The Tech - Online EditionMIT's oldest and largest
newspaper & the first
newspaper published
on the web
Boston Weather: 48.0°F | Mostly Cloudy

CORRECTION TO THIS ARTICLE:
This article misspells the last name of Chandrakasan’s co-author in “Low-Power CMOS Digital Design.” He is Samuel Sheng, not Cheng.

This article also imprecisely describes parallelism as running “computations on two processors at the same time.” Parallel computing is more general — it is the idea of running computations on multiple processors simultaneously, not just two.

Article Tools

They told him that it wouldn’t work. “When we went to conferences, they did not believe that complex digital circuits could reliably operate at low voltages.”

Then a graduate student at the University of California, Berkeley, Professor Anantha P. Chandrakasan wasn’t entirely sure it would work either.

In the end, though, his low-power chip designs did work, and one year later, Chandrakasan gave one of the most influential talks in the history of the International Solid-State Circuits Conference. So many people attended that some couldn’t even get close enough to hear, so the organizers asked him to give the talk a second time — a first for the conference.

A couple of decades later, Chandrakasan holds the chair for the International Solid-State Circuits Conference, held the directorship of MIT’s Microsystems Technology Labs, and has been recently appointed as the head of MIT’s Department of Electrical Engineering and Computer Science (EECS).

But how did it all start?

Chandrakasan’s interest in circuits was sparked in his college years. When Chandrakasan was a senior at Berkeley, he began working with Professor Robert Brodersen, who became his thesis advisor.

Brodersen had an idea for a device: a multimedia pad that people could carry around and use to access text, graphics, video, and audio. This was almost two decades before the release of the iPad, and much of the technology required to make such a device didn’t exist, so Brodersen and his research team had to come up with much of it themselves.

The project would later be known as the “InfoPad,” and would grow to incorporate over 50 students, faculty, and employees from several departments and companies before coming to an end in the mid-’90s.

Brodersen’s device was designed to connect to a network and exploit its computational resources. But certain processes would still have to be computed locally on the device, requiring chips that were powerful enough to perform them but efficient enough to not drain the battery. There was only one problem: these chips didn’t exist. For the InfoPad to work, someone would have to invent them. This was the task Professor Brodersen gave to then-grad students Chandrakasan and Samuel Cheng over the summer of ’91.

Chandrakasan and Cheng spent that summer trying to find a solution. In the end, they found one: parallelism — run computations on two processors at the same time to conserve energy. When the team tested their designs and they worked, Chandrakasan knew it was something huge. He co-wrote “Low-Power CMOS Digital Design,” submitting it to the IEEE’s Journal of Solid-State Circuits. The paper became the second-most cited paper in the journal’s history.

After getting his PhD in 1994, Chandrakasan left Berkeley to join MIT as a professor. He became the director of MIT’s Microsystems Technology Labs in 2006, and in July he replaced Chancellor Eric Grimson PhD ’80 as head of MIT’s Department of Electrical Engineering and Computer Science.

In charge

Heading an academic department is no small task, let alone heading MIT’s largest. “The thing about being a department head is that you have to touch pretty much every aspect of the department,” said Chandrakasan.

That means overseeing undergraduate education, graduate education, and different subgroups and tasks; getting to know and supporting faculty; preparing cases for faculty promotions; and interfacing with research labs, just to list a few.

And on top of that, Chandrakasan still heads his research group, Terminal 2020, which is trying to answer the question, “What is the iPad of the future?” Plus, Chandrakasan continues to be an advisor for undergraduates.

“Any hobbies?” I asked him.

“MIT,” he responded jokingly.

As the new EECS head, Chandrakasan wants to get one point across: he really wants to connect with students. “Any time a student wants to meet with me, I’m completely open,” he said.

“We want to have students — undergrads and grads — really work with us, provide suggestions, be able to walk into the department head’s office or go to the gym and suggest ideas. I would be very excited about students engaging in the implementation of new initiatives.”

The professor even goes as far as meeting students during his morning workouts at the Z Center around 7:45. “Actually, sometimes I invite some of my grad students. … We’ve had meetings to discuss research ideas, so if students want to find me there, that’d be a great way.”

Chandrakasan also accepts friend requests on Facebook — just as long as they’re from people at MIT. “There was a recent press article, and then a whole bunch of random people added me and I didn’t accept,” he recalled. “But if any of our students want to friend me, I’d be happy to accept.”

There is definitely more to Chandrakasan’s life than MIT. Chandrakasan has three children — two girls, five and eight years old, and a son who is in high school. Outside of MIT, Chandrakasan does everything from taking them to Red Sox games to watching Khan Academy videos with them.

“When I’m not here [at MIT], I like to spend time with my family,” he said. “I just like to hang out with them.”

What’s the newly-minted department head’s best advice for current students? “Really be passionate about something, try to take lots of classes, and try to use every opportunity you can at MIT. There are so many different things that you can do, but try to pick the ones that you get most excited by and follow through. And have an open mind in terms of what kind of careers you could have.”