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

[Martin Losch]

Hi Jean-Michel,

yes, I didn’t think of that, I just wanted to get rid of the many cpp-ifs and the complicated statements before calling seaice_lsr.

So it seems that we need to have to reset SEAICEuseLSR along with SEAICEuseFREEDRIFT like this:

#ifdef SEAICE_ALLOW_FREEDRIFT
       SEAICE_LSR =  .NOT.SEAICEuseFREEDRIFT
#endif /* SEAICE_ALLOW_FREEDRIFT */

Could be done in the appropriate autodiff file. Do you agree? Should I try it?

Martin

> On 22 Jan 2016, at 15:18, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:
> 
> Hi Martin,
> 
> I have the impression that these changes are breaking the AD verification 
> experiment: global_ocean.cs32x15.seaice_dynmix
> 
> On baudelaire, reference compiler+option -devel, from this morning:
> Y Y Y Y 16> 7<16 FAIL  global_ocean.cs32x15.seaice_dynmix
> If I go back and remove these pkg/seaice changes from yesterday,
> but keep the updated (from yesterday also) output_adm.seaice_dynmix.txt, I get:
> Y Y Y Y 16>16<16 pass  global_ocean.cs32x15.seaice_dynmix
> 
> Now, since this test is using a "funny" set/re-setting/re-re-setting of
> switches: in global_ocean.cs32x15/input_ad.seaice_dynmix/data.autodiff
>> SEAICEuseFREEDRIFTswitchInAd = .TRUE.,
> could this interfere with the new logical "SEAICEuseLSR" ?
> 
> Cheers,
> Jean-Michel
> 
> On Thu, Jan 21, 2016 at 11:31:05AM -0500, Martin Losch wrote:
>> Update of /u/gcmpack/MITgcm/pkg/seaice
>> In directory forge:/tmp/cvs-serv7057/pkg/seaice
>> 
>> Modified Files:
>> 	SEAICE_PARAMS.h seaice_dynsolver.F seaice_readparms.F 
>> 	seaice_summary.F 
>> Log Message:
>> add two new parameters
>>  - SEAICEuseLSR to simplify the logic in seaice_dynsolver (this 
>>    parameter is not in any namelist)
>>  - SEAICEusrPicardAsPrecon to allow s/r seaice_lsr to be used as 
>>    a preconditioner for the non-linear Newton iteration of the JFNK 
>>    solver
>> 
>> 
>> _______________________________________________
>> 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