Marios C. Papaefthymiou - Research Summary


Papaefthymiou's research interests are in the design of energy-efficient computers. Together with his students, he has pioneered the exploration of charge-recovery (a.k.a. adiabatic or energy recycling) technologies, demonstrating over the span of almost two decades a broad variety of charge-recovery silicon prototypes that achieve substantial (sometimes order of magnitude) energy savings over their conventional counterparts. He has also pioneered the commercialization of charge-recovery technologies, co-founding Cyclos Semiconductor to develop charge-recovery design solutions for energy-efficient high-performance clocking. The Cyclos technology marks the first-ever successful deployment of charge recovery in a high-volume commercial product and is now used in multi-GHz server chips from AMD and IBM.

In addition to energy-efficient design, Papaefthymiou has made contributions to the design and optimization of high-performance digital systems, with a focus on asymptotically efficient retiming algorithms. He has also made contributions to other areas of design automation, including algorithms and tools for power estimation and design complexity management.

Papaefthymiou has co-authored award-winning papers at the 2013 International Conference on Architectural Support for Programming Languages and Operating Systems, the 2012 IEEE Symposium on High Performance Computer Architecture, the 2001 ACM/IEEE Design Automation Conference, and the 1995 ACM/IEEE Design Automation Conference. Among other distinctions, he has received the Arthur Greer Memorial Prize from Yale College, the Outstanding Achievement Award from the EECS Department at U. Michigan, the Ted Kennedy Family Faculty Team Excellence Award from the College of Engineering at U. Michigan, the Faculty Recognition Award from the Graduate School at U. Michigan, a Young Investigator Award from ARO, CAREER and ITR Awards from NSF, and multiple IBM Faculty Partnership Awards.

Over the years, Papaefthymiou's research has been supported by NSF, ARO, DARPA, SRC, Broadcom Foundation, IBM, and Intel.