In Unix, how should I submit CPU-intensive jobs?

Several mechanisms are available for handling large jobs in Unix, including reducing the priority of a process, running a process in the background, and, on some systems, submitting batch jobs.

On this page:

• Batch jobs
• Changing a job's priority and running a job in the background

Batch jobs

For more information about the at and batch commands, see In Unix, what are at and batch, and how do I use them to submit non-interactive job requests?

For information about using LoadLeveler on the Libra Cluster at Indiana University, see Using LoadLeveler on Libra at IU.

For information about using LoadLeveler on Big Red at IU, see Getting started on Big Red.

For information about using PBS on Quarry, see:

Changing a job's priority and running a job in the background

To immediately execute a process at a lower priority, use the nice command. For example, to run a program called a.out at lower priority, at the Unix shell prompt, enter:

To let the job run in the background (so that you can keep working interactively), add  &  (an ampersand) to the end of the command line:

When the job is complete, you will see something like the following:

For more information on the nice command, see the manual (man) page, which you can access by entering:

For more information on how to work with background processes, see the man page for your shell (e.g., csh, tcsh, ksh, bash).

At Indiana University, to get support for personal or departmental Linux or Unix systems, see At IU, how do I get support for Linux or Unix?

Also see:

• In Unix, how do I cancel a batch job?
• How do I run stat/math jobs, including SAS and SPSS, under AIX LoadLeveler?
• On the Libra Cluster, how do I submit a Matlab batch job?
• In Unix, how do I avoid stty errors when I submit batch jobs?
• In Unix, what are at and batch, and how do I use them to submit non-interactive job requests?
• Introduction to Unix commands