasks on a given node and component threads are launched using pthreads. This allows for each component to specify its thread count independent of the other components. It also changes the definition of ROOTPE. The pelayout was changed to add PSTRID and to document the value of ESMF_AWARE_THREADING and how that value affects the ROOTPE variable. Test suite: scripts_regression_tests.py + manual tests of pelayout script Test baseline: Test namelist changes: Test status: bit for bit Fixes [CIME Github issue #] Depends on ESCOMP/CMEPS#144 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 05 April 2021, 21:23:42 UTC |
b86d2e1 | jedwards4b | 05 April 2021, 21:23:12 UTC | Merge pull request #3908 from jedwards4b/izumi_nag_mpiserial_fix Izumi module system requires that you explicitly unload mpi to build mpi-serial. Test suite: Hand tested SMS_Vnuopc_D_P1x1.f10_f10_mg37.I2000Clm50Sp.izumi_nag.clm-default Test baseline: Test namelist changes: Test status: bit for bit Fixes #3907 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 05 April 2021, 21:23:12 UTC |
6eac99c | Jim Edwards | 05 April 2021, 21:00:16 UTC | fix workflow | 05 April 2021, 21:00:16 UTC |
e0e4939 | Jim Edwards | 05 April 2021, 15:03:36 UTC | remove explicit RLIMIT | 05 April 2021, 15:03:36 UTC |
46ba776 | Jim Edwards | 03 April 2021, 18:28:33 UTC | fix github workflow issue | 03 April 2021, 18:28:33 UTC |
098261a | Jim Edwards | 03 April 2021, 18:23:53 UTC | fix issue with build of nag mpiserial | 03 April 2021, 18:23:53 UTC |
4b04870 | Jim Edwards | 01 April 2021, 16:46:53 UTC | reset R_LIMIT to -1 | 01 April 2021, 16:46:53 UTC |
c7dd84a | Jian Sun | 01 April 2021, 15:05:05 UTC | Bug fix for pgi-gpu compiler modified: config_compilers.xml | 01 April 2021, 15:05:05 UTC |
4fc6a89 | Jian Sun | 01 April 2021, 14:47:29 UTC | Add missing information for pgi-gpu and nvhpc-gpu compilers modified: config_compilers.xml | 01 April 2021, 14:47:29 UTC |
a65152b | Jian Sun | 01 April 2021, 02:34:31 UTC | Add compiler names for GPU codes modified: config_batch.xml modified: config_compilers.xml modified: config_machines.xml | 01 April 2021, 02:34:31 UTC |
aa890f5 | Jian Sun | 01 April 2021, 02:26:09 UTC | Rename the depends files renamed: Depends.nvhpc -> Depends.nvhpc-gpu renamed: Depends.pgi -> Depends.pgi-gpu | 01 April 2021, 02:26:09 UTC |
000b3c7 | James Foucar | 31 March 2021, 18:48:59 UTC | Merge pull request #3903 from jedwards4b/domain_root_fix handle case when domain_root is not present domain_root node may not always be present, need to handle this case. Test suite: scripts_regression_tests.py Test baseline: Test namelist changes: Test status: [bit for bit, roundoff, climate changing] Fixes #3902 User interface changes?: Update gh-pages html (Y/N)?: Code review: | 31 March 2021, 18:48:59 UTC |
b050b7f | Jim Edwards | 31 March 2021, 18:13:50 UTC | handle case when domain_root is not present | 31 March 2021, 18:13:50 UTC |
af4d0ae | mvertens | 30 March 2021, 16:53:43 UTC | Merge pull request #3901 from billsacks/dlnd_new_glc_forcings Point to new dlnd scpl forcing data These data are used to force cases with GLC - i.e., T compsets. Test suite: Just ran `SMS_Ly35.f09_g17_gl4.T1850G.cheyenne_intel` in the context of CESM's nuopc_dev branch Test baseline: Test namelist changes: Test status: climate changing - just for T compsets Fixes #3835 User interface changes?: N Update gh-pages html (Y/N)?: N Code review: | 30 March 2021, 16:53:43 UTC |
64952c7 | jayeshkrishna | 30 March 2021, 16:34:42 UTC | Save Scorpio I/O stats Save Scorpio I/O performance stats for E3SM runs. All stats are saved in spio_stats*.tar.gz | 30 March 2021, 16:36:16 UTC |
b3dc428 | Chris Fischer | 30 March 2021, 15:56:05 UTC | Update for cime5.8.40 | 30 March 2021, 15:56:05 UTC |
a900c8e | jedwards4b | 30 March 2021, 13:02:45 UTC | Merge pull request #3900 from mvertens/feature/nuopc_grids_bugfix bug fix for generating C/G nuopc compsets This is a minor bugfix that only effects nuopc compsets. Without this fix C/G compsets could not be created. Test suite: Test baseline: Test namelist changes: Test status: [bit for bit, roundoff, climate changing] Fixes: None User interface changes?: No Update gh-pages html (Y/N)?: N Code review: @jedwards4b | 30 March 2021, 13:02:45 UTC |
cee699d | Mariana Vertenstein | 30 March 2021, 01:22:13 UTC | bug fix for generating C/G nuopc compsets | 30 March 2021, 01:22:13 UTC |
cb0b8ea | Bill Sacks | 29 March 2021, 22:59:29 UTC | Point to new dlnd scpl forcing data These data are used to force cases with GLC - i.e., T compsets | 29 March 2021, 22:59:29 UTC |
1e754bf | Jian Sun | 29 March 2021, 16:00:46 UTC | Add openacc directives to macrop_driver.F90 modified: Depends.nvhpc modified: Depends.pgi | 29 March 2021, 16:00:46 UTC |