[MITgcm-devel] [MITgcm-cvs] MITgcm/pkg/seaice CVS Commit

[Jean-Michel Campin]

Hi Martin,

I had a quick look at this, and it's true that it does not affect
machine truncation in the forward part, but in the code that TAF returns,
it's likely that the truncation errors are different (new S/R, less
arrays than before, e.g., DRAGS is gone) and change a little bit (not
much, just getting 12 instead of 13-14 before) the adjoint results.

So, going to update this output (lab_sea/results/output_adm.evp.txt).

Cheers,
Jean-Michel

On Fri, Oct 19, 2012 at 04:05:18PM +0200, Martin Losch wrote:
> Hi Jean-Michel,
> 
> I am sorry, I only checked the lab_sea adjoint on my machine and there I did not observe any changes.What I did is moved some code (computation of DWATN) into a subroutine, that can be called by all three solvers. I thought is was a good idea, maybe not? Would you expect changes in truncation? I wouldn't, but maybe I am wrong.
> 
> Martin
> 
> On Oct 19, 2012, at 3:34 PM, Jean-Michel Campin wrote:
> 
> > Hi Martin,
> > 
> > I found few changes in some seaice experiments results
> > from the testreport output run over night (I don't have all 
> > the output yet):
> > - lab_sea.evp (adjoint, all tested platforms)
> > - lab_sea (tlm)
> > - global_ocean.cs32x15.seaice_dynmix (adjoint, different only on old 
> >      aces cluster)
> > Was this something expected (changes in machine truncation) ?
> > Given that evp is ultra sensitive, it would not be surprising.
> > 
> > I might generate new lab_sea/results/output_adm.evp.txt 
> > since now we only have 12 matching digits on baudelaire with gfortran 
> > (ref. results).
> > 
> > Cheers,
> > Jean-Michel
> > 
> > On Thu, Oct 18, 2012 at 06:06:42AM -0400, Martin Losch wrote:
> >> Update of /u/gcmpack/MITgcm/pkg/seaice
> >> In directory forge:/tmp/cvs-serv5849/pkg/seaice
> >> 
> >> Modified Files:
> >> 	seaice_preconditioner.F seaice_lsr.F seaice_init_varia.F 
> >> 	seaice_evp.F lsr.F dynsolver.F SEAICE.h 
> >> Log Message:
> >>  - simplify (?) ocean-drag coefficent computations by using
> >>    the same routine in all three solvers (saves some code lines)
> >>  - remove global fields DRAGA and DRAGS from C-grid code and
> >>    declare them locally in lsr.F (the only place where they are
> >>    still used)
> >> 
> >> 
> >> _______________________________________________
> >> MITgcm-cvs mailing list
> >> MITgcm-cvs at mitgcm.org
> >> http://mitgcm.org/mailman/listinfo/mitgcm-cvs
> > 
> > _______________________________________________
> > MITgcm-devel mailing list
> > MITgcm-devel at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-devel
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel