The good folks at HPC Carpentry have posted a new set of teaching materials designed to help new users take advantage of high-performance computing systems. No prior computational experience is required – these lessons are ideal for either an in-person workshop or independent study.
HPC Carpentry is not an organization – it is merely a set of publicly available teaching materials designed to make the task of teaching HPC a little easier. We welcome all contributions, in particular adaptations of our Intro to HPC lesson for other schedulers besides SLURM.
Each of the lessons are designed to teach how to interact with a compute cluster in a specific way. The first lesson, Intro to HPC, is a general introduction to the bash command line and submitting jobs on a typical HPC cluster. The second two lessons focus on two separate use cases for HPC: running large numbers of compute jobs (often with complex inter-job dependencies), or creating parallel programs that execute across one or more compute nodes. Each lesson takes roughly a full day to teach (or work through independently), and it is intended that the Intro to HPC lesson be taught together with either the Python or Chapel component to form a two-day workshop.
Workshops are intended to be fully interactive and taught in the “Software Carpentry” format: instruction is done live, with learners coding along with workshop instructors. Frequent excercises and test problems have been added to ensure students have a chance to try things on their own. No prior computing experience is required or expected.
Check out a more in-depth overview of HPC lessons, look into hosting a workshop, or consider contributing to the lessons on Github.
