Chun-Yi Su of Blacksburg, Va., a doctoral candidate from Virginia Tech's Department of Computer Science has been selected for the prestigious Lawrence Scholar Program at Lawrence Livermore National Laboratory.

Founded in 1952, the Lawrence Scholar Program recruits new scientific and engineering talent from the world's leading universities. Appointed candidates work closely with their academic thesis advisor and a technical mentor within the scholar program.

One of a dozen doctoral students to be selected nation-wide, Su will be provided a monthly salary, totaling approximately $200,000 for a total four-year period. He will work 25 hours per week in the laboratory, while also maintaining full-time enrollment at Virginia Tech.

"This is a wonderful opportunity for Chun-Yi [Su] and an honor for our department to have a student involved in such a selective program at one of our prestigious national labs," said Barbara Ryder, the J. Byron Maupin Professor of Engineering and the department head of computer science, part of the Virginia Tech College of Engineering.

The doctoral candidate will focus his research on memory locality models for supercomputers while completing his thesis.

"Modern supercomputers are energy/power monsters," said Su. "Running a top-10 supercomputer uses up 20 million watts and costs $10 million per year. Data movement and memory consume 40 percent to 60 percent of the power source. Data locality is key to improving performance and reducing power consumption of systems."

"The scholarship will provide him [Su] access to cutting edge hardware and software unavailable anywhere else in the world," said Kirk Cameron, a professor of computer science and a research fellow in the College of Engineering. "The opportunity will enable him to develop techniques optimizing emerging architectures that will become common in the next three to five years."

Cameron is Su's academic thesis advisor and director of the both the SCAlable Performance Laboratory, or SCAPE for short, and Center for High-End Computing at Virginia Tech in which Su works and utilizes emerging technologies to research the design, analysis, and improvement of scalable systems and applications.

Su received the 2012 best poster award at the annual Lawrence Livermore National Laboratory Student Poster Symposium. And in 2011, Su completed an internship at Pacific Northwest National Lab where he worked to develop a MODA, short for Memory Centric Performance Analysis tool. MODA is designed to instrument, collect, and analyze streams of memory requests on high performance computing systems.

Su received his bachelor's degree in computer science from the National Chiao Tung University, Taiwan and a master's in computer engineering from the National Cheng Kung University, Taiwan.

"His [Su's] work in the laboratory will directly improve the systems we use today to simulate the effects of global warming, model human gene sequences, and develop alternative energy technologies," said Edgar A. Leon, computer scientist and Su's technical mentor at Lawrence Livermore.

The findings of Su's research will be presented at the Virginia Tech colloquia series to an audience of faculty and students. At the national laboratory, the research will be disseminated through yearly seminars for laboratory scientists, postdoctoral appointees, and students. Additionally, the resulting software will be released to the high performance computing tools community for both feedback and broader impact.