RCFoC for March 5, 2001 - From the Bottom -- Up.

CPU Update (Not What You Think...)

"I have never ceased to be amazed at the progress we have been making, and the astonishing rate at which it has all been happening. What was only a dream yesterday turns into a reality today -- and tomorrow would seem to promise even greater rewards."

J. Fraser Stoddart, Saul Winstein, Professor, UCLA

You might expect, with a title and a quote like that, that we're about to explore the latest and greatest and fastest and lowest-power multi-million transistor CPUs, such as NEC's new "Vr4131" -- a specialized 64-bit RISC chip designed for embedded applications and pocket devices, which runs at 1.545 billion instructions per second while consuming only 1.5 watts of power (http://www.infoworld.com/articles/hn/xml/01/02/05/010205hnnecchip.xml?0206tuam).?

Or, perhaps, this might be about Compaq's new 833 MHz general-purpose 64-bit Alpha chip, the 21264, or its even faster 1GHz samples, which are now being made (http://news.cnet.com/news/0-1003-200-4797930.html?tag=mn_hd).?

Or it might be about the new 1 GHz notebook chip from Transmeta, which draws less power than conventional notebook CPUs (http://www.techweb.com/wire/story/TWB20010202S0006), or about Intel's new 700 MHz notebook Pentium III chip with "SpeedStep" technology (which reduces its speed, and power consumption, when running on batteries - http://www.intel.com/pressroom/archive/releases/20010227comp.htm).

Or, maybe this article relates to Intel CTO Pat Gelsinger's honey-and-vinegar announcements that, by 2010, 1.8 billion transistor chips (up from earlier estimates of 1 billion transistors) may be running at 30 GHz (up from previous estimates of 10 GHz) and executing one trillion operations per second (ten-times faster than previous expectations of a "mere" 100,000 MIPS). That certainly does sound like "honey," but the "vinegar" is that with today's techniques, these incredibly fast chips would also be producing "as much heat, for their proportional size, as a nuclear reactor." Which means that we do have to find a better way. (http://www.zdnet.com/zdnn/stories/news/0,4586,2681998,00.html and http://www.eetimes.com/story/OEG20010205S0083)

But it's not. So go read the latest edition of the Rapdily Changeing Face of Computing. Hint: Professor Stoddart is a professor of organic chemistry.