Remembering Lynn Conway, who revolutionized chip design

Lynn Conway, co-developer of very large scale integrated circuits (VLSI), died on June 9 at the age of 86. VLSI technology revolutionized microchip design by combining thousands of transistors into a single chip to create integrated circuits.

Conway, an IEEE Fellow, a transgender woman, and a transgender rights activist, played a key role in updating the IEEE Code of Conduct to prohibit discrimination based on sexual orientation, gender identity, and gender expression.

"The impact and personal courage that Lynn Conway has demonstrated in the field of engineering has been a source of inspiration to me and countless others," Michael Wellman, a professor of computer science and engineering at the University of Michigan in Ann Arbor, told Michigan Engineering News. Conway is a professor emeritus at the university.

The following profile of Conway is based on an interview the institute conducted with her in December.

Some engineers dream of their groundbreaking technology being recorded in history. But what happens if your contributions are overlooked because of your gender identity?

If you’re like Lynn Conway and facing the same dilemma, you fight back.

Conway helped develop very large-scale integration: the process of creating integrated circuits by combining thousands of transistors into a single chip. VLSI chips are at the heart of electronic devices used today. The technology provides processing power, memory and other features for smartphones, laptops, smart watches, televisions and home appliances.

Conway developed VLSI in the 1970s while working at Xerox Corp.’s Palo Alto Research Center in California with her research partner Carver Mead, then a professor of engineering at Caltech. For years, Conway’s role was overlooked, partly because she was a woman and partly because she was a trans woman.

Since coming out publicly in 1999, Conway has fought hard to get her contributions recognized, and she’s succeeded. Over the years, the IEEE Fellow has been honored by various organizations, most recently the National Inventors Hall of Fame, which inducted her last year, nearly 15 years after Mead received the honor.

Conway was initially interested in the study of physics because of the role it played in World War II.

“After the war, physicists became famous for blowing up the world to save it,” she said. “I was naive and thought physics was the source of all wisdom. I went to MIT, but I didn’t fully understand my chosen major at the time.”

She took a lot of electrical engineering classes because she said they allowed her to be creative. It was through these classes that she found her calling.

She left MIT in 1957 and subsequently earned bachelor’s and master’s degrees in electrical engineering from Columbia University in 1962 and 1963. While at Columbia, she did independent research under the direction of Herb Schorr, an adjunct professor and researcher at the IBM Research Center in Yorktown Heights, N.Y. The research involved installing a list-processing language on an IBM 1620 computer, “which was the most mysterious machine to try to do this on,” she says with a laugh. “It was a cool language that Maurice Wilkes at Cambridge had developed to experiment with self-compiling compilers.”

She must have impressed Shore, she said, because after she earned her master's degree, he recruited her to join his research center. While working on the Advanced Computing Systems Project there, she invented multi-out-of-order dynamic instruction scheduling, a technique that allows the CPU to reorder instructions based on their availability and readiness, rather than strictly following program order.

This work led to the creation of superscalar CPUs, which can manage multiple instruction pipelines to execute multiple instructions simultaneously.

The company eventually transferred her to its Bay Area office in California.

Despite her thriving career, Conway struggled with gender dysphoria, the distress people feel when their gender identity differs from the sex they were assigned at birth. In 1967, she began gender-affirming therapy “to address the horrible existential dilemma I had faced since childhood,” she said.

She informed IBM of her intention to make the transition, hoping the company would allow her to do so quietly. But IBM fired her because she believed her transition would cause “significant emotional distress to colleagues,” she said. (In 2020, the company apologized for firing her.)

After completing her transition, Conway resumed her career as a contract programmer in late 1968. By 1971, she was working as a computer architect at Memorex in Silicon Valley. She joined the company in what is known as "stealth mode." No one except close family and friends knew she was a transgender woman. Conway said she was afraid of being discriminated against and losing her job again. She said that because of her decision to keep her transition secret, she could not take credit for the technologies she invented at IBM Research because they were all attributed to the name she was assigned at birth, her "dead name."

In 1975, she was recruited to join Xerox PARC as a researcher and manager of the VLSI system design group.

It was there that she made history.

Concerned about how Moore's Law would affect microelectronics performance, the Advanced Research Projects Agency (now known as the Defense Advanced Research Projects Agency) formed a consortium of companies and research universities, including PARC and Caltech, to improve microchip design. Conway joined PARC's VLSI System Design Group and worked closely with Carver Mead on chip design. Mead, now an IEEE Life Fellow, is credited with creating Moore's Law.

At the time, making chips required manually designing transistors and connecting them to circuits, a process that was time-consuming and prone to errors.

“There were a whole bunch of different designs done at different levels of abstraction, including basic architecture, logic design, circuit design, and layout design—all done by different people,” Conway said in a 2023 IEEE Annals of the History of Computing interview. “People at different levels passed the designs down in a paternalistic, top-down system. People at any one level might not know what people at other levels in the system were doing or what they knew.”

Conway and Mead decided that the best way to solve the communication problem was to use CAD tools to automate the process.

The two also introduced a structured design approach to chip manufacturing that emphasizes high-level abstraction and modular design techniques, such as logic gates and modules, to make the process more efficient and scalable.

Conway also created a set of simplified chip design rules that enabled integrated circuits to be digitally encoded, scaled, and reused as Moore's Law progressed.

The approach was so radical that it needed help to become popular, she said. Conway and Mead wrote Introduction to VLSI Systems to bring the new concepts directly to the next generation of engineers and programmers. The textbook covered the basics of structured designs and how to verify and validate them. Before it was published in 1980, Conway taught the first VLSI course at MIT in 1978 to test how well it explained the approach.

The textbook was so successful that it became a foundational resource for teaching the technique, and by 1983, nearly 120 universities were using it.

The work of Conway and Mead led to the so-called Mead and Conway Revolution, which enabled the development of faster, smaller, and more powerful devices.

Throughout the 1980s, Conway and Mead were known as the creators of VLSI. Together, they received several awards, including the 1981 Achievement Award from Electronics magazine, the 1984 Pender Prize from the University of Pennsylvania, and the 1985 Weatherill Medal from the Franklin Institute.

In 1983, Conway left Xerox PARC and joined DARPA as Assistant Director for Strategic Computing. She led planning for the Strategic Computing Initiative, which was designed to expand the technological basis for intelligent weapon systems.

Two years later, she began her academic career at the University of Michigan as a professor of electrical engineering and computer science. She served as associate dean of the university's engineering department and taught there until her retirement in 1998.

Public account ID: icbank