-
EE 598: ELECTRICAL ENGINEERING RESEARCH SEMINAR COURSE #7
Thu, Feb 28, 2013 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Daniel Wong, PhD Student, Electrical Engineerin, USC Viterbi School of Engineering
Talk Title: Energy Proportional Datacenter Servers
Series: EE598 Seminar Course
Abstract: Energy proportionality (EP), the notion where power consumption should be proportional to utilization, is becoming an increasingly important concept in datacenter servers. In this talk, we introduce metrics to accurately quantify energy proportionality and analyze historical EP trends to identify opportunities to improve EP. We find that there exists large energy proportionality gap at low utilization and we present KnightShift, a server-level heterogeneous server that introduces an active low-power mode through the addition of a tightly-coupled compute node called the Knight. We evaluated KnightShift against a variety of real-world datacenter workloads using a combination of prototyping and simulation, showing up to 75% energy savings with tail latency bounded by the latency of the Knight and up to 14% improvement to Performance per TCO dollar spent.
Biography: Daniel Wong is a PhD student working with Prof. Murali Annavaram in the Ming Hsieh Department of Electrical Engineering at the University of Southern California, Los Angeles. His research focuses on novel energy efficient architectures spanning from datacenter servers to GPGPUs. His research has been recognized as one of IEEE Microââ¬â¢s Top Picks in Computer Architecture for 2013. He earned his BS degree in Computer Engineering/Computer Science and MS degree in Electrical Engineering from USC in 2009 and 2011, respectively.
Host: Professor Viktor K. Prasanna
More Information: Course Announcement_EE598_Focused on parallel and distributed computing_(Spring 2013).pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Janice Thompson