Invited Talk (1)

主講人介紹：

Tim Cheng received the B.S. degree in Electrical Engineering from National Taiwan University in 1983 and the Ph.D. degree in Electrical Engineering and Computer Science from the University of California, Berkeley in 1988.

He worked at Bell Laboratories in Murray Hill, NJ, from 1988 to 1993 and joined the faculty at the University of California, Santa Barbara in 1993 where he is currently Professor and Chair of the Electrical and Computer Engineering Department. He was the founding director of UCSB's Computer Engineering program. His current research interests include design verification, test, silicon debug, and multimedia computing. He has published over 300 technical papers, co-authored three books and holds ten U.S. Patents in these areas.

Dr. Cheng, a fellow of IEEE, received Best Paper Awards at the 1994 Design Automation Conference and 1999 Design Automation Conference, 2001 Annual Best Paper Award in Journal of Information Science and Engineering, Best Paper Award in 2003 Conference of Design Automation and Test in Europe (DATE 2003), and the Best Paper award at 1987 AT&T Conference on Electronic Testing. He currently serves as Editor-in-Chief for IEEE Design and Test of Computers, Editor for IEEE Transactions on Very Large Scale Integration (VLSI) Systems, Associate Editor for ACM Transactions on Design Automation of Electronic Systems, Associate Editor for Formal Methods in System Design, Editor for Journal of Electronic Testing: Theory and Applications, and Editor for Foundations and Trends in Electronic Design Automation. He has been General Chairs and Program Chairs for a number of international conferences on design, design automation and test.

演講摘要：

Future hardware systems must have sufficient robustness to cope with failures resulting from the increasing variability and reliability concerns. This requirement not only applies to high-end systems but also becomes a necessity for consumer electronics. Failures due to design bugs, manufacturing defects and variations, and environmental noise are becoming facts to be dealt with, not just problems to be solved.

Built-in redundancy to tolerate errors or built-in self-recovery from errors will become necessary to ensure sufficient system yield and reliability. Designing a robust system with spares and self-reconfiguration capability could also alleviate the need for burn-in/stress test in the manufacturing line, which has become extremely costly. For field-reconfigurable hardware with spares, the hardware with infant mortality failures can be reconfigured in the field by using the error-free spares to improve reliability and lengthen the life cycle. For high-precision and high-performance analog functions, incorporating a self-tuning capability could compensate variations and dynamically adapt to environmental noise and transient errors. It would require on-chip sensors to sense errors and, in turn, tune the functions to minimize the errors. The component must incorporate in the design some tuning “knobs” which will take instructions from the sensors for self-tuning. Such an approach can boost the precision of an intrinsically imprecise operation under a wide range of variations and noise.