[MITgcm-devel] isomip verification experiment

[Jean-Michel Campin]

Hi Martin,

I would like to reduce the size of the tiles (to have more tiles)
in both the forward and AD isomip set-up.
thinking of changing to sNx,sNy=25,25, and changing SIZE.h_mpi
to nPx,nPy=2,2.

The main motivation would be to test MPI with more than 2 procs
(and still have more than 1 tile per proc, since will still
have nSy=2), and since this set-up is not too small (and tested
with TAF & OpenAD), would be a good one.

Do you agree with these changes ?

Cheers,
Jean-Michel