NSF CAREER award to bolster cyberinfrastructure performance

“It is often said that a decade in computer science is a century in other sciences,” said Xun (Steve) Jian, an assistant professor in the Department of Computer Science, of the aspects of his field that fascinate him. “New computer hardware technologies constantly emerge.”

“Such a dynamic field provides many new problems to explore and solve, which is exciting for researchers,” said Jian. And Jian has much to be excited about: this year, he received a five-year Faculty Early Career Development (CAREER) program award from the National Science Foundation (NSF). The CAREER award is NSF’s most prestigious award in support of early-career faculty.

The project’s sole principal investigator, Jian is working within NSF’s Software and Hardware Foundations program, which supports potentially transformative research in the design, verification, operation, utilization, and evaluation of computer hardware and software through novel approaches, robust theories, high-leverage tools, and lasting principles.

The award will fund Jian’s research on boosting cyberinfrastructure performance, with direct implications for sustaining the growth of the digital economy in the United States.

“The performance growth of cyberinfrastructure systems, such as supercomputers and cloud, is important to the U.S. economy and well-being,” said Jian. “It helps boost scientific and engineering productivity, which is also vital for its global competitiveness. This research will bridge many research communities, spanning architecture, operating systems, supercomputing, and cloud computing.”

Jian will explore how to co-design central processing units and operating systems to maximize memory utilization in cyberinfrastructure systems, with the aim to unleash the 50 to 300 percent performance/dollar they are currently missing.

MemMax consists of two research thrusts, one targeting high-performance computing systems and the other targeting cloud systems, as these two systems have different causes for their memory underutilization. The research methodology consists of real-system measurements to characterize the behavior of existing systems; hardware prototyping to validate the functional correctness of MemMax; and architectural simulations to quantify the performance improvement it achieves.

To achieve these results, Jian will partner with collaborators from Advanced Micro Devices and Los Alamos National Laboratory to help maximize the real-world impact of the proposed research, as well as receive technical feedback to guide the research.

As part of its education plan, MemMax will engage K-12 students, as well as undergraduate students, in the area of computer architecture research. Partnering with Virginia Tech’s Center for the Enhancement of Engineering Diversity, Jian and his Ph.D. students will perform live demos with some of its hardware artifacts with area K-12 students.

In addition, Virginia Tech undergraduate students will get first-hand experience with computer architecture research through the incorporation of the prototype systems in class projects in two computer organization courses. The resultant class project submissions will also help identify undergraduate students to recruit for research.

Jian’s research has received international recognition, with an accepted paper published in the Top Picks of the IEEE Micro Journal in 2019. IEEE Micro Top Picks honors the top ~10 papers in the architecture computer research community each year

While at the University of Illinois at Urbana-Champaign, Jian was honored by the Young Researcher at the Heidelberg Laureate Forum, a gathering of laureates of some of the most prestigious prizes in mathematics and computer science, including the Abel Prize, Fields Medal, Turing Award, and Nevanlinna Prize. Additionally, Jian was awarded the M.E. Van Valkenburg Graduate Research Award.

Jian earned a bachelor’s in electrical engineering in 2010 and a doctorate in computer engineering in 2017, both from the University of Illinois at Urbana-Champaign.

###