[MITgcm-devel] [MITgcm-cvs] MITgcm/doc CVS Commit

[Jean-Michel Campin]

Hi Dimitris,

I am curious why we have to add CPP options (11 more) instead of
run-time parameters, which is generally a better solution.
Is it motivated by Adjoint issues ? (seems to me that, regarding adjoint,
 skipping some parts should work the same way as when every is used).

The other thing is regarding names: 
 #define ALLOW_OBCS_U_AT_NS
 #define ALLOW_OBCS_U_AT_EW
 #define ALLOW_OBCS_V_AT_NS
 #define ALLOW_OBCS_V_AT_EW
and
 #undef USE_OBCS_LINEAR
I am concerned that OBCS_SPONGE options that are named without the word 
"SPONGE" could be confusing (interpreted as a general OBCS options instead
of sponge only).

Cheers,
Jean-Michel

On Thu, Sep 25, 2014 at 08:06:03AM -0700, Dimitris Menemenlis wrote:
> JM, thanks for heads up.
> I had run all the forward tests that use obcs before committing but neglected to run the adjoint tests.
> I am in meeting in the AM but will work and fix this problem this evening.
> 
> Dimitris Menemenlis
> 
> On Sep 25, 2014, at 8:00 AM, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:
> 
> > Hi Dimitris,
> > 
> > All the Adjoint testreport that have been run last night
> > show a FAIL for experimeobcs_ctrl:
> > Y Y Y Y  2> 0< 0 FAIL  obcs_ctrl  (e=0, w=0)
> > as opposed to before:
> > Y Y Y Y 16>16<16 pass  obcs_ctrl  (e=0, w=0)
> > Did you check what you changes did to this experirment ?
> > 
> > Cheers,
> > Jean-Michel
> > 
> > On Wed, Sep 24, 2014 at 08:40:49PM -0400, Dimitris Menemenlis wrote:
> >> Update of /u/gcmpack/MITgcm/doc
> >> In directory forge:/tmp/cvs-serv5274
> >> 
> >> Modified Files:
> >> 	tag-index 
> >> Log Message:
> >> adding compile-time options for sponge boundary layers
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel