[MITgcm-support] MPI rules to save compute time?

Gustavo Correa gus at ldeo.columbia.edu
Mon Nov 14 12:18:26 EST 2011


Hi Chun-Yan

Any chance that in run1 used two separate computers, one processor on each [say, in a cluster]?
If so, communication across the network, even a fast network, is slower than in 
a single computer, which might contribute to a larger execution time than you expected.

Also, for all runs, is it on the same type of computer, same processors, 
same amount of memory? 
Was run1 sharing or not sharing resources with other jobs? 
Or perhaps there was somebody running Matlab, or viewing streaming video while the model ran?

If you are using a single computer, and mostly running only the model, 
than the rationale has to be different, maybe the cache
evictions that Dimitris mentioned.

Gus Correa

On Nov 14, 2011, at 11:10 AM, Menemenlis, Dimitris (3248) wrote:

> Consider number of grid points per process, "including" overlap regions.
> The narrower "run1" has larger overlap regions, hence more computations
> per process, possibly impacting processor cache too.
> 
> Dimitris Menemenlis
> 
> On Nov 14, 2011, at 7:47 AM, Chun-Yan Zhou wrote:
> 
>> Dear all,
>>  I use verification/exp4 to test the MPI.
>>  The the Nx=80, Ny=42 and Nr=8.
>>  I use differnt nPx and Npy to run the same iteration number and the data files are same.
>>  &PARM04
>>  usingCartesianGrid=.TRUE.,
>>  delX=80*5.e3,
>>  delY=42*5.e3,
>>  delR= 8*562.5,
>>  &
>>            sNx   sNy  nPx nPy processes    time cost
>> run0    80      42   1       1          1             20mins
>> run1    80      21   1       2          2             20mins
>> run2     40     42   2       1          2             12mins
>> run3     40      21  2        2          4            10mins
>> 
>> comparing run0 and run1,  nPy=2 didn't speed up the computation. But nPx=2 case is much faster than NPx=1(run0 VS. run2). Since the grid spaces are the same in the X and Y direction, what causes the compute time difference?  Any idea about this? Any rules to assign the nPx and nPy values in order to save time?
>> 
>> Thanks in advance!
>> chunyan
>> 
>> The University of Dundee is a registered Scottish charity, No: SC015096
>> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support




More information about the MITgcm-support mailing list