Re: [Condor-users] Builing GPU Cluster Using Condor

[Rob Rati <rrati@xxxxxxxxxx>]

Based upon the README I browsed through, that should be run on nodes
that will be running jobs (ones running a condor_startd).  It looks like
it queries the machine and adds attributes to the machine's ad that is
advertised on the collector.  For how to configure via
condor_config.local (as in the README), take a look at:

http://www.cs.wisc.edu/condor/manual/v7.6/3_3Configuration.html

Newer versions of condor are moving to using LOCAL_CONFIG_DIR though,
and I'd recommend doing that instead.

In that same section, lookup STARTD_CRON and read the associated text to
understand what is being done.

Rob

On Thu, 2011-05-26 at 09:07 +0700, Lucky wrote:
> Hi,
> 
> I am very new about condor. Recently, Condor using CPU as processing the 
> job. How about using GPU on every computer node?
> 
> Has anyone try to build GPU Cluster?
> Is there any clue how to build it?
> I have been using this script
> http://sourceforge.net/projects/condorgpu/
> 
> And I don't know how to deploy it. Is that script is running on computer 
> node not on condor pool? Thank you for your advice. Looking forward for 
> more about this project.
> 
> -Luky-
> _______________________________________________
> Condor-users mailing list
> To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> 
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/condor-users/