[MITgcm-devel] slow lab_sea adjoint

[Martin Losch]

Patrick,

I am sorry about the exf_getyearlyfieldname and not updating  
exf_ad_diff.list
Apparently I still don't understand the logic of the surface flux  
control variable and why details of the calendar and reading part of  
the exf package actually affect the adjoint, when the control  
variables are really the deviations from the forcing fields.

Martin

On 26 Jan 2008, at 07:10, Patrick Heimbach wrote:

> On Jan 25, 2008, at 8:54 PM, Dimitris Menemenlis wrote:
>
>> The lab_sea adjoint is slow again.
>>
>> I think the problem is with this morning's changes to exf_set_uv.F  
>> and exf_set_gen.F and with the addition of  
>> exf_getyearlyfieldname.F, not with the modifications to  
>> exf_getffieldrec.F.
>
> Should be fixed.
>
>> Using older versions of the exf_set routines but new versions for  
>> everything else, the lab_sea adjoint test completes OK.
>>
>> Patrick, how can we go further than above in helping you to  
>> diagnose adjoint problems as we change the forward code?
>
> We could have testreport grep for a few more specific error  
> messages in tad_ad.log.
>
> -p.
>
>> D.
>>
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>
> ---
> Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
> MIT | EAPS 54-1518 | 77 Massachusetts Ave | Cambridge MA 02139 USA
> FON +1-617-253-5259 | FAX +1-617-253-4464 | SKYPE patrick.heimbach
>
>
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel