KU professor of engineering wins NSF CAREER Award for research into high-performance computing

LAWRENCE — An assistant professor of engineering at the University of Kansas who is working to boost efficiency in supercomputers and other high-performance systems is the winner of a five-year, $558,000 grant from the National Science Foundation.

Hongyang Sun, faculty member in the Department of Electrical Engineering & Computer Science, is the recipient of an NSF Early Career Development (CAREER) award for his efforts to strengthen high-performance computing (HPC) systems by allowing them to do more work as they continue to expand. His goal is to combine algorithm and machine learning to empower HPC systems to reduce runtime variability, conflicting demands and the differing structures of dynamic workflows.

Eliminating such sources of uncertainty makes it easier for big computers — actually, clusters of networked servers together containing millions of nodes — to run together more smoothly, clearing the way for advancements in everything from neuroscience and medical research to climate modeling, artificial intelligence and beyond.

“These sources of uncertainty are becoming increasingly prevalent in today’s HPC platforms, and being able to effectively manage them will provide a significant improvement in runtime performance for a wide range of scientific applications,” Sun said. “This five-year grant will allow me to start an exciting new chapter in my research career and set the stage for influencing how future HPC systems operate at an unprecedented scale.”

Sun’s focus is on resource scheduling, which HPC systems use to govern which applications they run and when. He and his team look to find efficiencies in a combined approach: developing algorithms that will provide more accurate approximations of a range of metrics, from job completion times to resource utilization, while the researchers also use machine learning to train models to deliver accurate predictions for jobs runtimes, performance degradation and structural variability.

“Our hope is to achieve a new level of performance that neither framework — algorithm or machine learning — is able to accomplish alone,” Sun said. “We expect this integrated approach to offer unique perspectives and lay the foundation for the future of HPC resource management.”

Considered to be among NSF’s most prestigious awards, CAREER grants are given each year to about 500 early-career faculty who have the potential to serve as academic role models in research as well as education. The foundation expects recipients’ work to build a firm foundation for a lifetime of leadership in integrating education and research.

The NSF first issued CAREER awards in 1995. Since then, KU’s Lawrence faculty has included 70 recipients: 28 in the School of Engineering, two in the School of Pharmacy and 40 in schools and departments across the College of Liberal Arts & Sciences.

Sun, who arrived at KU in 2021, plans to use the grant to empower what he sees as a series of synergistic activities: outreach programs, curriculum development, student recruitment and more. He aims to engage students in K-12 through graduate school with programs and research opportunities that foster success in science, technology, engineering and math (STEM) and computer science.

Sun said he is grateful for the support he has received from his department, the School of Engineering and KU, and he’s ready to strengthen the performance, reliability, resilience — even energy and thermal efficiency — of large-scale computing systems and applications.

“The CAREER award will allow us to tackle a grand challenge in high-performance computing resource management,” Sun said. “I’m thankful.”