As the 2015 president of the IEEE Computer Society, Tom Conte, professor in Georgia Tech’s Schools of Computer Science and Electrical & Computer Engineering, is leading a national initiative to rethink how computers compute.
The initiative—“IEEE Rebooting Computing”—proposes a fundamental, holistic reexamination of the computer, including all aspects from device to user interface. The “reboot” is necessary because single-core processor performance stalled in 2005. The hardware industry created multicore processors, but these have limitations.
Now, even multicore processors are having performance issues because of the anticipated end of Moore’s Law -- a widely accepted observation that the amount of transistors on an integrated circuit would double every two years. Regularly increasing a circuit’s transistor capacity helped computers achieve faster processing speeds, but increases in transistor speeds have only been marginal for years. Resultant, necessary shortcuts have become ineffective.
“Until now, we’ve always been able to proceed anyway, knowing computers would go twice as fast in 18 months,” Conte said. “We cannot rely on that any longer.”
That’s why IEEE chose to act.
IEEE’s Rebooting Computing Initiative and its collaborator – the industry-led International Technology Roadmap for Semiconductors (ITRS) – have thus far hosted four summits of government, industry, and academic thought leaders in Silicon Valley and Washington, D.C. Summit participants focused on the three pillars of future computing: what computers will be used for (i.e., human/computer interface and applications), how energy-efficient computers can be made, and how secure they can become.
“We hope to influence industry and policymakers to change the direction of computing,” Conte said. “We’re trying to make people realize that they cannot be complacent; we have to act and find a solution.”
The possible solutions are a bit unconventional: Allow computers to produce approximate results, rather than computing to the customary 100th decimal point? Use non-digital computation? Or use models of the human brain in hardware? Some industry leaders consider all of these postulations “crazy,” but Conte says all options should be on the table.
“Skepticism of any idea is not a luxury we can afford at this point in time, with the challenges we have,” Conte said. “The possibilities are vast, as are the problems. Changing the technology that has permeated virtually every facet of the human condition is not going to happen without enormous effort and investment.”
“I believe Georgia Tech has the right expertise at the right time and is uniquely positioned to help lead a response to this challenge today.”
Last revised July 8, 2020