Titan Batch Queues
Categories: Running Jobs
Print this article
Queues are used by the batch scheduler to aid in the organization of jobs. Users typically have access to multiple queues, and each queue may allow different job limits and have different priorities. Unless otherwise notified, users have access to the following queues on Titan:
||No explicit request required||Default; most production work runs in this queue.||See the Titan Scheduling Policy for details.|
||Opportunistic; jobs start even if they will not complete before the onset of a scheduled outage.|
||Quick-turnaround; short jobs for software development, testing, and debugging.|
batch queue is the default queue for production work on Titan. Most work on Titan is handled through this queue.
At the start of a scheduled system outage, a queue reservation is used to ensure that no jobs are running. In the
batch queue, the scheduler will not start a job if it expects that the job would not complete (based on the job’s user-specified max walltime) before the reservation’s start time. In contrast, the
killable queue allows the scheduler to start a job even if it will not complete before a scheduled reservation.
debug queue is intended to provide faster turnaround times for the code development, testing, and debugging cycle. For example, interactive parallel work is an ideal use for the
debugqueue may have further access to the queue denied.
More detailed information on any of the batch scheduler queues can be found on the Titan Scheduling Policy page.