IT News & Events

News about IT at Indiana University and the world

Menu

Bentley University adds HPC cluster with aid from Campus Bridging

HPC cluster provides an important educational resource for growing Business Analytics program.

In January 2016, IU’s Campus Bridging staff assisted in building a new HPC cluster at Bentley University in Waltham, MA. Jason Wells and David Oury at Bentley recently started a project to make supercomputing resources available to their students and community, and reached out to campus bridging staff asking for guidance. Bentley University has many computing resources available to students, but the growing business analytics program did not have access to a real HPC resource until this project was completed. Campus bridging staff helped build the first HPC resource at Bentley over the course of a week, resulting in a 16 node cluster (currently named Rocks-at-Bentley) with 2.8 TFLOPS and 544 GB RAM. Eight more nodes will be added in the future by Bentley, now that staff there has become familiar with cluster building and administration.

The new HPC cluster provides an important educational resource for Bentley's growing Business Analytics program, in addition to providing researchers with a much-needed tool for performing cutting-edge economics and data science research. Researchers now have access to a real parallel environment with all of the scientific software provided by XSEDE, and some licensed software that Campus Bridging staff was able to integrate into the cluster scheduler. For some researchers at Bentley, calculations that once took months to complete can now be split over several machines, resulting in greatly increased productivity. This project also helped staff at Bentley gain familiarity with cluster technology, both from working on the build and from impromptu talks with Campus Bridging staff. Bentley University is engaging with other Boston-area colleges to improve HPC utilization.

Caption: Bentley University staff with IU’s Eric Coulter building the new HPC cluster