Advances in reliable computing draw recognition for doctoral student
Posted: October 04, 2011
Reiley Jeyapaul’s work in pursuit of a doctoral degree in computer science and engineering focuses on finding solutions to the problems that cause computers to crash, lose data or behave erratically.
His progress has been impressive enough to earn invitations to present research papers on his projects at two of the most prominent international computer science and engineering conferences. He presented at the International Conference on Parallel Processing in Taipei, Taiwan in September. He’ll return to Taipei in October to present at the International Conference on Compilers, Architectures and Synthesis of Embedded Systems.
The invitations to present at these high-profile gatherings of computer scientists are notable achievements, and especially rare for a student, says Aviral Shrivastava, Jeyapaul’s doctoral studies advisor.
Shrivastava is an assistant professor in the School of Computing, Informatics, and Decision Systems Engineering, one of ASU’s Ira A. Fulton Schools of Engineering. He notes that only about 20 percent of the scientists and engineers who submit papers to the conferences are chosen to present.
“Reiley submitted two separate papers and they were both accepted, which is extraordinary. Overall this year, four of his papers have been accepted for publication and presentation at prestigious computer design conferences,” Shrivastava says.
Jeyapaul, who has a master’s degree in electrical engineering from ASU, is working in an emerging specialty called reliable computing.
Reliability challenges have arisen in the wake of the development of smaller and smaller transistors used in computers. The tiny transistors have enabled computers to perform more rapidly and to use less power to operate. But at the same time they present the drawback of more frequently exposing computers to disruptive electrical phenomena.
Jeyapaul’s goal is to keep computation reliable despite that the smaller transistors make computing systems more prone to malfunctions. He has been developing new multi-core processors for computers. Multi-core processors are comprised of two or more independent cores that read and execute programmed instructions.
Jeyapaul has been able to redesign computer components so that a system can diagnose and fix problems without slowing down the cores. This enables computer systems to perform the check-and-correct function while computational activities proceed at a normal pace.
The big challenge is to be able to do this without slowing computing operations when even more checks and corrections will be needed as additional new functions are loaded on to computers in the future.
His solution helps ensure more dependable computer operation, and improves on earlier methods that caused computing to slow down or cease while problems were resolved.