Undergraduate ARCSIM intern helps install HPC resources
In 2023, the Advanced Research Computing, Security, and Information Management (ARCSIM) group partnered with Computer Science and Electrical & Computer Engineering faculty at UMaine to source a new GPU cluster from Cambridge Computer Services in Boston. Expanding GPU resources was given high priority to provide relief to faculty who are engaged with Artificial Intelligence (AI) and Machine Learning (ML) research, both of which greatly benefit from GPU-based hardware, which in recent years has been a critically scarce resource nationally.
To bolster the startup of the new cluster, ARCSIM brought on board an undergraduate intern, Nathan Price, in late 2023. Originally from Delaware, Price is studying computer engineering at UMaine, and brings to his position deep levels of experience and interest. As a hobby, Price maintains a Linux cluster at home, gaining skills and knowledge that directly translate to his internship.
Working closely with ARCSIM staff, Price has assisted with the physical buildout of the cluster. Price describes it as a puzzle. Much of his time is spent making sure all of the different components are communicating with each other, and this is not always easy. Configuration issues can take time to resolve, as he engages with vendors and seeks solutions. He is also working with ARCSIM staff to understand and define a suite of target user experiences, which is being leveraged to optimize the setup of the cluster.
Price takes pride in this work because of the importance of the research activities that the new cluster aims to support, and he sees his internship as a real growth opportunity. “This type of work is not something anyone gets to do as an undergraduate student,” remarked Price. “I am building a system that faculty will be using as an integral part of their research.” While there is some pressure in that, Price finds it exciting that he gets to use his knowledge to take on this challenge and grow through it.
The internship is also helping Price consider his future. Price, a US ARMY veteran, explained, “I am not a traditional student, and this project has opened a lot of doors for me. I get to meet people whose full time job is working in high-performance computing at national centers like the Ohio Supercomputer Center. I am developing first hand experience in my chosen field by supporting this project.” This internship has also helped Price better understand what he likes about the HPC field and what steps he will need to take to reach his long term career goals.
Once up and running in Spring 2024, the GPU cluster will be supported by Price and ARCSIM staff. This new computing platform will enable the work of AI/ML researchers, and provide the community with a co-located, accessible resource at UMaine.