[MITgcm-devel] [MITgcm-cvs] MITgcm/model/src CVS Commit

[Jean-Michel Campin]

Hi Martin,

You can ignore this message: There was something broken
with staf connection (but not in the usual way,
I was still getting some output "ad_input_code_ad.f"
but this one was wrong/incomplete). And because of this connection
problem, it slows down all the tests of last night and 
did not get the other in time. I also had some AD tests on
the old aces cluster that suffer from the same problem.
Now it looks better (gfortran ad test on baudelaire seems OK, 
execpt the thing I pointed to in the other email).

Cheers,
Jean-Michel

On Tue, Dec 20, 2011 at 04:08:37PM +0100, Martin Losch wrote:
> Hi Jean-Michel,
> 
> I admit that I did not check, because there where no store directives associated with this call, but I cannot reproduce the problem on my machine (runs fine without problems). It probably is specific to baudelaire? The error message does not look like a fortran problem (in <http://mitgcm.org/testing/results/2011_12/tr_baudelaire_20111220_0/global_ocean.90x40x15/make.tr_log_tail>):
> >  Transformation of Algorithms in Fortran (TAF)
> >  Copyright 2000-2010 FastOpt GmbH, Hamburg, Germany
> >  All rights reserved.
> >  URL: http://www.FastOpt.de, Email: info at FastOpt.de
> > 
> >  Processing files at fastopt.de, please wait.
> > 
> > Write failed: Broken pipe
> > 
> > gzip: stdin: unexpected end of file
> > tar: This does not look like a tar archive
> > tar: Exiting with failure status due to previous errors
> > cat: taf_output: No such file or directory
> > ls -l ad_input_code_ad.f
> > ls: cannot access ad_input_code_ad.f: No such file or directory
> > make: *** [ad_taf_output.f] Error 2
> 
> 
> 
> I will repair the data.seaice issue for the adjoint.
> 
> Martin
> 
> On Dec 20, 2011, at 3:57 PM, Jean-Michel Campin wrote:
> 
> > Hi Martin,
> > 
> > Did you check how this works with the adjoint ?
> > I am getting many AD tests that fail to compile (on baudelaire):
> > 
> > Y Y N N .. .. .. N/O   1D_ocean_ice_column
> > Y Y N N .. .. .. N/O   bottom_ctrl_5x5
> > Y Y N N .. .. .. N/O   global_ocean.90x40x15
> > Y Y N N .. .. .. N/O   global_ocean.cs32x15
> > Y Y N N .. .. .. N/O   global_ocean.cs32x15.seaice
> > Y Y N N .. .. .. N/O   global_ocean.cs32x15.seaice_dynmix
> > Y Y N N .. .. .. N/O   global_ocean.cs32x15.thsice
> > Y Y N N .. .. .. N/O   global_ocean_ebm
> > 
> > (I need to look to the details of this)
> > And the other are very-very slow to run (some store dir to
> > adjust ?).
> > 
> > Can you have a look at this ?
> > 
> > Thanks,
> > Jean-Michel
> > 
> > On Mon, Dec 19, 2011 at 06:47:04AM -0500, Martin Losch wrote:
> >> Update of /u/gcmpack/MITgcm/model/src
> >> In directory forge:/tmp/cvs-serv17078/model/src
> >> 
> >> Modified Files:
> >> 	do_oceanic_phys.F calc_ivdc.F 
> >> Log Message:
> >> use sigmaR in calc_ivdc rather than rhokp1 and rhokm1
> >> 
> >> 
> >> _______________________________________________
> >> 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