Job Submission Variations

...in plain english...

(note: this page is indexed as "Job Submission Examples" on the main page)

"I want my job to run on any available batch resources (with the understanding that there is a small chance that I may be preempted by the owner of a node):"

 qsub qsub.scr

"I want to submit my job so that it runs on 1 node with all 8 processor cores for 2 hours:"

 qsub -l nodes=1:ppn=8,walltime=2:00:00 qsub.scr

"I want to submit a 2-node job to the 'debug' queue:"

 qsub -q debug -l nodes=2:ppn=8,walltime=1:00:00 qsub.scr

"I want to submit a 2-node interactive job to the 'debug' queue:"

 qsub -I -q debug -l nodes=2:ppn=8,walltime=1:00:00 qsub.scr

"I want to submit a 16-node, 2-core-per-node, 6-hour job to the 'scavenge' queue (with the understanding that it may be preempted by debug jobs [if it ends up running in the debug_nodes partition] or the owners of a node in the batch_nodes partition:"

 qsub -q scavenge -l nodes=16:ppn=2,walltime=6:00:00 qsub.scr

"I want my job to run on my nodes the next time they're available:"

Set your "qos=" for my group (astro, cenpa, int, physics). If I am in physics:

 qsub -l qos=physics qsub.scr

If I am in astro:

 qsub -l qos=astro qsub.scr

"I want my job to run on my nodes right now, even preempting people in my own group:"

Set your "qos=" for your group with "_now" appended. If I am in physics:

 qsub -l qos=physics_now qsub.scr

If I am in astro:

 qsub -l qos=astro_now qsub.scr