[Pw_forum] MPI & disk unavailability

[Konstantin Kudin]

 Axel, thanks!

 As far as the NFS is concerned, we have quite a heavy load on the
system. I've seen a few jobs that terminated prematurely for no reason
with the MPI errors, using the code from the CVS from mid-August '05.

 I have not seen the same thing from a very recent version, so I'd like
to think that if there was a problem, it was somehow fixed in the
meantime.

 My question was caused by the head node being unavailable for more
than an hour, so probably expecting MPI to behave nicely with such a
long outage is not really reasonable.

 Kostya 


--- Axel Kohlmeyer <akohlmey at vitae.cmm.upenn.edu> wrote:

> this handled by the MPI _implementation_. you may want to try a 
> different package (there are several others besides MPICH which is 
> IMNSHO somewhat clunky , see e.g.: 
> http://www.lam-mpi.org/mpi/implementations/shortlist.php ),
> or hack your local installation to increase the delay before 
> there is a timeout and the library considers the cpu0 process
> as dead.



	
		
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com