Women’s History Month Spotlight: Lynn Conway, Pioneer of VLSI Design

In celebration of Women's History Month, All About Circuits has been highlighting remarkable women in the engineering world. Last week, we spotlighted Grace Hopper, the namesake of NVIDIA's processor line. 

Making engineering history is hard enough, but American computer scientist and electrical engineer Lynn Conway smashed through more than one glass ceiling to leave her permanent mark.

 

Lynn Conway. Image used courtesy of the News House

 

Conway was born in the state of New York in 1938. Among her many notable accomplishments is her role in the Very-Large-Scale-Integration (VLSI) revolution in the late 70s, which changed computer architecture design forever.

However, a decade before this feat, Conway was working as a successful engineer at IBM by a name given her at birth: Robert Saunders. Conway's journey to revolutionize the engineering world as a transgender woman is an incredible story of perseverance and brilliance worthy of reflection during Women's History Month and beyond. 

 

Childhood and Education

Conway’s childhood took place during the tail-end of World War II. Her father was a chemical engineer and introduced her to the world of science, mathematics, and electronics at a young age, while her mother was a teacher. Conway considered Charles Steinmetz and Edwin Armstrong among her heroes (the fathers of alternating-current and FM radio, respectively). 

Steinmetz’ story seemed to have left a particular impression on young Conway. Steinmetz inherited kyphosis, a condition causing excessive spinal curvature. In one of Conway’s reflections, she mentions that his story “carried an embedded message … [someone] who was somehow perceived as different might become liked, even honored, if they made valuable contributions”. 

Conway has written candidly about her childhood and gender dysphoria growing up. In parallel, she excelled at school and continued to foster her interest in the sciences.

In 1955, Conway was accepted into the Massachusetts Institute of Technology (MIT) to study physics. During her time at MIT, she began to transition but ultimately halted due to the lack of medical support. Uncertain about her future as a scientist and still struggling with her identity, she dropped out of MIT in 1959 and went to work as an electronics technician.

 

Columbia University and IBM

After a short while, Conway decided to return to college. She recalled the circuit theory and electronics classes she had taken during her physics studies at MIT and decided to apply to the electronics engineering program at Columbia University in 1962. She completed her bachelor’s degree within a year and followed with her master’s degree the year after. 

In 1965, IBM recruited Conway to work on the Advanced Computing System project, with the goal of designing the world’s most powerful scientific computer. The project was relatively secretive, and the first iteration focused on innovations such as the instruction set and branching hardware. 

Conway observed that the architecture suffered from problems that resulted in processing bottlenecks while waiting for instruction results. Eventually, she came up with a dynamic instruction scheduling scheme (DIS) that could issue multiple instructions out of order. 

ACS would never be built, though, and IBM eventually shelved the project along with Conway’s DIS invention.

 

Starting Over in Secret

In 1968, Conway was finally in the position to go through with her gender transition with more medical support available. She advised IBM’s human resources team about her transition and was fired shortly after by the CEO. 

 

Conway at Xerox PARC (1977). Image used courtesy of IEEE Solid-State Circuits

 

Conway was left to start over again, and due to this experience, held her past critically secret. Now fully living as Lynn Conway, she took a programming contract job at Memorex from 1969 to 1972. Over this period, Conway began to rebuild her credibility as an engineer and, in 1973, found herself at the Xerox Palo Alto Research Center (PARC). In 1976, she took on the challenge of investigating ways to improve silicon design during a period when the number of transistors that could be manufactured on a chip was quickly increasing but outpacing fabrication techniques.

 

Conway's Indelible Mark on VLSI Systems

Her main collaborator would be Carver Mead, another American scientist and engineer. Together they came up with a simplified set of layout design rules that utilized the concept of ratios for spacing and widths of cell topologies instead of fixed lengths for a specific process. This way, the design rules could be applied for different designs more easily and only require a two-page set of design rules to check as opposed to the 40 pages commonly required at the time.

 

Some examples of the VLSI design rules. Image used courtesy of IEEE Solid-State Circuits

 

To ease the adoption of the rules, Mead and Conway published a textbook together called Introduction to VLSI Systems. To accompany this book, Conway also developed a lecture that she taught first at MIT. The course was designed in two parts, the first focusing on theory and the second on applying the theory to an actual design. Conway believed this would be a good way to prove their new system worked. 

After most students succeeded in designing a chip and had it manufactured with minimal issues (including one student who designed a LISP microprocessor), Conway and Mead moved on to making the course materials more widely available for other educators to adopt.

More significantly, Mead spearheaded the concept of foundries to allow designers to submit their designs over the ARPANET to a manufacturer. Xerox PARC took on the responsibility of developing this system so that universities using the Mead-Conway course materials could similarly have their students manufacture their designs. In the following few years, many of the most important institutions in engineering education were offering this course, and the VLSI revolution took off. 

 

Conway's Ongoing Legacy

Conway spent most of her career quietly taking on challenges and triumphs. With the attention that the Mead-Conway VLSI course was bringing her and sudden renewed interest in the DIS invention she came up with at IBM, Conway felt it was necessary to reveal her past identity on her own terms. In 1998 she began to share her involvement with the ACS project at IBM. 

In 2020, IBM issued an apology to Conway for discriminatorily firing her in 1968. Since then, Conway continued her successful career as a scientist, engineer, professor, and dean. She has received many accolades for her technical successes and, after coming out, has also acted as an advocate for transgender people. 

Throughout her career, Conway has remained dedicated to her identity as an inventor, scientist, engineer, and woman.