[MITgcm-devel] recomputation ?

[Jean-Michel Campin]

Hi Patrick,

It seems that since Sept 12, the AD testing on faulks is taking
a long time ( > 5 hours), instead of ~45 mn before.
And the one that takes longer is global_ocean.90x40x15.

Just the day before (Sept 11), I made this change
(in dynamics.F & implidiff.F):
> call impldiff with tracerId=-1,-2 for gU,gV resp. (to do diagnostics inside)
And the same day, we have also few changes related to useSETRLSTK
but those are unlikely to generate recomputation.

Thanks,

Jean-Michel