[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] 'parallel' universe job submission crashes SCHEDD
- Date: Thu, 30 Dec 2010 12:47:38 -0500
- From: Michael Hanke <michael.hanke@xxxxxxxxx>
- Subject: Re: [Condor-users] 'parallel' universe job submission crashes SCHEDD
On Thu, Dec 30, 2010 at 12:13:01PM -0500, Erik Aronesty wrote:
> > Yes, sometimes we need a multiple of that -- even though it is only a
> > single cpu job. But at the same time we also have multi-threaded
> > applications that are relatively gentle in terms of memory consumption.
> > Having 8GB statically assigned to each slot would be suboptimal in both
> > cases.
> Is there a version/configuration of condor which doesn't static-assign
> memory on a
> per-slot basis? Dedicated?
Somewhat. With partitionable slots a slot can be created dynamically
that matches a job's requirements regarding CPUs and memory -- that
opens the door to have other dynamic slots with different properties
for jobs with other requirements from the remaining resources.