The CPU time limits are given relative to a CPU hour on a single processor on Tempus. Since the Alpha cluster CPU's are about four times slower than a Tempus CPU (actual figures may vary), the CPU limit in the Alpha cluster is really about four times longer than these figures.

The long-tempus, 3day-tempus and week-tempus queues are intended for general batch jobs on Tempus requiring a single processor.

The wide-tempus queue is intended for users with multiprocessor applications. The priority-tempus queue is for users who require realtime access to Tempus. Jobs submitted to this queue will be quickly dispatched on Tempus and other running jobs may be suspended. For permission to use either of these queues, send email to HPC-support

There are several restrictions placed on the queues that may affect job scheduling. For example, there is a cluster wide limitation of two jobs per user at any one time, so if you already have two jobs running anywhere on the cluster and your job is next in line, it will be bypassed by a later job unless one of your running jobs finishes first. To find out why your job is pending, use the bjobs -l command.

These configurations are subject to change. Use bqueues -l <queue> for the current configuration information on a particular queue.