Robert Jacob

Make sure you use the new maint-2.0 branch for any v2.0 runs. It has that fix. Confirmed that file isn't on the server but if you're continuing a run, why...

Do you still have the old case directory that worked?

Also noticed: this is an F-case: --compset F20TR --res northamericax4v1pg2_WC14to60E2r3. That's not in our test suite. The coupled version is (--compset WCYCL1850 --res northamericax4v1pg2_WC14to60E2r3). So this may have been broken...

Looking at your old case directory, /lcrc/group/e3sm/ac.qtang/E3SMv2/v2.NARRM.amip_0201/fba8cd0a29/case_scripts, it was using 3840 tasks for all components.

That's good. But that case should default to a workable layout when no pelayout is specified. I'll change the title of this issue and assign it.

@amametjanov did you confirm that pesize=M is picked up when a layout is not specified? Qi's testing indicated that a 128-task layout was what was chosen. It would be find...

On the other hand if 128 tasks is the default on Chrysalis that we should have a 128 task graph for all MPAS components.

This summary is from the test suites unless indicated. Its not just gnu Fails with: pgi 21.11 (ascent) gnu 9.1 (ascent) ibm 16.1.1 (ascent) gnu 11.2 (crusher) gnu 11.2 (perlmutter...

What does it mean to select "atmospheric conditions" when this is part of a global coupled run?

Is this a feature need for v2 BGC runs?