[torqueusers] C scheduler -> maui help

Jeff Blasius jeff.blasius at gmail.com
Mon Oct 3 08:41:02 MDT 2005


Using this method I believe it's possible the first job can run a load
> 1.0 and prohibit the next job from being run. Is this the case?

Maybe this is better:
 Example 1:  Favor the fastest nodes with the most available memory
which are running the fewest jobs.

NODEALLOCATIONPOLICY PRIORITY
NODECFG[DEFAULT] PRIORITYF='SPEED + .01 * AMEM - 10 * JOBCOUNT'

Maybe this should be discussed on the maui list.

Thanks,
                  jeff

On 10/3/05, Chris Johnson <johnson at nmr.mgh.harvard.edu> wrote:
> On Mon, 3 Oct 2005, Chris Johnson wrote:
>
> > On Sun, 2 Oct 2005, Jeff Blasius wrote:
> >
> > > I'm new here too, so this may not be correct. I believe what you want
> > > can be accomplished by setting
> > > NODEALLOCATIONPOLICY CPULOAD
> > > See:
> > > http://www.clusterresources.com/products/maui/docs/5.2nodeallocation.shtml
> > >
> > >        -jeff
> > >
> > >
> >
> >      Closer.  No cigar yet.  Some jobs seem to get stuck in Q state if
> > too many are queued at once.  For example, I have three nodes,
> > queueing 4 jobs does the right thing in that one cpu from each node
> > was sceduled before the second cpu on any node was scheduled.  However
> > queueing 5 jobs put two in Q state and left them there.
> >
>
>      Closer.  Puting two together gets it almost right.  I'm using,
> over and above the out of the box config,
>
> JOBNODEMATCHPOLICY     EXACTNODE
> NODEALLOCATIONPOLICY   CPULOAD
>
> -------------------------------------------------------------------------------
> Chris Johnson               |Internet: johnson at nmr.mgh.harvard.edu
> Systems Administrator       |Web:      http://www.nmr.mgh.harvard.edu/~johnson
> NMR Center                  |Voice:    617.726.0949
> Mass. General Hospital      |FAX:      617.726.7422
> 149 (2301) 13th Street      |Knowing what thou knowest not
> Charlestown, MA., 02129 USA |is in  a sence omniscience.  Piet Hein
> -------------------------------------------------------------------------------
>
>


--
Jeff Blasius / jeff.blasius at yale.edu
Phone: (203)432-9940  51 Prospect Rm. 011
High Performance Computing (HPC)
UNIX Systems Administrator, WorkStation Support (WSS)
Yale University Information Technology Services (ITS)


More information about the torqueusers mailing list