[torqueusers] RE: launching GM jobs is too slow

Maestas, Christopher Daniel cdmaest at sandia.gov
Thu Nov 10 21:36:37 MST 2005


Quick question ... Why do enable filesync by default if it bites us for
large systems?
It says that it makes things more reliable ... But I don't see the
reasoning.  I would suggest that we disable file system blocking by
default.


-----Original Message-----
From: Garrick Staples [mailto:garrick at usc.edu] 
Sent: Thursday, November 10, 2005 9:28 PM
To: Maestas, Christopher Daniel
Cc: mpiexec at osc.edu; torqueusers at supercluster.org
Subject: Re: launching GM jobs is too slow

On Thu, Nov 10, 2005 at 09:17:42PM -0700, Maestas, Christopher Daniel
alleged:
> Garrick,
> 
> In fixing some scaling issues recently with Pete on ib, we found that 
> changing the following code in the attached torque patch the pbs_mom 
> with launch issues.  I would also suggest testing against the mpiexec 
> in cvs as well.  Pete was going to release a new mpiexec rsn ... :-)

The O_Sync issue is already covered by --disable-filesync, but I will
try CVS right now.

--
Garrick Staples, Linux/HPCC Administrator University of Southern
California



More information about the torqueusers mailing list