September 08, 2003
Musings at MIT
Please note that contributed articles, blog entries, and comments posted on EDACafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.
The best thing about MIT is the Charles River.
The mile-long campus of the Massachusetts Institute of Technology is both defined and constrained by the gently flowing waters of this ancient river, which grows broad as it passes under the bridge at Mass Ave - the portal to Cambridge and MIT.
Within its dark stately flow, the leafy conceits of Harvard and the urban sensibilities of Boston University mingle with the ascetic intellects of MIT and the tasteful energies of Boston's Back Bay. The river moves on downstream and adds the gritty working class accents and vaguely puritanical and democratic ideals of downtown Boston to the flow before joining with the Mystic and passing grandly through Boston Harbor and on out to the chilly waters of the Atlantic and rugged realities beyond.
MIT is a place within context, anchored to the soil of Cambridge by the silent, heavy symmetry of the dome and its pillared, orderly attendants that evoke the names of Archimedes, Newton, Franklin, Brahe, Gutenberg, Plato, Des Cartes, Fulton and others. The dome wants you to believe that this place is a shrine to mathematicians, scientists, and engineers stretching back through history - male, confident, and proud.
But don't be fooled by the public face of MIT. Behind that classical façade, it's a place of chaos and industriousness, disorder and irregular, non-linear progression toward real solutions to real problems. It's a place where the elegant proof meets the unruly world and designs are optimized within an acceptable tolerance level.
It's a campus that was founded on applied technology, where the basic sciences arrived late, and where a hostile take-over bid by erudite Harvard in the early 20th century failed because MIT was unwilling to play second-fiddle to cerebral theory and the ivory tower. It's a place where people with a passion for the sciences are mandated to get their hands dirty. It's a place where boys can be boys - and now, happily, girls can be girls - if that means thinking and scrambling and making a mess in order to understand what's what in the world around us. It's a place where there are six major areas of study numbered I to VII.
On a rainy September morning, I found the Computer Science Department at MIT and Professor Srinivas Devadas by wandering past the disarray of the emerging Stata Center on campus, soon to be home to both the EE and CS Departments. The disarray of Stata is not due to the army of construction equipment and people vainly attempting to meet their year-end completion deadline for the building. The disarray is designed right into the structure, which is suffering - or celebrating - from a distinct lack of vertical lines or right angles. Very MIT.
When I reached Dr. Devadas a block later in Technology 100, he was hunkered down over his computer trying to sort out a Windows problem, which was a nice way to meet one of the best and brightest of MIT and EDA.
Good to know that accolades and accomplishments notwithstanding - PhD in CS at the tender age of 24 from U.C. Berkeley, NSF Young Investigator Award (1992), Best Paper Awards from both IEEE Transactions of VLSI Systems (1996) and Computer-Aided Design (1990), Best Paper Award at DAC (1998), Fellow of the IEEE (1998), author of Logic Synthesis with Ghosh and Keutzer, and now Chair for MIT's Area II - Srinivas Devadas suffers from the same, mundane irritations that the rest of us enjoy when dealing with the sloppy code that emanates from Bellevue, WA.
Dr. Devadas has been on the Computer Science faculty at MIT since he finished at Berkeley in 1988. During his time at Berkeley and in the years following, he was extremely busy, along with his peers, sorting out many of the challenges in computer-aided design of logic systems and their ilk.
Since the late 90's, however, Devadas, has wandered away from his early passion for EDA and today teaches undergraduate classes and sponsors graduate research that more closely mirror his current interests. He's busy teaching what he's preaching - that a strong foundation in math, computer science, and electronics are all necessary to understand the modern world.
He told me: “In the early days it was really exciting in EDA. Cadence was just starting, and Synopsys as well. When I started, people were working on logic design, layout, schematics, RTL, behavioral synthesis. It was beautiful from an academic point of view. There were a good 10 years in there, but those days are past.”
about the aesthetics of the problem for me, and it's just not there [anymore in EDA].”
“In academics, you define a problem and work on challenges that are not necessarily important to industry. Once a problem is established, industry will take over. In academia, however, it's nice to enjoy the problem for a while [before that happens].”
“So I decided I didn't want to work on [existing] CAD problems anymore, but wanted to work on a high-performance processor, which has its own painful set of problems. But for a number of reasons, I decided instead to work on building a processor that isn't high performance or low power, but is a secure processor - one that would protect against network attacks and copyright issues. This is [allowing me to address] problems that are broader than just EDA. It's nice to do EDA for that particular type of system, but there are generic EDA tools available - the companies have done a good job of providing useful tools to support the work.”
“I took a year's leave, 2000 to 2001, to work at Sandburst Corp. [Devadas served as Principal Engineer at the fabless start-up founded by MIT's Arvind, Johnson Professor of Computer Science and Engineering.] However, for many reasons, I just didn't enjoy it. We've got VCs walking around the halls all the time here at MIT, but after my [recent] experience in industry, now I just get up and shut my door when I see them running by. I made the decision when I came back to MIT in 2001 that I would be a scholar for another 10 years before I ever considered leaving again.”
“[Meanwhile], after Jonathan Allen died [Professor and Director of MIT's Research Lab of Electronics], our VLSI design course kind of went dormant - from 1999 until 2002 or so. It's true that VLSI design can be seen to be somewhat of a trade skill, perhaps inappropriate [to some] for the atmosphere at MIT. However, we know that the subject is important and we're in the process of reviving the course. We're making the course more modern, to teach about FGPAs and new synthesis tools. These are important things for people to know.“
“Meanwhile, I've decided to concentrate on teaching undergraduates - in particular, freshman and sophomores. We've got very sharp undergraduates here at MIT and I really like the younger students. They're not [jaded] like the older students.”
“I'm teaching a basic course on Java, and one on Boolean logic, and one on hardware. I've become interested in the power of the material and am interested in how to teach students to think [simultaneously] across all three disciplines - mathematics, software, and hardware. If students don't think about math, software, and hardware at the same time, they're starting off [on the wrong foot]. Right now the CS department and EE departments are in different buildings on campus. When we're all [housed together] in the new Stata Center, it will help us to think about embedded systems, across hardware/software [boundaries], reusable IP, behavioral synthesis, SystemC.”
“Right now, I'm seeing a lot of crummy CAD coming out of the systems area and the conferences aren't helping. The CAD guys go to their own conferences and have good CAD, but crummy designs. The Design guys are off at their own conferences and have good designs, but crummy CAD. People see the same faces at their own conferences, a small group of guys that they're comfortable with. They [resist] going to conferences where they don't know anybody and they're not known. I want to go to conferences where people don't know me and I'm trying to do that more these days.”
You can find the full EDACafe event calendar here.
To read more news, click here.
-- Peggy Aycinena, EDACafe.com Contributing Editor.
Be the first to review this article