[MITgcm-devel] lab_sea

[Jean-Michel Campin]

Hi Dimitri,

I checked the differences 1 tile / 2 tiles for the lab_sea experiments,
and I can see a "line" in the sea-ice area differences, after just
1 iteration. I think this is the reason why at the 2nd iteration,
we already loose many digits of precision in the cg2d_ini_res output.

When I turn off the dynamics (by the way, it would be nice to have 
a run-time flag to turn off the seaice dynamics), the comparison
is much better, and after 10 iterations we still have 12 digits
identical for cg2d_ini_res output (this is what I expect, given
the fact that some part of the sea-ice is quiet sensitive, so that
a small error can grow).

See you,

Jean-Michel