To God, there is no zero. Robert Matheson in the Incredible Shrinking Man.
That’s how Pat Gelsinger, Intel vice-president and chief technology officer
started his address to over 600 developers at the Intel Developer Forum in
Bangalore this May. Well, God’s prophet never came in a more unassuming form
and developers never listened to a messiah in more rapt attention. A 20-year old
veteran at Intel and among its youngest vice president at 32, Pat is as big as
it gets for developers. His bottom line message: Not only is Moore’s law alive
and kicking but Intel is well on its way to extending and expanding it beyond
limits earlier imagined.
|
Gordon Moore, now chairman emeritus of Intel, was asked by an interviewer to
predict microprocessor trends (there were then only 30 components on a chip).
Moore saw 60 components on a chip in his lab and in a complete leap of faith
predicted—"I blindly extrapolated" as he said in his recording for
Gelsinger’s presentation—60,000 components in 10 years. It turned out to be
a lot more precise than I estimated." A passing observation turned into law
for the technology industry—both technologically and economically. Today,
Gelsinger says, "Our technologies have touched a billion people but we till
have five billion to go—one billion in India."
The road here and beyond
It had not been an easy journey. Gelsinger recounts how he and Avtar Saini
(Intel India head who was working with Gelsinger) for years labored for speeds
of 25 megahertz which they pushed to 50 Mhz. "Now we add 25 mhz a
week," Gelsinger said, "and it won’t be long before we add 25 mhz a
day." Similarly, the big question in the 1980s was could Intel possibly go
past 1 micron? In the 1990s it got pushed to 0.1 micron. "To us
today," said Pat, "There is no end to Moore’s law."
Among other things, Intel’s mid-term forecast includes Silicon Radios. It’s
a vision of what Gelsinger calls "Radio Free Intel"—essentially,
radios that are "free, integrated, ubiquitous and part of every chip we
build." Ongoing areas of research include MEMS or micro electro-mechanical
systems that use mechanical instead of the electrical properties of silicon.
Gelsinger says Intel will apply Moore’s law in the realm of silicon photonic
research "bridging the gap between photons and electrons" and building
highly integrated components that combine digital logic functionality with
silicon-based opto-electronic devices on a single chip. One the biggest issues
with optical technologies today is the prohibitive costs and Intel hopes
research in silicon photonics will reduce those costs.
Betting on technology
Ultimately, the vision Gelsinger painted for developers is one that lies at
the core of Intel’s philosophy—that technology waits for no one.
"Always bet on technology," Gelsinger told Dataquest. "Or you’re
setting yourself up for a fall if not a demise."
That is not an easy vision to uphold in the face of technological and
business issues. "Sometime ago I spoke of power being the fundamental
stumbling block to Moore’s law. But we’ve had a lot of breakthroughs in the
past few years, some of which are hyper-threading, terrahertz technologies and
low power high performance circuits.
Lately Intel has been accused of releasing new technologies much before the
product and market are ready. For instance, Itanium I was released with compiler
issues still unresolved and in the thick of an economic recession. Many
predicted, and quite correctly, that it’s off take wouldn’t really be
spectacular. So is Intel happy with how the way Itanium performed? "Well,
yes and no," says Gelsinger. "Yes - because it is being accepted, we’re
getting customers on it and we’re seeing applications ported to it. No -
because it came out later than it was supposed to in the marketplace and it hasn’t
quite ramped up at the speed that we expected it to." However, Intel felt
there were numerous needs that forced the release of Itanium (at the time it
did) but "the need to get the industry started" was most strong.
According to Gelsinger, sometimes it actually made business sense not to
introduce a new technology. "When we introduced the Pentium, the 486
actually made better business sense for us." It would not, however have
been in line with the company’s vision of technology. With or without a
downturn, he said one can look forward to an infinite future with infinite
resources and infinite possibilities.
Sarita Rani in Bangalore