NSF Grant Funds Supercomputer for Sciences

Emma Paul

Professors from the College’s Biology, Physics and Chemistry departments were the recent recipients of a $486,256 grant by the National Science Foundation to build a supercomputer, or a high-performance computing cluster, which will allow students to process data sets of an unprecedented size.

The computer, which is slated to be built by the end of next summer, is replacing an older high-performance computing cluster, which has been used in the science departments for the past nine years.

Nearly 350 Chemistry students per year currently use the old HPC cluster, and the computer also has applications in physics, astrophysics, and computational biology.

Matt Elrod, Biggs Professor of Natural Science, said he hopes that the new system will bring that opportunity to a larger portion of the student body.

“This is a natural evolution in how chemistry has been taught,” Elrod said. “People have used computers since the ’70s, but they were difficult to use and not very powerful, so you couldn’t do many interesting things. Having this supercomputer resource will let us do [computations] in a way that is easy and in a way so we can actually tackle relevant problems, as opposed to super simple demonstrations of principal problems, so we can actually do things that connect to the things students are doing in the experimental part of the curriculum.”

Oberlin students will be some of the few liberal arts students able to work with a HPC cluster, especially at an undergraduate level. Oberlin’s commitment to undergraduate research helped distinguish the school during the competitive review process for the National Science Foundation’s Major Research Instrumentation grant.

“We have a very long history of supporting serious undergraduate research,” said Mike Moore, associate professor of Biology and co-author of the proposal. “We have the existing supercomputer, and the usage Oberlin has gotten out of that for the

past nine years has helped us show NSF that, in fact, we do use this thing quite extensively. I think it’s worth pointing out that not many liberal arts colleges have any computing clusters at all. Oberlin is a leader in this kind of computing among liberal arts colleges.”

The Chemistry department isn’t the only one that will benefit from the HPC cluster. Aaron Goldman, assistant professor of Biology and one of the authors of the proposal, said he believes the HPC cluster will affect the process of research across multiple departments.

“I would say computational research is now a major part of biology and has been a major part of chemistry and physics for a long time, but is now becoming a major part of research in social sciences, as well as humanities research,” Goldman said. “And so, broadly speaking, it opens up research opportunities for students.”

The supercomputer’s advantage comes from its structure. It is not one large computer, but many computers synced together. The computers used are called nodes, each of which will contain at least 64 processors. The nodes are organized on a rack, with each in its own slot. When the HPC cluster is given a calculation, it divides the tasks among processors.

“For a large-scale computational job, there are a couple different things you might need. The one people are most familiar with is storage space, but different from that is the amount of memory the job may require, and also processors that actually do the computation. What a HPC cluster does is actually sync the computers together in order to take those jobs and split them across a bunch of different computers in order to get the most efficient mixture of processing and memory to complete a job.”

For Elrod the new technology is a step toward creating more opportunities for the students of Oberlin College.

“I think having the system is also important for recruiting faculty, which leads to courses being developed that use computational approaches, which leads to research projects for students that involve computational aspects,” he said.