[MITgcm-devel] pkg/diagnostics

[Andrea Molod]

hi dimitris (and baylor),

maybe i am not getting things exactly, but i think that
the question that dimitris addressed is different from the
one that baylor is addressing.

dimitris - yes, what you are describing is indeed what is
in the code. that you ask for only one level of THETA as
output does not change the fact that space will be allocated
in the model internal array (that numdiags is used for) based
on all the levels that the diagnostic called THETA is defined 
for. it would take a bit of doing in terms of coding in the
package to change this .... if this something that is in your 
way and needs to be addressed soon, maybe it would be best to
define a new single-level diagnostic for this purpose.
(SSTHETA, or something like that). if this is not reasonable
to you, lets hear about it.

and baylor - the idea to increase the 'default' size of
numdiags would be so that you don't end up having to include
DIAGNOSTICS_SIZE.h every time you run an experiment? (ie, use
a more 'default' default size?).

just trying to get things clear....

andrea



On Fri, 16 Sep 2005, Baylor Fox-Kemper wrote:

> Yes,
>    I often have the same problem.  Can't we set numdiags by default to 
> be a bigger value, like 1000?
>    -Baylor
> 
> On Sep 16, 2005, at 3:08 PM, Dimitris Menemenlis wrote:
> 
> > Hi Jean-Michel, I was updating the hires cubed sphere setup with new 
> > pkg/diagnostics capability and noticed that even when dumping a single 
> > layer, SETDIAG requests sufficient room to store the whole 3-D array. 
> > For example the SST diganostic below requires numdiags=50 instead of 
> > numdiags=1.
> >
> >   frequency(1) = 21600,
> >    fields(1,1) = 'THETA   ',
> >    levels(1,1) = 1.,
> >    filename(1) = 'SST',
> >
> > Am I missing something in the data.diagnostics file?  That is, is 
> > there a way to reduce the size of the "large array" to a single level 
> > for this variable?
> >
> > Dimitris
> >
> > -- 
> > Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
> > Jet Propulsion Lab, California Institute of Technology
> > MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
> > tel: 818-354-1656;  fax: 818-393-6720
> > _______________________________________________
> > 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
> 

-- 
----------------------------------------------------------------
Andrea M. Molod, PhD.
Department of Earth, Atmospheric & Planetary Sciences 
MIT
Tel: (617) 253-5458, Email: amolod at mit.edu 
WWW:http://paoc.mit.edu/paoc/people/person.asp?position=Postdoc&who=molod
----------------------------------------------------------------