Webinar 2016 What Happened to My Job? Cluster Scheduling In Detail

From SHARCNETHelp
Jump to navigationJump to search

The scheduler of an HPC cluster accepts job submissions, which are later given resources on compute nodes and allowed to run. Waiting for your job to run can be annoying, and when it eventually runs, there may be problems. This seminar will examine what the scheduler is doing with your job, how it decides when and where to run it, and how you can minimize scheduler-related frustration.