The majority of computational work on the Digital Research Alliance of Canada (DRAC) compute clusters is achieved by submitting jobs to the Slurm scheduling software. The scheduling software is what researchers use to access the vast and diverse compute nodes made available by DRAC. The scheduling software is also responsible for how fairness of the access to resources is achieved on these clusters that are freely available to all academic researchers in Canada. This course introduces the role of the scheduler on compute clusters before covering topics on its use. Beyond the general content of using a Slurm scheduler this course covers topics that are specific to the DRAC scheduling environment configuration and offers strategies for getting the most out of the national clusters.
Prerequisite: no prerequisites but an understanding of the "New User" and "Introduction to Shell" Self-Paced Courses is helpful.
Estimated time: 1-4 hours depending on time spent exploring the examples.
- Teacher: James Desjardins