Diana Marculescu Carnegie Mellon University

Energy Aware Computing: Synchronous vs. Partially Asynchronous Processor

This talk will be centered around various themes in energy aware computing, ranging from fine-grain resource scaling to partially asynchronous processors with full local speed and voltage control. I will provide an overview of our research in microarchitecture-level power management and will discuss in detail the effect of using a Globally Asynchronous Locally Synchronous (GALS) organization for a superscalar, out-of-order processor, both in terms of power and performance. To this end, we propose a novel modeling and simulation environment for multiple clock cores with static or dynamically variable voltages for each synchronous block. Using this design exploration environment we were able to assess the power/performance trade-offs available for Multiple Clock, Single Voltage (MCSV), as well as Multiple Clock, Dynamic Voltage (MCDV) cores.

Our results show that MCSV processors can be more power efficient than their fully synchronous counterparts, at the expense of some performance loss. However, for some applications, a MCDV design can provide more than 22% average power savings with less than 12% performance reduction when compared to the fully synchronous core. I will conclude the talk with some recent results on the effect of GALS organization in multithreaded and clustered processors

Bio
Diana Marculescu is an Assistant Professor of ECE at Carnegie Mellon University. She has received her Ph.D. in Computer Engineering in 1998 from University of Southern California and her M.S. in Computer Science from “Politehnica” University of Bucharest in 1991. Diana Marculescu is a recipient of a NSF Career Award (2000), a member of the organizing committee of the ACM/IEEE Intl. Symposium on Low Power Electronics and Design and general or program chair for several workshops on low power related topics. She also serves on the technical program committee of several conferences, including IEEE/ACM Intl. Conference on Computer-Aided Design and IEEE Design, Automation and Test in Europe Conference. Her research interests are in the area of energy aware computing, VLSI, computer architecture and CAD for power modeling and estimation.