Strelka, Swarthmore’s high-performance computing (HPC) system, is set for an exciting expansion next month. This upgrade is needed because faculty and students are increasingly using the system to conduct research, teach courses, and other uses, resulting in bottlenecks and resource limitations. In some ways, this is a good problem to have! We love it when users reasonably push the system. But, we want to ensure that robust use doesn’t result in substantial wait times, or running out of storage space.
The October upgrades consist of three primary items:
- A parallel storage node, which will offer an additional ~850TB of space for our users, representing a 14x increase in usable space! This will facilitate, for example, the analysis and storage of larger data sets, while also providing greater redundancy and resiliency against hardware failure.
- Three additional compute nodes, allowing a greater number of concurrent users and jobs to interact with the system, resulting in a 33% growth in the number of general-purpose compute nodes.
- A new management node, which will simplify administration and monitoring of the cluster as well as provide greater reporting of, and insight into, metrics and usability statistics.
If you are interested to learn more about Strelka, or how you can use it for a research project or as part of a class, please consider attending the upcoming HPC 101 session on Monday, September 18th, or reach out to Jason Simms (email@example.com).