drying_slope/1km/single_layer/loglaw fails

[gcapodag]

I am getting a runtime error in the forward step when running this test. This is the error:

At line 1074 of file mpas_ocn_vmix.F
Fortran runtime error: Index '0' of dimension 1 of array 'kineticenergycell' below lower bound of 1

It looks like those lines have been modified by @cbegeman with this commit.

@cbegeman or @xylar, would you mind checking if you are getting the same failure? Thank you.

[cbegeman]

I'll look into it. This likely an E3SM issue, not a compass one. I'll confirm first, though. I have this note on L1072:

! For some reason we can use k-1 even when N=Nsurf

So it looks like I was aware that this could be a problem but didn't encounter that error. What machine, compiler combination are you using? Is debug on?

@xylar wasn't involved in that work so he doesn't need to look into it unless he wants to.

[xylar]

@cbegeman and @gcapodag, I'll leave this up to the 2 of you for now but ping me again if you would like my input.

[gcapodag]

I am on perlmutter with gnu and debug on.

[gcapodag]

make gnu-cray DEBUG=true OPENMP=false

[cbegeman]

@gcapodag Given the code freeze, I only opened a PR on ocean discussion E3SM-Ocean-Discussion/E3SM#71. If you would like to use log-law drag in the meantime, I suggest cherry-picking this commit c551420785c3be55010082325f691ad8331c7705.

[cbegeman]

Migrated to E3SM-Project/E3SM#6055