[mpich-discuss] Strange behavior of mpd

[Bellanca Gaetano]

Hi Gus and list,

I came back to my problem with mpd and tested my code on a small PC 
cluster of 10PEs.
In this case, the first one with a 'true' cluster (in other cases I 
tested the code on a dual core and a two processors quad/core 
machine), I was unable to run the simulation with a 'big' data 
set  also after a mpd reboot.

Therefore, following the suggestion of Gus, I modified the code 
reducing the message size sending data by 2D slices. This solves my 
problem. Now I'm able to run the simulation with a small number of 
PEs also with a big data set.

Thanks for your help in solving the problem.

Gaetano


----------
Gaetano Bellanca - Department of Engineering - University of Ferrara
Via Saragat, 1 - 44100 - Ferrara - ITALY
Voice (VoIP):  +39 0532 974809     Fax:  +39 0532 974870
mailto:gaetano.bellanca at unife.it

----------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20081010/58572c88/attachment.htm>