Time： 10:00-12:00 a.m., June 18, 2012 (Monday)
Place：Room 446, ICT. CAS
Speaker: Prof. Kathryn McKinley
Power and energy constraints are now the driving force in all device markets from the cloud to desktop to mobile. To deliver energy efficiency however requires both hardware and software innovations, and in particular, vastly more capable system software (compiler, runtime, OS) together with new programming models. This requirement comes in part from the challenge of optimizing power and performance together, which is substantially more difficult than managing performance alone. This talk presents quantitative power, performance, and energy measurements on a range of workloads and devices, from server class machines to mobile. We show how a managed runtime can exploit differentiated workload characteristics to significantly improve energy efficiency. Our results suggest that software must play a larger role in order to achieve energy efficiency in future systems.
Kathryn McKinley is a Principle Researcher at Microsoft Research and an Endowed Professor of Computer Science at The University of Texas at Austin. She received her Ph.D. from Rice University. Her research interests span architecture, compilers, memory management, runtime systems, reliability, security, and software engineering. She and her collaborators have produced widely used tools: the DaCapo Java Benchmarks, TRIPS Compiler, Hoard memory manager, MMTk garbage collector toolkit, and Immix garbage collector. Her service includes program chair for ASPLOS '04, PACT '05, PLDI '07, ISMM'12, and CGO'13;Editor-in-Chief of ACM Transactions on Programming Language Systems(TOPLAS) (2007-2010); CRA-W co-chair (2011-present), and CRA Board member (2012-present). Her awards include an NSF Career Award, IBM Faculty Fellows, an ASPLOS Best Paper Award, CACM Highlights, IEEE Top Picks, and the SIGPLAN 2011 Distinguished Service Award. She has graduated 14 PhD students. She is an IEEE Fellow and an ACM Fellow.