| IBM Research: Press Resources | |
Return to Web version |
IBM Fellow Emeritus
A native of Gwalior, India, which is located near the Taj Mahal city of Agra, Agarwal received a Bachelor of Technology degree in electrical engineering from the Indian Institute of Technology in Bombay, where he also received The President of India Gold Medal.
He traveled to the United States to earn first his M.S. (in information theory) and then his Ph.D. (in electrical engineering/digital signal processing), both from Rice University. He won the Sigma Xi Award for the best Ph.D. thesis in Electrical Engineering and an IEEE Signal Processing Society Best Paper Award (1974) for his published doctoral papers
While a post-doctoral researcher for IBM Fellow Richard Garwin in the mid-1970s, Agarwal developed an algorithm that helped determine the three-dimensional molecular structure of insulin from X-ray diffraction data that reduced computing times 50- to 100-fold over conventional methods. Agarwal's algorithm made it possible to refine large protein molecules and is still being used in many laboratories around the world.
In 1978, he returned to India for family reasons and rejoined IIT Delhi as an associate professor, where he had worked earlier before getting his doctorate. But tragedy struck in 1981. An out-of-control car plowed into Agarwal's scooter and he lost his right leg. In the aftermath, Agarwal decided to return to research at IBM.
Working again with Garwin, Agarwal helped the National Academy of Sciences analyze the acoustic tapes related to the assassination of President Kennedy to learn if there was any evidence for a second gunman. Agarwal observed that the recordings containing the so called "shots" were made about a minute after the assassination, and eliminated the possibility of a second gunman or the conspiracy that had been built up around that conjecture.
In 1994, he analyzed the floating-point divide flaw in the Intel's Pentium chip and showed that the probability of the divide error increases by several orders of magnitude for spreadsheet calculations using decimal numbers.
His primary research has been in algorithms and architecture for high performance computing on RISC workstations and scalable parallel machines. specific research topics have included developing efficient computational algorithms that:
