Update CTM to be in line with GEOSgcm 10.21.0 (and a bit more) #42
Conversation
mathomp4
added
0 diff
mathomp4
added
the
Contingent - DNA
|
Blocking this as the build tests seem to have failed... |
|
Also converting to draft so it can't be merged. |
mathomp4
removed
the
Contingent - DNA
|
Ohh. I think I know the answer to this. I might need to wait until @bena-nasa is back from holiday break. I'll make my best attempt to figure out the issue, but I think I'll need to edit some new(-ish) code of his in FV3. |
@mmanyin Okay. I mainly made the PR so I didn't forget anything myself. I'll probably just use this one as the omnibus PR for various non-CTM updates (MAPL, etc.) |
|
Okay. This has fixes from @bena-nasa in MAPL 2.15.0 and FV3 GC 1.6.0 that should let GEOSctm build! |
This PR updates the CTM to be inline with GEOSgcm 10.21.0. Actually a little more than GEOSgcm 10.21.0. There have been some changes to GEOSgcm
mainand GEOSgcm_App that are to better support working at NAS. As these changes were (trivially) zero-diff, I decided to add them in this PR rather than wait for the next stable version of GEOSgcm (which will be next year).Note 1
I know that #41 exists. But I didn't want to "overwrite" that in case @mmanyin or @JulesKouatchou were currently testing it. This PR is sort of built on #41 so if #41 goes in first, this PR should get "smaller" as we merge in all the changes already in #41.
Note 2
I am not sure how to label this. GEOSgcm 10.21.0 is non-zero-diff to GEOSgcm v10.19.5, but I'm not sure if any of that non-zero-diff-ness spreads to GEOSctm. Much of the NZD was in GEOSgcm_GridComp, so that doesn't matter.
This does, however, update GEOSchem_GridComp to v1.7.0 which was NZD to the GCM.
It also updates ESMA_env and ESMA_cmake to change compiler to Intel 2021.3 and default Intel compiler flags to better generalize GEOS (makes things easier at NAS). These two changes were NZD to the GEOSgcm, but weirdly zero-diff to the GEOSldas. So...not sure what will happen here.
I'm guessing...non-zero-diff???