Unless otherwise noted, all seminars will take place in the 6th floor conference room of Donald Bren Hall (DBH 6011). Refreshments will be served at 10:50am, and the seminar talks will run from 11:00am until noon.
For additional information, please contact CS Seminar Administrative Coordinator, Mare Stasik, at mstasik@ics.uci.edu or (949) 824-7651.
Sang-Woo Jun
Assistant Professor, Department of Computer Science
University of California, Irvine
October 5, 2018
11:00am - 12:00pm
Donald Bren Hall 6011
Title:
Cheaper, faster computers: Affordable Big Data analytics using hardware accelerators and NVM storage
Abstract:
CPUs are no longer becoming faster at a rate they used to a decade ago, which is a problem because the amount of data to process is exponentially growing. This gap opens up exciting opportunities for systems research, on how to incorporate new technologies like computation accelerators, memory-class storage, and networks to continue performance scaling at an acceptable cost. In this talk, I will present my work on systems using reconfigurable hardware accelerators and flash storage. I will also present how a PC with a custom storage device was able to outperform a small server cluster in graph analytics and other applications.
Speaker Bio:
Sang-Woo Jun is an Assistant Professor in the Department of Computer Science at the University of California, Irvine. His research interest is in more efficient, high-performance computer systems for Big Data analytics, aiming to boost performance and lower cost using FPGA-based application-specific hardware acceleration and NVM (non-volatile memory) storage. Sang-Woo has designed systems have improved performance at 1/10 cost and power consumption, on applications including graph analytics and bioinformatics