[MITgcm-devel] timing test

[Dimitris Menemenlis]

> 2) when I change to Olx=Oly=3 instead of =2, the total time increases
> by more than 50.%, and the parts that take more time
> (up to a doubling) are where 3.D exch are called.
> This slow-down surprises me (has something to do with
> the volume of data that are exchanged ?).

Hi Jean-Michel, I have survived my week with kids without Kate and am 
catching up with e-mail.  Just want to add that there are some 
unconfirmed reports (from Ou Wang and from Heidi) that high-resolution 
cubed-sphere configuration has doubled in wallclock time on Columbia 
compared to old average of 17 hours per year of computation on 216 
processors.  I will be investigating this soon, I hope, along with cause 
of configuration failure for most recent checkout.  But wondering 
whether possible slowdown might be related to pkg/diagnostics memory 
utilization and/or to issues you discuss in your e-mail of 10/15?

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