Speaker

Location

McDonnell Douglas Auditorium

Date & Time

Friday, January 16, 2004
Refreshments served at 10:00 a.m., lecture begins at 10:30 a.m.

Abstract

The term "system-level design" has been defined, used and abused countless times by designers, cad developers and most of all, tools sales people. By now we all know what is supposed to be, and it 'ain't pretty` (despite the claims by cad vendors). In this talk, I will argue that for the past 5 years (at least) we have tried to develop system-level tools by fattening the already beaten up ASIC design methodology, by exaggerating the advantages of high-level languages, by overblowing the benefits of software-related methodologies for designing hardware, among other very incremental approaches. The design methodologies and tools currently used for system-on-chip design are an evolution of ASIC and board-level design tools. Such an incremental approach has been ineffective in closing the productivity design gap. I will analyze the current push to use high-level languages and software approaches for hardware design, and try to convince the audience that such attempts, while in the right direction, are far from sufficient. I will argue that radical new approaches for system design are needed if we want to be able to design new products in shorter time-to-market. I will discuss some of the approaches we are developing at IBM Research and speculate on possible new research avenues.

Biography

Reinaldo Bergamaschi graduated in Electronics Engineering from the Aeronautics Institute of Technology, Sao Jose dos Campos, Brazil , in 1982, and in 1984 he received the M.E.E degree from the Philips International Institute, Eindhoven, The Netherlands. In 1989 he obtained the Ph.D. degree in Electronics and Computer Science from the University of Southampton, Southampton, England, and joined the IBM T. J. Watson Research Center, Yorktown Heights , USA. He has worked extensively on RTL, behavioral and system-level design tools and methodologies. He was the main architect of IBM's high-level synthesis system and he is currently developing system-level modeling and synthesis tools for core-based, platform-based designs. He has participated in numerous conference committees and given several tutorials and lectures on high-level and system-level design issues and tools.