[torqueusers] simulating worst case behavior from my cluster. request 8 cpus such that no two are on the same node.

Victor Gregorio vgregorio at penguincomputing.com
Mon Apr 20 14:53:23 MDT 2009


Hello Rahul,

I believe you can do one of the following...

1) Request 8 nodes at 1 processor each:
   #PBS -l nodes=8:ppn=1

2) Request each node by hostname and use 1 processor each:
   #PBS -l nodes=node1:ppn=1+node2:ppn=1+node3:ppn=1 ...

-- 
Victor Gregorio
Penguin Computing

On Mon, Apr 20, 2009 at 03:46:34PM -0500, Rahul Nabar wrote:
> I am trying to simulate worst-case behavior of a job on our cluster
> where one job would get 8 cpus but one each from a different
> compute-server. Each server has 8cpus.
> 
> How can I do that? Server names: noce01 through node23. Scheduler PBS-Torque
> Super-scheduler MAUI
> 
> Normally I request using:
> #PBS -l nodes=8          #####any 8 cpus
> 
> #PBS -l nodes=1:ppn=8 ########## 8 cpus on a single server
> 
> #PBS -l nodes=node17:ppn=8########8 cpus on a specific server node17
> 
> This time I want one each from any 8 servers from the cluster. Without
> blocking the remaining 7 cpus on that node. The nodes are indeed set
> up as shared by default.  Any way to do this?
> 
> -- 
> Rahul
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers


More information about the torqueusers mailing list