Dear All,

I have two types users in my Cluster:

1. Running serial jobs (single core).

2. Running parallel jobs (multi core).

Users running parallel jobs requires the use of an entire machine (8-12
cores).

I would like to configure the queue such that once a parallel job is
submitted a machine would not accept any more sequential jobs such that the
parallel job can start.

the situation now is that such jobs wait a very long time till a specific
machine is completely empty.

Thanks

Snir
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers

Reply via email to