Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] Factor of 10 loss in performance with 1.3.x
From: Ralph Castain (rhc_at_[hidden])
Date: 2009-04-07 16:23:45


It isn't in a file - unless you specify it, OMPI will set it
automatically based on the number of procs on the node vs. what OMPI
thinks are the number of available processors. The question is: why
does OMPI not correctly know the number of processors on your machine?

I don't remember now, but did you send us this info:

   http://www.open-mpi.org/community/help/

If not, could you please do so? It would help us track down the problem.

Thanks
Ralph

On Apr 7, 2009, at 2:16 PM, Steve Kargl wrote:

> On Tue, Apr 07, 2009 at 01:40:13PM -0600, Ralph Castain wrote:
>> I would really suggest looking at George's note first as I think you
>> are chasing your tail here. It sounds like the most likely problem is
>> that OMPI thinks you are oversubscribed and is setting sched_yield
>> accordingly. which would fully account for these diffs.
>>
>> Note that the methods for setting sched yield are different between
>> the various releases, so something could be catching you here. If you
>> are sure you aren't oversubscribing, then set -mca
>> mpi_yield_when_idle
>> 0 on your cmd line to override any internal settings.
>>
>
> '-mca mpi_yield_when_idle 0' returns the performance of 1.3.1 to
> that of 1.2.9. In fact, 1.3.1 may be marginally better. A comparison
> can be see at
>
> http://troutmask.apl.washington.edu/~kargl/ompi_cmp_yield.jpg
>
> So, in what file is sched_yield set?
>
> --
> Steve
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users