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

[Jean-Michel Campin]

Hi Gael,

I did check that this modification did not impact the current verification 
experiments as well as global_oce_cs32 (in contrib verification_other).
However, it does affect global_oce_llc90 (at least the forward results, did
not try ADM) and it's due different machine truncation in computing the 
N.pole average value.

New output (on baudelaire.csail, gfortran -ieee, + mpich2 on 6 procs):
Y Y Y Y>14<13 16 16 16 16 16 16 16 14 13 13 13 13 14 16 16 pass  global_oce_cs32
Y Y Y Y>10<16 16 16 16 16 16 16 16 14 14 14 16 14 14 13 16 pass  global_oce_llc90
Y Y Y Y> 9<16 16 12 12 16 16 13 11 13 13 11 11 13 12 11 11 FAIL  global_oce_llc90.core2
Y Y Y Y> 3<12 13 10 10 11 16 11  9  9 10  9  8 11 11  8  9 FAIL  global_oce_llc90.ecco_v4
Y Y Y Y> 6<16 13 12 12 12 13 13 10 10 12 10 11 11 11 10 11 FAIL  global_oce_llc90.ecmwf

Repeating same test with previous code:
Y Y Y Y>14<13 16 16 16 16 16 16 16 14 13 13 13 13 14 16 13 pass  global_oce_cs32
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass  global_oce_llc90
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass  global_oce_llc90.core2
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass  global_oce_llc90.ecco_v4
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass  global_oce_llc90.ecmwf

And with a different compiler on different platform (before code changes):
Y Y Y Y>14<13 16 16 16 16 16 16 16 14 13 13 16 13 14 16 16 pass  global_oce_cs32
Y Y Y Y> 6<16 16 12 12 16 16 13 11 13 12 11 12 12 13 11 11 FAIL  global_oce_llc90
Y Y Y Y> 9<16 16 12 12 16 16 14 11 13 13 13 13 13 14 12 13 FAIL  global_oce_llc90.core2
Y Y Y Y> 3<12 13 10 10 11 16 11  9  8 10  8  9 10 11  9  9 FAIL  global_oce_llc90.ecco_v4
Y Y Y Y> 6<13 13 11 11 12 13 13 10 10 11 10 11 11 11 10 11 FAIL  global_oce_llc90.ecmwf

Likely that we will have to update few output*.txt there.

Cheers,
Jean-Michel

On Wed, Jun 14, 2017 at 09:03:55AM -0400, Jean-Michel Campin wrote:
> Update of /u/gcmpack/MITgcm/pkg/exf
> In directory forge:/tmp/cvs-serv6822
> 
> Modified Files:
> 	exf_interp_uv.F 
> Log Message:
> - fix bug in calculation of South-Pole mean value.
> 
> 
> _______________________________________________
> MITgcm-cvs mailing list
> MITgcm-cvs at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-cvs