310 Larsen Hall
310 Larsen Hall
Gainesville, Florida 32611
With the explosive growth of data traffic in the Big Data era, data movement/communication in both high performance computing systems and mobile platforms are starting to consume significant portion of the system power. This continuing trend urgently calls techniques for energy efficient data communication before energy spent on moving data dominates the information processing stack. This talk discusses potential strategies to improve data movement efficiency and presents two design examples for energy proportional links (EPLs) and near-storage computing, respectively. On one hand, EPLs address one limitation of data movement efficiency that improvement on energy efficiency of link building blocks is not fully translated to power savings at system-level, especially in many practical applications where links are only sporadically utilized. On the other hand, motivated by the excessive cost in data movement and the bandwidth bottleneck in memory/storage, the near-storage computing project tackles data movement challenge by introducing local computation close to SSDs to directly reduce the movement of raw data. Silicon measurements of the energy proportional links and a FPGA prototype of near-storage computing project will be presented to prove the effectiveness and efficiency.
Guanghua Shu (S’10-M’16) is currently a Research Scientist in Oracle Labs, Belmont, CA, USA. He received the Ph.D. degree in the Department of Electrical and Computer Engineering from University of Illinois, Urbana-Champaign, IL, USA.
In the summer of 2014, he was a Research Intern in Xilinx, San Jose, CA, developing power and area-efficient parallel link architectures. He worked on 56Gb/s wireline receivers (both electrical and optical) in IBM Thomas J. Watson Research Center, Yorktown Heights, NY, at Mixed-Signal Communication IC Design group in the fall of 2014 and summer of 2015. His research interests are energy-efficient wireline communication systems, clocking circuits, power converters, and hardware accelerations for efficient computing systems.
He is a recipient of the Dissertation Completion Fellowship (2015-2016) from University of Illinois and the IEEE SSCS Predoctoral Achievement Award (2014-2015).
Dr. John Harris