Scheduler Selection Beta Now Available
Choosing a HPC/AI scheduler is often a once per career type decision and getting it wrong can be expensive. Want a hand?
Nivida acquiring SchedMD is old news now, but maybe you’re still spooked by it and want to explore your options. Selecting an HPC scheduler is something most people never have to do (you use what you’re given) and even those of us that do have to make that decision, will generally only make it once or twice in our careers.
It’s also a decision that has a massive impact. Not only on how well you manage to utilise your expensive supercomputers (everyone’s aiming for 100% utilisation right? 😆) but it also determines what your users can and can’t achieve. Whether that’s science, finance, or AI.
Let’s be honest, most of us probably don’t even know what the options to pick from are!
Want a hand? We’ve worked with the vendors of many of the most popular schedulers to understand their feature sets and come up with a way to help you narrow down the choice based on answers to some questions.
We first started working on this early last year but we’ve been a little slow in getting a beta in front of you all I’m afraid.
You can use the tool here. If we’ve missed a scheduler or made a mistake, it’s all open source so please feel free to raise a PR or an Issue here. I know we’re a bit light on the AI specific schedulers/ stacks (not that I’m biased against them or anything 😁) such as run:AI and Anyscale’s Ray.
Feedback and input are very welcome so please do share.
https://scheduler-select.hmxlabs.io
hmxlabs
