[MITgcm-devel] gad_som_advect etc

[Jean-Michel Campin]

Hi Martin,

Just a guess here: There is a case where SOM could be used for (some) ptracers
but not for T or S. In this case, the present code allows to compile with
#undef GAD_ALLOW_TS_SOM_ADV while using gad_som_adv* routines for ptracers
advection.

Cheers,
Jean-Michel

On Mon, Jun 13, 2016 at 10:41:20AM +0200, Martin Losch wrote:
> Hi Jean-Michel,
> while search for the reason why the cray-compiler tests on ollie keep failing while compiling seaice_obcs (actually while compiling gad_som_advect.f), I noticed that the CPP flag GAD_ALLOW_TS_SOM_ADV is not used in routines like gad_som_advect.F, gad_som_adv_x.F (was removed 8 years ago with this statement: "change option: GAD_ALLOW_SOM_ADVECT to GAD_ALLOW_TS_SOM_ADV which only applies to files where Temperature & salinity 2nd Order moments are used???). As a consequence, we have these fields smTp(snx,sny,nz,nsx,nsy,nSOM) and similar, that the flag was supposed to avoid (memory). I am not sure if this is intended. I am also not sure if this has anything to do with my problem on ollie, but maybe I am using too much memory while compiling large fields, just a guess ???).
> 
> Do you see a way that we can avoid these large fields when GAD_ALLOW_TS_SOM_ADV is undefined? Or is this a useless suggestion?
> 
> Martin
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel