The Oregon Regional Computing Accelerator (Orca) will build on PSU's existing campus cyberinfrastructure, including the Coeus high-performance computing cluster pictured above.
Portland State will power up a new high-performance computing cluster that will give researchers at universities and colleges across Oregon the ability to advance computing-intensive research projects by processing large datasets and performing complex computations in a fraction of the time — thanks to a nearly $1 million grant from the National Science Foundation’s Campus Cyberinfrastructure program.
The Oregon Regional Computing Accelerator (Orca) aims to provide free-of-cost computing resources and cyberinfrastructure to colleges in the region, with a special emphasis on rural, regional and minority-serving institutions that otherwise wouldn’t have access to cost-prohibitive computing resources.
“Orca embodies our commitment to inclusive innovation by providing critical resources to both rural and minority-serving institutions, enhancing statewide research in vital areas like data science and cybersecurity,” said Rick Tankersley, vice president for Research and Graduate Studies. “This effort not only strengthens our academic partnerships but also equips us to lead in computational research and education.”
The project is designed to support research and educational projects at Heritage University on the Yakama Indian Reservation in Washington, Lewis & Clark College, Oregon State University, Portland Community College, Reed College, Southern Oregon University, University of Oregon, Western Oregon University and Willamette University.
Will Pazner, principal investigator and an assistant professor of applied and computational mathematics at PSU, said data and computational-enabled science are playing an increasingly important role across almost all domains of research and education and computationally intensive fields such as machine learning demand high-performance hardware — the cost of which is often out of reach for smaller institutions.
“We believe this cluster will be instrumental in advancing scientific research, improving computation-related educational opportunities, and enhancing workforce development,” said Pazner, who joined PSU in 2022 as part of a cluster hire in computational science for a sustainable future.
Orca is not PSU’s first high-performance computing (HPC) cluster. In 2016, PSU’s Portland Institute for Computational Science procured the Coeus HPC cluster with the support of an NSF instrumentation grant. Orca, building on PSU’s existing infrastructure, will feature both GPUs — short for graphics processing units, advanced chips used to perform mathematical calculations at high speed — and high-performance CPU cores, a central processing component of a computer.
Its design and specifications have been created through collaboration and consultation with researchers from across the region. Some of the research projects that will benefit with Orca’s acquisition include climate and atmospheric modeling, computational analysis of legal text, geohazard simulation, and graph algorithms for molecular systems biology. The cluster will be designed with user-friendly features such as a web-based interface to facilitate classroom usage, providing students with opportunities to build machine learning models, work with large datasets and experiment with advanced algorithms.
Orca will leverage the network of Link Oregon, a nonprofit consortium that provides high-speed, fiber-optic broadband connectivity to the state’s public and nonprofit sectors.
“High performance and high throughput computing (HPC/HTC) capabilities demand state-of-the-art networks that can cost effectively deliver fast, reliable and secure transport of extremely large datasets without degrading overall network performance,” said Steve Corbató, executive director of Link Oregon and a co-investigator for this project. “We are proud to support Portland State University's expanded HPC efforts and the Orca collaboration with high-speed network services that will enable their researchers to scale their compute and data-intensive initiatives.”
The cluster is expected to be installed and operational by 2025. The project’s co-investigators include Josef Dufek, professor and chair of earth sciences from the University of Oregon; Mark Keever, executive director of digital research infrastructure at Oregon State University; and Greg Anderson, an assistant professor of computer science at Reed College.