Forty years of Moore's Law
Forty years ago this month, Intel co-founder Gordon Moore made a historic prophecy that foretold how a bucolic valley of plum and cherry...
Knight Ridder Newspapers
SAN JOSE, Calif. — Forty years ago this month, Intel co-founder Gordon Moore made a historic prophecy that foretold how a bucolic valley of plum and cherry orchards would be transformed into Silicon Valley.
The number of transistors on a computer chip had and would double every year, observed Moore in a paper published April 19, 1965. A decade later, he revised what had become known as Moore's Law: The number of transistors on a chip would double every two years. The prediction was an elegant statement of how semiconductor chips would become cheaper, faster, smaller and more reliable over time.
The result is a world unimaginable four decades ago: Computers once the size of refrigerators fit in the palm of your hand. A $40 talking Furby toy packs four times the processing power of a 1960s Apollo moon lander. A cellphone that cost $2,000 in 1984 and had all the sex appeal of a brick can now be had for a hundred bucks and slips into your pocket.
It is why technology has infiltrated the lives of everyday people, from iPods and Xboxes to camera pills that wirelessly transmit photos of patients' digestive tracts. We don't have flying cars yet, but from a 1965 vantage point, the present looks like "The Jetsons."
To put it another way, in 1971 Intel was able to put 2,300 transistors on its first microprocessor. Later this year, the Santa Clara, Calif., chip giant will unveil a processor crammed with 1.7 billion transistors.
"If the automobile industry moved this fast, your car would move at a million miles per hour and it would get 50,000 miles per gallon," said Moore, 76, in a recent press conference.
But Moore's prediction turned out to be much more than a codification of the exponential speed of technological progress.
"It became an almost religious faith in human ingenuity and a belief in the future," said Carver Mead, the former professor at the California Institute of Technology who coined the term "Moore's Law." "It spoiled everyone into thinking that this would go on forever."
Now, some fear it won't.
By most accounts, Moore's Law will collide with the laws of physics sometime in the next 15 years. That's when it will become physically impossible to squeeze more transistors onto a single chip.
To keep Moore's Law alive, scientists and engineers are working to develop new technologies to enhance or replace silicon chips.
That race is another indication of how Moore's prediction introduced a sense of Darwinian competition to a valley that at the time was filled with orchards and just a few small semiconductor factories.
It put Silicon Valley on a technological treadmill: Companies either had to match the pace set by Moore's Law or suffer the consequences in the marketplace.
"Keeping up with Moore's Law required a tremendous amount of work for the engineers in the trenches. It set the bar for what you had to do, though it didn't say how you had to get there," said Stan Williams, who has toiled for more than a decade at Hewlett-Packard in Palo Alto, Calif., to find alternatives to current chip technology.
That sense of competition still keeps the valley innovating today, whether the competitors are down the street in shiny office buildings where the orchards once stood, or across the ocean in China or India.
As the 40th anniversary of the publication of Moore's paper approaches, the semiconductor business has grown into a behemoth with $213 billion in annual sales that powers the $1 trillion electronics industry.
Computing spinoffsThe chip industry has also laid the foundation for advances in many other areas, from the study of genetics to automobile safety, said Dan Hutcheson, president of market researcher VLSI Research.
"The Lewis and Clarks of today's time don't use optics to map the landscape," he said. "They use computer visualization tools to map the human genome."
Because Moore's Law keeps pushing chip costs down, it has been a deflationary force in the global economy.
The cost per transistor has fallen from $5.52 in 1954 to a minute fraction of a penny in 2004, according to a calculation by Hutcheson.
Thanks to this trend, the information-technology industry accounts for a quarter of the growth in the gross domestic product even though it is only 3 percent of the GDP, according to Dale Jorgenson, an economics professor at Harvard University. In other words, technological advances are key to keeping the economy going, despite the small size of the tech industry.
Moore's Law pumps up the economy as each new generation of ever-more-powerful chips spurs consumers to replace their "old" computers and gadgets with the latest versions.
All of the fuss over Moore's Law comes as a bit of a surprise to Moore himself.
He says he was embarrassed when his friend Mead coined the term "Moore's Law," and he couldn't bring himself to utter the phrase for about 20 years.
Moore said he recently reread the original paper, entitled "Cramming More Components Into Integrated Circuits," and was surprised that he predicted home computers would be one way the world would make use of all of the electronics that Moore's Law would make available.
"I wanted to get across the idea that integrated circuits were a way to make electronics cheap," Moore said. "I blindly extrapolated for 10 years continuing to double every year, from 60 components to 60,000, not thinking that it was going to be especially accurate but just trying to get the idea across that it would be significantly more complex and a lot cheaper."
Moore said he can foresee Moore's Law continuing for the next five years, but he can't really see beyond that and can't yet predict when his own law will stop describing advances in chip processing.
Intel and all chip makers have armies of researchers trying to find replacements for basic silicon chip manufacturing, which may not be needed until 2020.
Said Texas Instruments Chief Executive Richard Templeton: "Moore's Law is looking gray, but it's alive and well."