[MITgcm-devel] following on verif_basic

[Martin Losch]

Hi,

I think that Jean-Michel's selection is good, I would not reduce the  
number of experiments any further. Reducing global_ocean.90x40x15 to  
under 10MB would be nice, though ...

Martin

Martin
On Jul 9, 2006, at 6:41 PM, chris hill wrote:

> > to (what I propose):
> > 13504   aim.5l_cs/
> > 21460   global_ocean.90x40x15/
> > 660     hs94.128x64x5/
> > 528     ideal_2D_oce/
> > 6664    lab_sea/
> > 820     rotating_tank/
> > 720     tutorial_plume_on_slope/
> > 44392   .
> >
>
> I would rather
>
> > ?????   global_ocean.90x40x15/
> > 660     hs94.128x64x5/
> > 528     ideal_2D_oce/
> > 6664    lab_sea/
> > 820     rotating_tank/
> > 720     tutorial_plume_on_slope/
> > 332     tutorial_baroclinic_gyre/
>
> ????? above is because I think we can make global_ocean/ smaller.  
> For example without qshyd and compressing the large .txt results sets.
>
> I don't think having aim_5l/ in the basic tests is needed as long  
> as we make sure it stays on the main testing we do.
>
> tutorial_baroclinic_gyre/ is a good very basic test.
>
> Chris
>
> Jean-Michel Campin wrote:
>> Hi Martin (& Ed),
>> Ed proposed to remove the 2 CVS modules:
>>  MITgcm_verif_atmos
>>  MITgcm_verif_ocean
>> And to keep an updated version of:
>>  MITgcm_verif_basic
>> If we agree on a small number of small (let's say, below 10.M)
>> experiments in verif_basic (that will always remain independent),
>> the way we use "prepare_run" is no longer an issue.
>> Regarding the modification of MITgcm_verif_basic, the current list  
>> of experiments is:
>> global_ocean.90x40x15
>>   <- keep it for now, until we switch to tutorial_global_oce_latlon
>>      (without the input.qshyd and result_ad*, size will go down,  
>> below 10.M)
>> aim.5l_cs   <- too big ?
>> hs94.128x64x5 <- ?
>>       I would keep those 2 atmospheric set-up (one on Lat-Long,
>>       the other on CS grid), since aim.5l_cs is testing plenty of AIM
>>         / Land / thsice stuff.
>>       alternatively, we can put insead aim.5l_LatLon &  
>> tutorial_held_suarez_cs,
>>       but these 2 together take more space than aim.5l_cs +  
>> hs94.128x64x5);
>> front_relax
>>  <- quiet simple, would prefer ideal_2D_oce (test diagnostics,
>>     mnc, all gmredi options not already tested in global_ocean. 
>> 90x40x15)
>> tutorial_barotropic_gyre
>>   <- so simple that I am ready to remove it
>> tutorial_plume_on_slope
>>   <- keep it (test NH, variable dX, AdvScheme=33)
>> tutorial_global_oce_biogeo
>>   <- far too big
>> I would add 2 small tests (removing tutorial_barotropic_gyre &
>>  tutorial_global_oce_biogeo, the total number would be unchanged).
>> lab_sea
>>   <- test sea-ice, exf
>> rotating_tank
>>   <- test cylindrical grid & NH
>> Now, in term of disk space (& data to download),
>> we will go from (currently):
>> 13504   aim.5l_cs/
>> 492     front_relax/
>> 21460   global_ocean.90x40x15/
>> 660     hs94.128x64x5/
>> 332     tutorial_barotropic_gyre/
>> 26244   tutorial_global_oce_biogeo/
>> 720     tutorial_plume_on_slope/
>> 63468   .
>> to (what I propose):
>> 13504   aim.5l_cs/
>> 21460   global_ocean.90x40x15/
>> 660     hs94.128x64x5/
>> 528     ideal_2D_oce/
>> 6664    lab_sea/
>> 820     rotating_tank/
>> 720     tutorial_plume_on_slope/
>> 44392   .
>> What do you (and anyone who use those CVS alias) think of this ?
>> Cheers,
>> Jean-Michel
>> _______________________________________________
>> 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