Alfred University students and faculty will have new opportunities to work with large-scale computing resources following computer science faculty Chandan Kumar’s participation in the competitive 2026 OSG School at the University of Wisconsin-Madison.
OSG is a consortium dedicated to the advancement of open science via the practice of distributed high throughput computing, and the advancement of its state of the art. Kumar, assistant professor of computer science, was selected to attend the weeklong program July 13-17, where he trained directly with OSG staff in high-throughput computing using HTCondor and the OSPool, a national distributed computing resource available to U.S.-based academic researchers.
Kumar gained long-term access to the OSPool and the ability to introduce Alfred students and faculty to the platform, opening the door to computing capabilities that would be difficult to replicate using local hardware alone.
“The biggest shift was in how I think about computing itself,” Kumar explained. “Instead of running one large task on one powerful machine, you break the work into many independent pieces and run them in parallel across a huge pool of shared machines.”
That approach, known as high-throughput computing, is especially useful for research that requires hundreds or thousands of independent calculations or experiments to be completed. Rather than waiting for one computer to process each task sequentially, researchers can distribute those tasks across a large network of machines. For Kumar, the model is a natural fit for both his research and teaching.
“It fits my research almost perfectly, since I run many separate experiments, and it fits teaching just as well, because it lets an entire class compute at once instead of taking turns on one machine,” he said. “It is a genuinely different way of working, and it scales in a way local hardware never could.”
The OSPool access adds another layer to Alfred University’s growing access to national research computing infrastructure. Kumar also recently received an allocation through the National Science Foundation’s ACCESS program, which provides researchers with time on some of the country’s advanced supercomputing systems. While both resources expand the amount of computing power available to Alfred researchers and students, Kumar said they are designed for different types of problems. High-throughput computing is particularly effective when researchers need to run a large number of independent tasks simultaneously. Supercomputing resources, by comparison, are designed for individual problems that may require many processors or graphics processing units working together at the same time.
“One gives you breadth and volume, the other concentrated power,” Kumar said. “Having both means we can match the right tool to the right problem.”
For Alfred students, Kumar sees the OSPool as an opportunity to gain practical experience with computing infrastructure typically associated with much larger research institutions. Through coursework and future projects, students will be able to move beyond the limitations of a personal laptop and learn how to distribute computational workloads across national infrastructure. That could allow a student to run hundreds of experiments simultaneously while developing skills applicable to fields ranging from computer science and artificial intelligence to data analysis and scientific research.
“It levels the playing field for our students,” Kumar said. “An Alfred undergraduate gets free, hands-on access to the same national computing network that researchers across the country rely on, right inside their coursework.”
The resource also allows Alfred to expand access to advanced computing without first having to build and maintain a large computing cluster of its own. “For a university our size, that is a genuine edge. It is exactly the kind of opportunity students usually have to go to a large research institution to find,” he said.
Kumar plans to make high-throughput computing a continuing part of teaching and research at Alfred rather than limiting its use to individual projects. In the near term, he intends to incorporate the technology into additional courses and develop an introductory HTCondor seminar that would give students and faculty from across disciplines an opportunity to work with the system. Longer term, he sees access to national resources such as the OSPool as a foundation for building Alfred’s own computing capabilities.
“As more of our students and faculty start doing this work, the case for investing in local infrastructure to build on top of these national systems only gets stronger,” Kumar said.
Kumar’s goal is to make large-scale computing a regular part of the academic experience at Alfred University and create opportunities for undergraduate students to engage with tools and infrastructure they may later encounter in graduate research or industry. “My vision is an Alfred where working with large-scale computing, national and local, is simply part of what it means to be a student or researcher here. It is the kind of opportunity people usually associate with much larger institutions.”
Story by Andrii Maltzev ’27
