Where Microprocessor Technology is Taking Us Next

With new, fast-developing technology, designers can generally steer the technology towards the most lucrative ends. The history of the computer business has been one of architecture franchise owners (IBM, Intel, Microsoft) plowing huge profits back into technology developments aimed at expanding the franchise still further. Given the general proclivity of computer system buyers to reward such standardization, the only thing that could break this cycle is physics. And that is where we stand today: the basic physics of chip heat extraction have finally forced Intel to publically renounce their historical drive for much faster computers in every new generation. Power dissipation will no longer allow that game to move ahead. So Intel and AMD have announced a new multiple-core-on-die scheme, which they say will more than keep the industry on its habitual Moore's Law every-two-years fix. What's behind this multicore idea? What are the unknowns? Where will it lead, and do we really want to go there?
Dr. Colwell was Intel’s chief IA32 microprocessor architect, 1992-2000 and managed the P6 and Pentium 4 projects. In 1996 he was named an Intel Fellow, has published many technical papers and journal articles, inventor or co-inventor on 45+ patents, and has served on numerous panel sessions and invited talks. He is the Perspectives Editor for IEEE Computer Magazine and writes the At Random Column. He is currently an independent consultant.