batch compute cluster
SSCF operates a batch compute cluster which is available to all Division affiliates. Faculty, staff, and graduate students are provisioned with access automatically. Undergraduates (for example research assistants) can be provisioned upon request by a faculty member.
Our goal is to have a computing resource available to our faculty and students that doesn't require them to go ask for permission and justify what they want to do - we'd like access to a reasonable level of computing power to be something people can simply expect to be available to them when they join Social Sciences.
For our divisional users, we also hope we can have most of the common discipline-specific software readily available. When your compute needs grow large, we're lucky enough to have SDSC right here on campus, so we don't envision this growing into something at supercomputer scale - this is a complement to the services they provide, not a substitute.
A quickstart guide for the cluster is available, there is cluster-specific documentation for R, python, julia, etc. and for help you can always contact the infrastructure team.
This is the first time we've made a resource like this available across the entire division, so we're also looking to all of you for advice on what packages need to be made available (specific R or matlab addons?), which types of resources are more important (high single-thread performance? More RAM-per-CPU? More GPUs?) and we'd love feedback on the documentation as well.