IBM Research scientist Robert Dennard, who at age 80 still comes to work at the lab nearly every day, has been awarded the Kyoto Prize—one of the world’s most prestigious recognitions for personal achievement. He will receive the Advanced Technology Prize in Electronics at a November ceremony in Japan.
Dennard, an IBM Fellow, is best known for inventing the memory chip in 1967. The simplicity, low cost and low power consumption of his invention, dynamic random access memory (DRAM), opened the door to the personal computer. Today, memory chips are used in every PC, laptop computer, game console and mobile communications device.
“I knew it was going to be a big thing, but I didn’t know it would grow to have the wide impact it has today,” Dennard said in an earlier interview. He was awakened at 1:48 a.m. today by a call from Japan informing him he had won the prize. “I’m feeling quite elated,” he declared a few hours later.
In addition to DRAM, Dennard also pioneered the theory of device scaling—figuring out how to shrink microcircuits in every dimension. He and his team laid out the technical requirements for making memory chips smaller and faster, principles that were also applied to processor chips. Those rules provided a recipe that chip designers have followed ever since. Today, thanks to Dennard’s insights and much innovation by generations of scientists, billions of transistors can now be placed on a single chip.
Electronics industry pioneer Gordon Moore is famous for predicting in the 1960s that the density of microcircuits would double every two years—a pronouncement that has since been codified at Moore’s law. The work on scaling that Dennard launched has enabled Moore’s Law to hold for decades.
“For a guy to produce one of these two breakthroughs is extraordinary,” said Bernard Meyerson, IBM’s vice president for innovation. “Two from one guy is ridiculous.”
IBM Research Senior Vice President John E. Kelly III congratulated Dennard, and said he’s an exemplar of one of IBM’s core values–producing innovation that matters for IBM and society. “Bob is the consummate Wild Duck” he said. Kelly was referring to IBM’s tradition of encouraging scientists and engineers to try out unorthodox ideas and approaches to solving problems.
Sometimes the Wild Ducks really go it alone. That was true for the DRAM. The idea underlying his invention came to Dennard in 1966, when he was sitting on a living room couch in his Westchester County home. Earlier in the day at the IBM Research lab in Yorktown Heights, N.Y., he had listened to another research team discuss improvements they planned for magnetic core memory, a technology that was used in mainframe computers. He was impressed with the simplicity of their design. His approach for is his own project, designing memory technologies for metal-oxide semiconductors, was quite complex—using six transistors for each bit of information.
On the spot, he came up with a technique for using a single transistor for each bit of data. He drew sketches on a pad of paper to help him visualize how it would work. That was the breakthrough idea that paved the way to the invention of DRAM. He developed it further in his spare time while working on other “official” projects. The patent was issued in 1968.
But Dennard said he didn’t usually work by himself. “I’m not really a Wild Duck. Most of my work was done with teams of guys, which is really the most powerful way of solving problems,” he said. “You want to have groups of people with different kinds of knowledge, working together and sharing ideas.”
These days, Dennard is working on ways to improve today’s primary chip technology, Complementary Metal-Oxide Semiconductors, or CMOS. Because of the limitations of classical physics, chip designers are hard-pressed to achieve the kind of overall performance gains that were possible in the past. The problem, he said, is power consumption. Today’s chips consume too much electricity and throw off too much heat. So he’s working on new approaches to architecting chips- using extra-thin layers of silicon. “When you’re building the last generation of a technology, you want it to be the best,” he said.
Dennard received his B.S. and M.S. degrees in Electrical Engineering from Southern Methodist University and his Ph.D. from Carnegie Institute of Technology—now Carnegie Mellon University. He credited CMU with training him to approach complex problems in a systematic way.
One of his early jobs was at a Texas electrical utility, and he remains interested in energy issues all these years later. He’s concerned about society’s dependency on fossil fuels–and the effect on the climate. “We’re not only burning ourselves up; we’ll run out of the fuel to do it.” Dennard is enthusiastic about the potential of wind and solar power. And he figures he’s doing his part for the environment by trying to decrease the energy consumption of chips.
He has received many other awards during his long career, including the U.S. National Medal of Technology, in 1988, for his invention of the DRAM, and the Charles Stark Draper Prize from the U.S. National Academy of Engineering. The Kyoto Prize is presented by the Inamori Foundation, and was first awarded in 1985.