Wayne Wolf

Professor of Electrical Engineering
Ph.D. 1984, Stanford University

A major thrust of my research is hardware/software co-design of high-performance application-specific computing systems. We use multimedia systems as example systems in many of our studies, although the techniques can be applied to other domains as well. Application-specific systems often use heterogeneous multiprocessor architectures to meet cost/performance/power goals. Such systems are heterogeneous in both their software and hardware architectures. The sophisticated nature of the implementations, coupled with the importance of meeting hard design metrics such as cost and performance, make the analysis and synthesis of such systems especially challenging. My students and I have worked on and are continuing to work on a number of problems in hardware/software co-design, including performance analysis of heterogeneous multiprocessors, performance analysis of realistic communication links, memory system analysis and synthesis, and the integration of ASIC's and CPU's in heterogeneous multiprocessors.

As a separate but complementary activity, I am studying the architecture of programmable video signal processors (VSP's). Over the next few years, advances in VLSI technology will make it possible to fabricate large chips that include both multiple parallel computation units and significant amounts of on-chip memory. My students and I are looking into a number of problems related to single-chip VSP's, including VLSI studies of the properties of VSP subsystems, microarchitectural studies on the hardware requirements for implementing advanced multimedia applications, and compilation techniques that map programs onto the available memory and function unit parallelism.

I am also interested in video libraries, including architectures for video libraries and algorithms for analyzing video content. We expect video libraries to play an important role in education in the coming years as a means for making available the multimedia historical record of the twentieth century to students; video libraries will also play a growing role in commercial programming. My students and I have studied caching algorithms for efficiently distributing large video objects over the Internet. We have also studied several topics in content analysis, including shot segmentation, key frame extraction, and image search.