We’re setting up QuNex on AWS ParallelCluster and I was wondering if anyone has done this and has any tips, and more specifically, advice on what machine type to choose. (There are now several hundred choices for EC2 machine type!)
We’ve validated most qunex steps on AWS ParallelCluster, with a few exceptions including steps requiring GPUs. We use
r5a.large (2vCPU, 16GB of RAM) with SMT disabled, but x86 VM types with similar specs should work. We also don’t use parallelization within a node, meaning each node handles one job and one session at a time (i.e.