Presentation of the 2000 Phil Kaufman Award to Yen-Son (Paul) Huang

 

Dr. Yen-Son (Paul) Huang
2000 Phil Kaufman Award Recipient

Presented on October 5, 2000
A. Richard Newton

It is both a pleasure and an honor to present the seventh annual Phil Kaufman Award to Dr. Paul (Yen-Son) Huang—entrepreneur, inventor, technologist, philosopher, and friend to all who have had the pleasure of working with him over the years.

Paul has been intimately involved with the EDA industry since the early 1980’s and is perhaps best known as the developer of the Dracula physical design verification system and principal founder of ECAD, which is today part of Cadence Design Systems. Paul was the inventor, the architect and was in charge of development for the Dracula Physical Verification System for over ten years. Dracula was first introduced into the market in 1983 and is still sold by Cadence today.

Paul also built the engineering team that created the first timing-correct emulator as the principal founder of Pie Design Systems, which merged with Quickturn Design Systems in 1993. He managed the implementation the PiE emulators, which became the Quickturn System Realizer, and eventually formed the principal underlying architecture for all of Quickturn’s future in-circuit emulators. Most recently, Paul Huang founded Novas Software, which provides a unique integrated design debugging system for analysis and verification of complex system-on-chip designs.

If you consider each of these entrepreneurial activities—from layout verification, to system emulation at the gate level, to Novas’ system level debug at the RTL level—you soon see that Paul has be recognizing opportunities and moving up the EDA food chain for more than twenty years.

Paul actually began his lifelong love affair with design automation even before that–as an undergraduate at National Chiao Tung University in Taiwan, where his undergraduate research project in 1971 was focused on circuit simulation. After a brief stint designing high-capacity trunk networks between telephone exchanges for the Taiwan Telecommunication Laboratories, Paul came to Santa Clara University in 1975 and graduated with his Ph.D. His dissertation, completed under the supervision of Professor Shu-Park Chan in 1980, was focused on a very relevant problem of that day—physical mask resizing to compensate for the effects of photolithographic systems and IC processing. Paul, it seems it is an even more relevant problem today—just consider the market cap of Numerical Technologies! After graduating from Santa Clara, Paul joined National Semiconductor in their memory division and continued his career in the layout area by developing a suite of layout verification tools. It was during this period he first met Ping Chao, now president and CEO of Silicon Perspective, and this represented the beginning of what has become a lifelong friendship. In 1980, Paul was attracted away from National to the Gould Systems Engineering Laboratory (SEL), where they were developing a workstation based on this new-fangled technology called a microprocessor, and Paul wanted to learn about that. His boss in this new position was Glenn Antle, another person who was to become closely involved with Paul in the years to come. Now in June 1982, Gould bought AMI, decided they no longer needed to develop their own microprocessor and so laid off around 50 people at SEL, including Paul. But Paul and Glenn negotiated the rights to the design verification source code they had developed and decided to start a company.

It was around this time I first met Paul. Jim Solomon and I were in the process of forming what became SDA Systems and Jim knew Paul from his National days. Jim was interested even then in bringing Paul and his design verification code into SDA and asked my Ph.D. student Ken Keller and I to evaluate the code. With Paul’s permission, we loaded the FORTRAN source code onto a VAX computer and started reading it. Pretty soon we realized that although the code was very well written and easy to ready, and although it contained a lot of subroutines, virtually none of them were ever called! In fact, Paul and his team had used a feature of FORTRAN called an entry point, sort of like a glorified go-to, to jump in and out of the code at will. I recall that Ken and I were not only impressed with the code itself, but also with the fact that Paul and his team could even manage such a large code based using an approach like this. Of course, the coding style made the program considerably more efficient that if one had called the routines directly, and as Glenn Antle told me later “Much of this optimization was due to the 500Kbyte main memory limitation of the Gould machine on which the code was developed.” This code base soon became the Dracula program first sold by ECAD in 1993, and then eventually by Cadence after SDA Systems and ECAD merged in 1988.

Why was Dracula so successful and why has it been the benchmark in design verification for so many years? As Ping Chao told me, “Paul has that rare combination of technical excellence, architectural vision, and sheer passion and drive in whatever he does.” Ping and Glenn agree that Paul is easily the hardest working and most passionate entrepreneur they have ever met. When the team left Gould to start ECAD, Paul, Ping, Glenn and Wen-Jai Shieh mortgaged their homes to raise the capital they needed to buy the VAX 750 they used to develop the code. None of the venture capitalists the spoke to would fund them, until they were a profitable company, at which time all of the VC’s they had talked to wanted to give them money. Oh, how times have changed in the Valley!

One of the first Dracula customers was Ford Microelectronics in Colorado. When the engineers there said they needed some special features added to the program before they would buy it, Paul flew to Colorado with the source code in hand, borrowed their computer, worked through the night and when the Ford engineers arrived for work in the morning, Paul demonstrated everything they had asked for. Needless to say, they made the sale!

But it is not only technical excellence, passion, and drive that has made Paul the successful entrepreneur that he is today. Jim Solomon put it well when he said “Paul is an extremely honest, straightforward and sincere guy. I would trust him with anything.” I think all of us who have known Paul would certainly agree. Professor Ernie Kuh, himself a Kaufman Award winner, told me “Paul is very concerned about people in everything he does.” Again, I think the respect that all of us who have had the pleasure of working with him share in this regard is uncontestable. When Paul started PiE Design systems and asked me to join him as a Director, I remember asking him where the name PiE came from. I can still hear Paul’s words as he explained to me “Polish, Innovate, and Enjoy—P.I.E. First you polish the software, then you innovate some more, but the most important thing is to enjoy what you are doing. Polish, Innovate, and Enjoy.” Paul, what a great philosophy for any engineer, but especially so for an entrepreneur.

Again at PiE, the logic emulator Paul and the team developed was the first timing-correct implementation of the emulation concept, another breakthrough technology that allowed engineers to create a hardware model of their design, input live data real-time, debug operating systems and applications, and verify the integrated systems before committing to silicon. Quickturn’s System Realizer (& now Mercury) systems, based on the architecture developed by Paul’s team at PiE Design Systems, have the capacity to allow a user to run an entire state of the art design completely within the emulator.

Over the past few years, Paul has continued his lifelong commitment to improving the electronic design process. He was a founder of Novas, where his vision has led to the Debussy high-level design debugging architecture. He is also a director of Verplex, Nassda, Incentia, and of course Silicon Perspective, the company founded by his friend Ping Chao. Virtually every major electronics firm in the world has at one time used Cadence’s Dracula, the PiE-based System Realizer, or the Novas Debussy products to verify and improve the products that they deliver to their customers.

So, in summary, on behalf of Mrs. Kaufman, EDAC, and all present here tonight, I am honored to be able to present the Phil Kaufman Award to Dr. Paul Huang who, like those who have preceded him on this stage, has played a central role in creating the modern EDA industry, through his technical insights and attention to detail, his architectural and entrepreneurial vision, but most importantly through his belief in people and their ultimate importance to this industry.