[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/pkg/diagnostics CVS Commit

[Ed Hill]

On Sun, 2005-02-20 at 10:14 -0500, Andrea Molod wrote:
> hi there ed,
> can i ask about this?
> are we talking about the model state pickups? are does this
> mean a 'diagnostic pickup' of sorts? if it is the normal state
> pickup, is the diagnostic package the place to have an option
> to write model pickups?
> thanks,
> andrea


Hi Andrea & Jean-Michel,

I meant to send an email to the devel list about these changes but was
tired late last night...

Anyway, the diagnostics-pickup is there to save the "qdiag" array across
re-starts.  The idea is that a re-start won't truncate diags in
progress.  For instance, Dimitris and I are trying to do monthly
averages on his 510x510x6x50 cube setup and, if a run gets interrupted
sometime during a month, then the monthly diagnostics average is wrong
because the partial sums within qdiag aren't preserved.

So I have two quick questions:

 1) Is saving the qdiag array sufficient for all cases?  Or 
    do I need to save other items as well?  I looked in 
    GETDIAG() and it seems that it will work for both 
    "counter" and "non-counter" diagnostics -- but am I 
    overlooking anything?

 2) Can you think of ways to make the diagnostics pickups
    either: (a) more efficient and/or (b) have them save the 
    state by default?  For the time being, I've put the 
    routines inside IF's with off-by-default flags so they 
    won't disrupt anything.

Ed

-- 
Edward H. Hill III, PhD
office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
             Cambridge, MA 02139-4307
emails:  eh3 at mit.edu                ed at eh3.com
URLs:    http://web.mit.edu/eh3/    http://eh3.com/
phone:   617-253-0098
fax:     617-253-4464