Rating is available when the video has been rented.
This feature is not available right now. Please try again later.
Published on Dec 2, 2015
We have all been following the dictum of Moore's Law for longer than most engineers have been alive. Our focus on functionality, performance, and economics has yielded remarkable systems, from today's smartphones to supercomputers and internet servers, to engine controllers, anti-lock brakes, airbags, and navigation systems in our vehicles.
But software is notoriously buggy, and hardware isn't defect-free, either. Worse, we seem to only take security into account when some hacker perpetrates yet another outrage. The pending end of Moore's Law will greatly diminish the historical demand to replace systems every few years, which will put new pressure on system aging effects.
Performance and efficiency enable new applications, and in the past, the computer design community could simply stop when those targets were within reach. In this talk, Bob Colwell (formerly of DARPA and INTEL) argues that we cannot get away with that any longer, and need to turn our attention to areas of system design that are not good enough for how those systems will be used in the near future. He also discusses ideas on how to accomplish this without incurring additional development or product costs.
Bob Colwell, former director, DARPA's Microsystems Technology Office, former chief IA32 microprocessor architect, Intel