High Performance Computing | Invitation to try the "Unity Cluster"

For the past few years, a number of departments at UMass Amherst have been exploring different options for providing access to high-performance computing capabilities for scholars working with large, hard-to-wrangle datasets. Last year, the folks at Research Computing invited a handful of researchers to pilot-test a remote fast computing cluster solution (the “Unity Cluster”). The Unity Cluster has moved past the pilot stage, and is now in "beta" and open to a wider group of scholars. 

If you are interested in making use of this resource, please visit the Unity website, where you will be able to:

Mark Pachucki (Associate Professor of Sociology) was involved in the Unity pilot, and offers this introduction: 

You’ll be basically working with a better-user-interface version of R that’s piped through a tool called jupyterlab. In short, it’s better than the kludginess of working with raw R and avoiding the command-line interface, but not as nice as the user-friendliness of working with RStudio (though sort-of-like-it, after some effort). But if you have a 50GB data file, you’re stuck – your laptop or desktop will laugh at you before it blows a fuse, and you’ll need to do something. 

A few points: 
• The cluster does not support Stata, to my knowledge. 
• Also, the cluster does not provide for the security protocols of working with HIPAA-compliant data
• You can request a shared folder if you’re working on a project with multiple users (multiple faculty or faculty + grad student(s), etc.)
 
*Note: the Unity Cluster website is a work-in-progress, and while Unity is in Beta, some documentation may be missing or incomplete. While we offer offer a number of links to facilitate your getting where you need to go, we encourage you to contact the Administrators if you need further assistance.