This workshop uses AWS ParallelCluster to create a SLURM cluster and goes through how to run different HPC container runtimes/engines.
The inspiration for the workshop comes from a FOSDEM21 talk about the ‘State of the art’ in Containers in HPC. It’s goal is to allow the readers to go through different runtimes and experience them first-hand.
This workshop is for you if you want to experience how containerization in HPC can be implemented on top of a traditional HPC cluster like SLURM - using AWS ParallelCluster.
You should be familiar with HPC setups in general (specifically SLURM), have some knowledge about the linux command line to walk through the example. The workshop iterate through different runtimes and how they are used. A basic understanding of how containerization works is helpful, but not mandatory.
The duration and cost of the workshop depends on how much you want to do. :) Since the cluster is preconfigured you can get started right away. If you constrain yourself to run through what is shown here it should take 2h and cost less than $10.