ABSTRACT
Heterogeneous multicore architectures integrate energy-efficient wimpy (small) cores with highly-performing brawny (large) cores are emerging as a compelling server design option. In this talk, we explore the assignment of threads to cores in two ways. First, we consider the computational and memory bandwidth constraints of the threads and propose an optimization approach (model and experiments). Second, we describe a latency-sensitive task mapping that are least power-hungry and sufficient to meet the QoS requirements.

BIOGRAPHY
Daniel Mosse is Professor and Chair of the Computer Science Department at the University of Pittsburgh. The current major thrusts of his research are real-time systems and power/energy management issues, including sensor and wireless networks. Typically funded by federal agencies, his projects combine theoretical results and actual implementations. He received a BS in Mathematics from the University of Brasilia in 1986, and MS and PhD degrees in Computer Science from the University of Maryland in 1990 and 1993, respectively. Dr. Mosse received the Provost's Innovation in Education Grant/Award in 2007 for redesigning the Introductory Programming course for non-majors. He received the Tina and David Bellet Teaching Excellence Award in 2006 (one of two among over 500 faculty members in the School of Arts and Sciences). Dr. Mosse has served on the TPC and as TPC chair for most major IEEE- and ACM-sponsored real-time conferences. In 2011, he was co-chair of the International Green Computing Conference and General Co-Chair for the International Conference on Embedded and Real-Time Computing Systems

Conference Information