[MITgcm-devel] just checking: default code fails for verif exp ctrl_obcs

[gael forget]

Hi An, Jean-Michel,
I had forgotten to checkin a file revision in obcs_ctrl/input_ad. 
I did it this morning, which should resolve the FAIL from obcs_ctrl.
The warnings are benign I think but I will double check.
Cheers,
Gael

On Oct 25, 2015, at 10:59 AM, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:

> Hi An,
> 
> I am cc this to MITgcm-devel. 
> 
> I think all experiments were OK yesterday (Sat) early morning,
> given that the nightly testreport output were all good.
> 
> I suspect that some of Gael's changes later yesterday have
> changed the situtation, and ended the testreport from last night (early Sun)
> shows that obcs_ctrl has all zero digit of matching.
> for e.g., ref compiler (gfortran) on ref platform (baudelaire):
> http://mitgcm.org/testing/results/2015_10/tr_baudelaire-b_20151025_0/summary.txt
> 
> Also it seems that we have 1 or 2  more TAF recomp. warning in few experiements
> (1D_ocean_ice_column, lab_sea, obcs_ctrl).
> 
> Cheers,
> Jean-Michel
> 
> On Sun, Oct 25, 2015 at 12:04:04AM -0400, An Nguyen wrote:
>> hi Jean-Michel,
>> 
>> I was about to check in my code until i found that with the default MITgcm code the adjoint verif exp ctrl_obcs is failing.
>> 
>> Just making a note here in case you don't see the same failure.. i did cvs -q up -A to get latest, then ran the test below.
>> 
>> cheers,
>> 
>> ------
>> Sat Oct 24 23:45:47 EDT 2015
>> run: testreport -ad -t obcs_ctrl
>> on : Linux baudelaire 2.6.34.7-61.fc13.x86_64 #1 SMP Tue Oct 19 04:06:30 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux
>> 
>> No "OPTFILE" was specified ; genmake2 found and uses:
>>                         OPTFILE=../../../tools/build_options/linux_amd64_gfortran
>> 
>> Adjoint generated by TAF
>> 
>> default    10
>> G D M    C  A  F
>> e p a R  o  d  D
>> n n k u  s  G  G
>> 2 d e n  t  r  r
>> 
>> Y Y Y Y  0> 0< 0 FAIL  obcs_ctrl  (e=0, w=1)
>> Start time:  Sat Oct 24 23:45:47 EDT 2015
>> End time:    Sat Oct 24 23:49:05 EDT 2015
>> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel