-
Notifications
You must be signed in to change notification settings - Fork 360
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add DLND to moab coupler #6567
Add DLND to moab coupler #6567
Conversation
use existing lnd_domain attribute in infodata we do not differentiate on coupler side if data land or not still need to send the proper land domain file
domain file is actually set, in shr_strdata_type%domainFile all the other data models use the first stream file for domain information land is the one that actually has that member set in the SDLND structure it is easier to get it from there, then from stream the domain file in regular land case is defined by fatmlndfrac file
iMOAB_ApplyScalarProjectionWeights is not used anymore all maps are done now with seq_map method, and there is only one instance of iMOAB_ApplyScalarProjectionWeights in E3SM using moab driver
only intel on chrysalis complained about it It could be a cause for trouble later on
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed the code and made comments. I haven't run this branch and tested it just yet.
I tested this branch on different platforms, and chrysalis is working fine; gce, perlmutter, and my laptop (gnu compiler) all fail to compress the log files at the end. So the bug I saw earlier is not solved yet. Somehow, it worked on gce when I tested 2 weeks ago; needs more work |
compute rof lnd interaction only if rof_c2_lnd is true we were computing an unnecessary map always even if rof_c2_lnd is false (for most coupled cases it is true)
in case rof c2 lnd, we were not retrieving the the MPI group needed for comm graph computation, for same grid path we were retrieving it for intx only how come we never had samegrid land rof before ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Few more comments to address
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for addressing my comments!
Add data land capability for moab driver; case tested: --res r05_r05 --compset RMOSGPCC standard data moab model, mainly adapt from data ocean implementation major change: Initialize all factors set by MOAB area correction init routine to 1.0. It affects all models, but it seems that some values were not initialized for data land case. On chrysalis, with intel compiler, it actually complained about non initialized values when applying correction factors. The compressing file errors seen during development were caused by non-ascii characters in the log files. Those were caused by an uninitialized mapper, for which we were still creating a MOAB map. It was never used, it is fixed properly now, the moab map is not computed anymore. [BFB]
Add data land capability for moab driver;
case tested: --res r05_r05 --compset RMOSGPCC
standard data moab model, mainly adapt from data ocean implementation
major change: Initialize all factors set by MOAB area correction init routine to 1.0. It affects all models,
but it seems that some values were not initialized for data land case.
On chrysalis, with intel compiler, it actually complained about non initialized values when applying correction factors.
The compressing file errors seen during development were caused by non-ascii characters in the log files.
Those were caused by an uninitialized mapper, for which we were still creating a MOAB map. It was never used,
it is fixed properly now, the moab map is not computed anymore.
[BFB]