[MITgcm-devel] [MITgcm-cvs] MITgcm/model/src CVS Commit
Dimitris Menemenlis
menemenlis at jpl.nasa.gov
Thu Apr 21 11:13:09 EDT 2011
Jean, could you take a look at freeze_interior.F
(it's only 8 lines of code) and suggest how I should proceed.
Should I enclose the whole subroutine between
#ifndef ALLOW_AUTODIFF_TAMC
Will that flag also make the code invisible to OpenAD?
Dimitris
On Thu, 2011-04-21 at 07:17 -0700, Jean-Michel Campin wrote:
> Hi Dimitris and Jean,
>
> Looks like S/R FREEZE_INTERIOR is causing problems for OpenAD.
> The 2 tests from last night show error (at compile time)
> in freeze_interior_cb2m.f90 , don't know why.
>
> I have an other comment for Jean:
> When a test fail to compile, the compile log file
> (make.tr_log) is sent with the testreport output.
> But with OpenAD, this file is quite big (> 180.M) so that
> even when compressed, the email is too big to fit into
> 1 piece and got split into multi-parts.
> I changed the way to retrieve output to get back multi-parts
> messages (and was pleased today to see that it's working)
> but maybe, ome day, could think of ways to reduce the size of this
> compile-log file ?
>
> Cheers,
> Jean-Michel
More information about the MITgcm-devel
mailing list