Somewhere on your person you’re probably carrying the heritage of one of the most insightful articles written on technology.The cellular phone in your pocket or purse, the BlackBerry on your belt and the PocketPC device in your jacket are all are powered by a microprocessor with muscle almost unimagined in 1965.
That was the year Gordon Moore, then director of research for Fairchild Semiconductor, submitted an article to Electronics magazine.
It was a time when integrated circuits were just starting to be commercialized after being largely restricted to the military.
With ICs offering high reliability and cost savings in manufacture and design, Moore foresaw that the number of transistors on a silicon chip would double every 24 months, and as that number increased the cost of making them would drop.
Integrated circuits would become mass products, and “lead to such wonders as home computers — or at least terminals connected to a central computer,” he wrote.
It took a few years to prove, but what has been dubbed Moore’s Law has come true, to the great benefit of people around the world.
Three years after the article was published Moore co-founded Intel Corp., whose first microprocessor, in 1971, had 2,300 transistors and cost about US$300. By 1978, its 8086 chip, which powered many an XT clone, had 29,000 transistors.
Today, CPU density is doubling roughly every 18 months — thanks in no small part to competition — to the point where an Intel Pentium 4 CPU has 250 million transistors, and costs about US$300.
Of course, Moore’s Law isn’t really a law, it’s a prediction. Nor was Moore the first person to make the observation. But his name has stuck to a guideline that is at the heart of the information technology industry.
The obvious candidates who have directly benefited from his insight: Intel, AMD and Texas Instruments among the mass semiconductor companies; makers of PCs and servers; and finally, software companies. They make a circle, pushing each other to take advantage of the technology.
‘Living off Moore’s Law’
In this country, Research In Motion, ATI and Celestica have built empires around Moore’s Law.
But there are many high-tech firms you likely haven’t heard of who have benefited indirectly, such as SiGe Semiconductor of Ottawa, a fabless semiconductor design company whose integrated circuits are used in wireless LANs and cellular phones.
“One of the benefits of Moore’s Law is that everything that used to come in a big box is in a cellphone,” observed Jerry Lorraine, the company’s chief technology officer. “A modern cellphone has a 2-megapixel camera, an MP3 player, can play video, do diary and e-mail and runs on a battery that lasts a few days.”
“We’ve been living off Moore’s Law,” said Patrick Myles, director of corporate communications at Dalsa Semiconductor of Waterloo, Ont., which makes high-performance digital imaging systems used in everything from the Mars rover to Hollywood movie cameras.
“Part of our job is to spit out data from our cameras. What’s allowed us to grow has been Moore’s Law because computers catch that data and then make decisions. If we were to keep coming out with newer cameras that spit out more pixels and there weren’t computer technology that could grab it, our company probably wouldn’t have grown.”
Forty years is a long run, but the laws of physics are catching up to Moore. As chips shrink, the space between transistors becomes smaller, increasing the risk data will leak across the circuit.
Chips that are 90 nanometres in length today are about to give way to 65nms, said Howard High, Intel’s strategic communications manager. In 2007 they’ll be only 45 nm long and then shrink even more.
“We have at least another 10-15 years of standard improvements that we can make,” he said.
Increasing megahertz to hike chip speed is another problem, as it also increases heat. That’s one reason Intel and AMD are coming out with dual core CPUs.
Eventually, however, exotic technologies now being researched, such as spintronics, backed by IBM, and crossbar latches, being worked on by HP, will have to replace silicon.
Opportunities
“The challenges are getting greater,” said Dale Gale, vice-president and chief technology officer of CMC Microsystems Corp., a non-profit Kingston, Ont., agency that helps universities and industries doing leading-edge microelectronics research. “And any time one encounters a challenge it opens up new opportunities.”
Over time, he foresees fewer products taking advantage of silicon and more custom-built solutions emerging using light-based (photonics), fluidics, micromechanical and nanotechnologies, niches which Canadian companies are now working in.
“The microelectronics technologies that (today) follow Moore’s Law will be more foundational,” he said.
Meanwhile the race to create faster chips for less money, and to extend Moore’s Law, will continue.
After all, says California-based chip analyst Nathan Brockwood, “Nobody wants to be the first guy to say ‘Oh, we can’t do this any more.’”
