Conversation
TestingI successfully ran the ocean I also ran the land-ice |
|
The MALI |
c80f515 to
7965b24
Compare
Further testingI reran the |
|
@matthewhoffman and @trhille, if you could test this out on a more recent MALI-Dev than the Compass submodule and also check if generating a variable resolution mesh with the new JIGSAW works as expected for you, that would be great. Without these fixes, any new Compass environment is likely to result in failures. |
matthewhoffman
left a comment
There was a problem hiding this comment.
Testing
I confirmed that I can build MALI with this env and that the landice full_integration suite passes. I then set up and ran the koge_bugt_s/mesh_gen mesh gen case and it passes.
The mesh generated with this PR (bottom) differs in cell layout from one generated with 1.8.0-alpha.1 (top), including an uglier disconnected region, but I attribute that just to a different random realization of JIGSAW associated with the version of that code changing and not a problem.
This version updates:
mpas_toolsto 1.3.1 -- brings in:build_jigsawwrite_netcdf()to handle a problem in the latest xarray (2025.8.0) when unlimited dimensions are sliced.Timedimension when creating MALI initial conditionsjigsawandjigsaw-pythonto 1.1.0 -- includes threading with OpenMPChecklist
Testingin this PR) any testing that was used to verify the changes