Partition == Queue

SLURM allows the definition of different partitions. Think of these as job queues.

Our cluster has 4 different partitions:

• all
  • The default queue. Contains the 17 standard compute nodes.
• gpu
  • Contains the GPU machine only.
• bigmem
  • Contains the big memory machine only.
• short
  • Contains all nodes, including gpu and bigmem, but has a 2 hour time limit, and jobs in this queue have a lower priority than jobs in the other queues.

All, gpu, and bigmem have no time limit.

Note: right now there is an extra partition, wp, which was created just to ensure that some nodes are available for the workshop.