Ask DR Ambuj Goyal what was special about Deep Blue, the IBM su- percomputer that in 1996 defeated the reigning chess champion Gary Kasparov, and he says modestly “It was not , that sophisticated.” As someone who believes supercomputers need to be “not 10 or 100, but 1,000 times faster” than they are today to solve critical healthcare problems, Goyal can probably afford to say that.
He explains, “There were three main things work ing behind the Deep Blue — its capacity to look ahead at several situations, the algorithms developed by playing scores of games against masters , and a database of opening moves and end games.” Goyal did not invent the Deep Blue. What he, as a member of the IBM Research team that was interested in Game Theory, did was to champion the supercomputer’s use in engaging the world chess champion.
Goyal, who now heads information management at the IBM Software Group, says, “I invested in the concept of solving the ‘grand challenge’ problem. It was an amazing experience. What is even more exhilarating is that it encouraged many at IBM to take up other challenges and not to give up easily.” This ‘lifer’ at the Big Blue, who in his quarter century at the multinational has seen its revenues grow three-folds to almost $100 billion, takes that to be the bigger gain.
Goyal joined IBM in 1982 right after finishing his PhD in electrical engineering at the University of Texas , Austin. He says, “I wanted to work for a company that encouraged innovation, a company which invested in its people. I wanted my sandbox. IBM turned out to be the largest one I could play in.”
In the ‘sandbox’, Goyal has distinguished himself with quite a few notable playthings. His main research interests are in high-performance systems , databases and distributed systems. His work has led to IBM’s Universal Database family , WebSphere servers, and RS/6000 super computers. At present, some 1,500 researchers from seven IBM laboratories around the world report to him.
It is this diver sity of talent that still excites Goyal, a five-time winner of IBM’s presti gious innovation award. He says it has helped the century-old com pany reinvent it self time and again — “from selling meat slicers, to punch cards, to comput ing, to services”.
This push to take up the next challenge gave birth to the Blue Gene project five years ago. Its goal is to design a supercom puter that would help us un derstand the mechanisms behind protein folding through large-scale simulation, and explore new ideas in ‘massively parallel’ com puting. Goyal says, “Down the road, this research will help create drugs that are specifically designed to work for an individual’s genetic make-up.” In a life full of accolades, that would be his greatest treasure.