[Condor-users] out-of-memory issues in parallel universe

["Robert E. Parrott" <parrott@xxxxxxxxxxxxxxxx>]

Hi folks,

We have some macih es with 4 Gb RAM, and some with 6 Gb RAM all in the  
same cluster, and may be seeing some over-extension of memory for MPI  
jobs that are asssuming that there's 6 Gb of RAM available.

Is there some way of specifying the image size, and restricting jobs  
to larger memory compute nodes, for MPI jobs submitted in the parallel  
universe?

Also, what's the recommended way to enforce memory constraints on  
jobs, in general and in particular in the parallel universe? ulimit?  
something else?

thanks,
rob


==========================
Robert E. Parrott, Ph.D. (Phys. '06)
Associate Director, Grid and
       Supercomputing Platforms
Project Manager, CrimsonGrid Initiative
Harvard University Sch. of Eng. and App. Sci.
Maxwell-Dworkin  211,
33 Oxford St.
Cambridge, MA 02138
(617)-495-5045