add 05degree blom #377
add 05degree blom #377
Conversation
…using vcoord_type = cntiso_hybrid.
|
@matsbn @TomasTorsvik - I'd really like to have this as part of noresm2_5_alpha04 that I'm working on creating this week. Consequently, I thought we could move the HAMOCC capability for the half-degree blom grid to an upcoming PR and get this PR accepted first. I've implemented a new set of regression tests that more fully cover other new grids and archiving/resubmit capability as part of this PR. |
I agree it makes sense to deal with HAMOCC capability separately, and not halt the progress on the physics side. |
| @@ -401,7 +401,7 @@ subroutine InitializeAdvertise(gcomp, importState, exportState, clock, rc) | |||
| if (ChkErr(rc, __LINE__, u_FILE_u)) return | |||
| read(cvalue,*) nthrds | |||
| endif | |||
| !$ call omp_set_num_threads(nthrds) | |||
| !$ call omp_set_num_threads(nthrds) | |||
There was a problem hiding this comment.
For all other OMP calls, the prefix !$omp <something openmp> is used. I find it strange that this is not required here. Is this documented somewhere?
There was a problem hiding this comment.
Good question. This is not being used at this point. I'll raise it as an issue since as you mention it does not effect this PR.
|
@matsbn - yes - BLOM will still work with an MCT based NorESM - but my understanding (based on talking with @TomasTorsvik ) is that master will no longer be used in noresm2.1.x and noresm2.3.x developments - there are now separate branches for this development. Is this not the case? If this is the case - we should consider removing the mct code from master moving forwards. |
|
@matsbn , @mvertens I think we should discuss this in a BLOM-core meeting before making a decision. |
|
It is still very useful in ongoing projects to be able to swap in current master in a MCT based NorESM case. We don't really have separate branches fulfilling this purpose to my understanding. The "release-*" branches potentially could, but was intended for final testing before release and later maintenance, bug fixes or documentation enhancements but not answer-changing. A great discussion point for a BLOM-core meeting! |
|
I tried the PR with a NorESM2.0.6 setup and get a buildlib failure. Since I have ongoing projects and numerous development tests on this old NorESM base that could still benefit greatly from master compatibility, this PR should ideally not break it. |
|
@matsbn - sorry about that. I had assumed that we no longer needed the 2.0.6 functionality in master. Its easy for me to add it back in. |
|
@matsbn - I've back out the changes and this should work for you now. |
|
It builds now, but some of the modified compsets does not work with NorESM2. I think we should find a way to keep the old compsets in parallel for some time. Maybe the old compsets could be renamed to start with "N2", e.g NOIIAOC20TR -> N2OIIAOC20TR. For this specific compset, also directly creating a case with the complete old compset string "20TR_DATM%JRA_SLND_CICE%NORESM-CMIP6_BLOM%ECO_DROF%JRA_SGLC_SWAV" fails with a shr_stream_findBounds error. I wonder if the removed RUN_STARTDATE modification is the reason. |
|
@matsbn - I'll fix that as well. |
…resm2.3 with the prefix N2
|
@matsbn - I've pushed fixes to the compsets back that you suggested. Hopefully this all works now. |
add a 1/2 degree blom grid - also will require CICE and ccs_config PRs and is implemented in NorESM2.5 development branch (https://github.com/mvertens/NorESM/tree/feature/noresm2_5_alpha04_v3.
fix in nuopc cap for runid that has too short
addition of new tests for short term archiving, 2 degree, intel-oneapi, etc
simplification of buildlib since master will now only support nuopc coupling infrastructure and cdeps
ran new regression test successfully
NOTE: for HAMOCC need (this should be done in another PR) - fdepfile, swaclimfile, rivinfile, ndepfile