This session is designed for participants familiar with Linux, Shell Scripting and running simple jobs on a cluster (attending the earlier sessions HPC: Shell and HPC: Essentials will suffice).
 
We will cover the following topics:

-how the scheduler works on Compute Canada systems and how it picks (allocates) which jobs run next;
-why your jobs do not run, either at all or quickly, and what can be done about it;
-how to use resources more efficiently to maximize throughput; and
-how to troubleshoot and deal with crashes.

We will also touch on some specialized topics, such as running a large number of jobs, running interactive jobs, long-running jobs, large memory use, and Graphics Processing Units (GPUs).

As we will be doing hands-on work, participants are strongly encouraged to bring a laptop to submit jobs and examine the cluster. We recommend having a Compute Canada account in order to inspect a decently sized working cluster, but this is not a requirement.

If you are using Windows and do not already have a Shell installed (Linux and MacOS come with them built in), the option we recommend is to install the Git Shell from https://git-scm.com/download/win but you are welcome to use any option that will give you the ability to open a secure shell (SSH) to a remote system and also emulate a shell on in your desktop environment.