ARVIND MITHAL (1947-2024)

BY SRINI DEVADAS AND FRIENDS OF ARVIND
SUBMITTED BY THE NAE HOME SECRETARY

ARVIND MITHAL, a towering figure in computer architecture and digital systems, passed away on June 17, 2024, at 77. A pioneer of dataflow computing, parallel processing, and hardware design, Arvind was also a deeply committed educator, a cherished mentor, and an advocate for equality.

Born in India, Arvind earned his bachelor’s degree in electrical engineering from IIT Kanpur in 1969, where he developed an interest in parallel computing. He moved to the United States for graduate studies at the University of Minnesota, earning an M.Sc. and a Ph.D. in computer science in 1972 and 1973, respectively. His doctoral work focused on operating systems and mathematical models of program behavior.

He began his academic career at the University of California, Irvine, teaching there from 1974 to 1978, before joining the faculty at MIT. Initially reluctant to remain in the U.S., he was persuaded by his mentor, M.V. George, to reapply to MIT. This decision led to nearly five decades of transformative contributions at the institution.

As the head of the Computation Structures Group at MIT’s Laboratory for Computer Science — later part of the Computer Science and Artificial Intelligence Laboratory) — Arvind became a global leader in dataflow computing. Dataflow architectures optimize computation by leveraging parallelism, a focus that shaped his career and led to the Monsoon project in the 1980s and 1990s. In collaboration with Motorola, his team built 16 dataflow computing machines and their associated software. One of these machines is now displayed at the Computer History Museum in Mountain View, California.

Arvind’s work spanned programming languages, high-level synthesis, and formal verification of digital hardware. He led the development of Id and pH, declarative programming languages for parallel computing, culminating in his 2001 book Implicit Parallel Programming in pH (Morgan Kaufmann), co-authored with R.S. Nikhil.

In the 1990s, Arvind shifted focus when, as he explained in a 2012 interview for the Institute of Electrical and Electronics Engineers (IEEE), funding for parallel computing research declined. “Microprocessors were getting so much faster that people thought they didn’t need it,” he recalled. Instead, he applied his team’s techniques for parallel programming to the principled design of digital hardware, leading to groundbreaking tools for hardware verification, memory models, and cache coherence protocols.

In 2000, he founded Sandburst, a fabless semiconductor company, serving as its president before returning to MIT as an advisor. Sandburst was later acquired by Broadcom. In 2003, he co-founded Bluespec, Inc., building on his development of Bluespec, a language designed to automate chip design. Over the past decade, he introduced these language and design innovations into MIT’s undergraduate curriculum, modernizing courses such as Computation Structures and Deep Learning.

Arvind’s presence in the computer architecture community continues through his deep technical contributions and the legacy of his students. Over his career, he advised more than 35 Ph.D. students, many of whom became leaders in academia and industry.

A deep thinker with a keen sense of abstraction, he had a gift for simplifying complex systems. He pushed his students to explain their research at abstraction levels that hid the implementation complexity. Those who co-authored papers with him knew his single-minded devotion to fully understand each result at its most fundamental level. Even after research was complete and others thought the paper was ready for submission, Arvind would insist on further refining its insights, often pushing us to understand it at yet a deeper level. Then, no matter how late the hour, we had to articulate this new understanding to the world — to Arvind’s satisfaction. Two of his students recall working on a paper for weeks but failing to meet his exacting standards by the deadline. When they suggested submitting it anyway, he told them “You can submit it, but take my name off of it.” This was his way of teaching them the power of abstraction and clarity, often leading to insights and advancements that would have been difficult to uncover otherwise. Arvind would say that once a researcher became accustomed to this abstract thinking style, there was no turning back. This approach defined him in both his intellectual and interpersonal pursuits.

Arvind believed that knowledge was its own reward. Reflecting on the content of annual reviews, he once commented that if he gained one (or maybe two) deep insights in a year, he considered it a success. That perspective epitomized self-actualization.

He believed in the innate abilities of his students, encouraged them to take ownership of their brilliance, remarking, “A Ph.D. advisor shouldn’t believe that they had anything to do with the brilliance of the students they graduate when the students were already bright when they came to the advisor!” His sage advice to new professors was simple: “The best thing an advisor can do is to keep pointing a student in the right direction and hang on for the good ride!”

He treated his research group like family. His wife, Gita, often brought their sons to his office when they were young, as much to see the graduate students as to see Arvind. Discussions frequently lasted late into the evenings, and when hunger struck, Arvind would call on Gita to bring food for everyone — something she never failed to deliver!

Arvind’s mononym reflected his rejection of caste markers, a political statement affirming his egalitarian values. His marriage to Gita Singh Mithal defied conventions, rooted in love and shared ideals rather than societal expectations. Together, they created a welcoming home for students, colleagues, and friends.

An adventurer at heart, he explored the Serengeti, traveled overland from London to Kabul, and immersed himself in diverse cultures during sabbaticals in Tokyo, Barcelona, and Seoul. His curiosity and zest for life were matched by his optimism, encapsulated in his favorite saying, “Pessimists are more often right, but optimists live happier lives.”

Passionate about teaching, Arvind played a significant role in developing course content for computer architecture and hardware design, both at MIT and other universities. He prepared extensively for his lectures, sometimes for days, reviewing each slide in excruciating detail — even for lectures he had given many times before. His love for and dedication to teaching never wavered, even when he was ailing; in spring 2024, from a hospital bed, he watched and critiqued final project presentations and participated in final grading for the course he was teaching.

Over the past decade, he focused on advancing undergraduate education at MIT by introducing modern design tools to courses 6.004 (Computation Structures) and 6.191 (Introduction to Deep Learning) while incorporating Minispec, a programming language closely related to Bluespec.

Arvind’s contributions were widely recognized. He was elected to the National Academy of Engineering in 2008 and the American Academy of Arts and Sciences in 2012. A fellow of IEEE and ACM (Association for Computer Machinery), he received the Harry H. Goode Memorial Award in 2012 for “fundamental contributions to research in dataflow computing, memory models, and cache coherence protocols.” His alma mater, IIT Kanpur, honored him as a distinguished alumnus, and he was a member of the Indian National Academy of Sciences and other prestigious organizations.

On a personal level, Arvind was a wonderful person — truly kind, generous, and always smiling. He was an inspiring mentor and a true friend to countless students, colleagues, and researchers worldwide. Arvind and Gita opened their Arlington home to anyone they met, creating a welcoming space for all. Their annual Diwali parties, hosted for more than three decades, were highlights for those fortunate enough to attend. Guests were invariably amazed by the array of friends from all walks of life that Arvind and Gita had brought together. They were true pillars of both the MIT and Arlington communities, and in some circles, Arvind was simply known as “Gita’s husband.”

His contributions shaped modern computing, and his influence endures in his students, colleagues, and the broader academic community. His son Divakar described him as “an idealist before he was a mononym, an adventurer before he was a scientist, an egalitarian before he was a professor.” As one of his students put it, “Once you learn to think the way Arvind taught you, you can never go back.” His legacy is not just in his groundbreaking research but in the countless lives he inspired.

Arvind is survived by his wife, Gita; their sons, Divakar and Prabhakar; their wives, Leena and Nisha; and two grandchildren, Maya and Vikram. His warmth, humility, and intellectual brilliance will be cherished by all who knew him.

Rest in peace, Arvind. You will live forever in our hearts and minds.