Gordon Moore
“It’s not a law; just an observation.” So claimed Gordon Earle Moore, who, with others, founded Fairchild Semiconductor. But without that “law” the telecommunications industry would be decades less advanced. Consider many of the applications now taken for granted: conference calling, caller ID, speakerphone, and most especially cell phones and wireless networks and DSL. Oh, maybe we would have some of them, but the speakerphone apparatus would fit in a shoebox rather than in a tiny telephone, and cell phones would continue to occupy the major part of an automobile trunk. Who knows what a DSL router would look like?
But what, exactly, is Moore’s Law?
Back in 1965, when Moore was Director of Engineering at Fairchild Semiconductor, he noticed that semiconductor chips were getting smaller, and smaller, and smaller. In fact, the number of electronic appliances – transistors, capacitors, resistors, diodes – that were crammed onto a sliver of silicon seemed to be doubling every 18 months. (Patents for the integrated circuits were first filed in 1959.) Moore predicted that this would continue well into the future. He published this “law” in an article, ”Cramming More Components onto Integrated Circuits” in Electronics magazine on April 19, 1965. The article was only 3 and 1/2 pages long.
Gordon Moore was born in San Francisco, California, on January 3, 1929. He grew up in the San Francisco area, and at age 10 his family moved to Redwood City. It was here that he first became involved with science and engineering; his next-door neighbor had a chemistry set. The two had a grand time with the materials that were then available, and Moore quickly built his own home lab.
Moore was not an outstanding student, but his grades were good enough to get him in San Jose State for two years, then Berkeley, and finally graduate school at Caltech. (He was, as a matter of interest, the first member of his family to get a college degree.)
His doctoral thesis dealt with the infrared spectra and structure of a few simple molecules, and he was so sure that this work would provide direction for the rest of his career that he took a research position at John Hopkins University’s Applied Physics Laboratory.
In a 2002 interview with Caltech News he said, “I found myself calculating the cost per word of the articles I was publishing at APL and wondering if, at $5 per word, the taxpayers were getting their money’s worth. It was time to look for something more practical.”
That more-practical job was a stint at Shockley Semiconductor Laboratory in Palo Alto.
It was in the early ‘50s that great progress was being made with the transistor. The little device had been developed in late 1947 by William Shockley, Walter Brattain, and John Bardeen. In 1955 Shockley left Bell Labs to form Shockley Labs, and he attracted a number of the best and brightest; Moore was one of them.
Things didn’t go too well at Shockley Labs, and Moore is quoted to say that Shockley was responsible for two important changes in his life: “He got me into semiconductors, and he unwittingly gave me the push to go elsewhere.”
In October, 1957, seven employees including Moore and Robert Noyce left Shockley Labs to form Fairchild Semiconductor. Eleven years later – July 16, 1968 – the two left Fairchild to form Intel – the company today known throughout the world for its work in the semiconductor field.
Gordon Moore was named President and Chief Executive Officer of Intel in 1974, and later assumed the role of Chairman of the Board. In May 2001 he was named Chairman Emeritus, and left the board.
But what about the so-called "Law?" Does it remain valid? There are occasionally questions – after all, there is a limit to how small things can get. Isn't there?
Back in 1978 Intel introduced the 8088 integrated circuit. The device was home to 29,000 transistors. That number was, and is, astounding. But it certainly wasn't the end. The top-of-the-line Pentium integrated circuit, in June 1996, carried 3.1 million transistors. By 1999 the number had passed 10 million, and by 2003 100 million. By 2006 the number was 125 million transistors. And in 2007 the industry experienced a real breakthrough: a different element – hafnium – was employed to cut the size of the transistor even more.
These numbers are so big, and so small, that a bit of perspective is needed:
• If a house shrunk at the same pace transistors have, you would not be able to see a house without a microscope.
• The price of a transistor is now about 1 millionth the average price of a transistor in 1968. If car prices had fallen at the same rate, a new car today would cost about 1 cent. One scientist calculated that the cost of a single transistor on a chip cost only 1/10 as much as the staple holding 2 pieces of paper together.
• More than 2 million transistors could fit on the period at the end of this sentence. This has resulted in a density of 30 million transistors on a chip the size of the head of a pin.
• Modern transistors can switch on and off approximately 300 billion times a second. A beam of light travels less than 1/10 tenth of an inch during the time it takes the transistor to switch on and off.
Moore's Law lives on!
Endnotes
Definition of Moore's Law: "The number of transistors and resistors on a chip doubles every 18 months." By Intel co-founder Gordon Moore regarding the pace of semiconductor technology. He made this famous comment in 1965 when there were approximately 60 devices on a chip. Proving Moore's law to be rather accurate, four decades later, Intel placed 1.7 billion transistors on its Itanium chip. In 1975, Moore extended the 18 months to 24 months. More recently, he said that the cost of a semiconductor manufacturing plant doubles with each generation of microprocessor.
Source: June 1, 2009: http://www.pcmag.com/encyclopedia_term/0,2542,t=Moores+law&i=47229,00.asp.
Transistor Count Growth from 1971 Through 2006
Processor Transistor Count Date of Introduction Manufacturer
Intel 4004 2,300 1971 Intel
Intel 8008 2,500 1972 Intel
Intel 8080 4,500 1974 Intel
Intel 8088 29,000 1979 Intel
Intel 80286 134,000 1982 Intel
Intel 80386 275,000 1985 Intel
Intel 80486 1,200,000 1989 Intel
Pentium 3,100,000 1993 Intel
AMD K5 4,300,000 1996 AMD
Pentium II 7,500,000 1997 Intel
AMD K6 8,800,000 1997 AMD
Pentium III 9,500,000 1999 Intel
AMD K6-III 21,300,000 1999 AMD
AMD K7 22,000,000 1999 AMD
Pentium 4 42,000,000 2000 Intel
Itanium 25,000,000 2001 Intel
Barton 54,300,000 2003 AMD
AMD K8 105,900,000 2003 AMD
Itanium 2 220,000,000 2003 Intel
Itanium 2 with
9MB cache 592,000,000 2004 Intel
Cell 241,000,000 2006 Sony/IBM/Toshiba
Core 2 Duo 291,000,000 2006 Intel
Core 2 Quad 582,000,000 2006 Intel
Dual-Core
Itanium 2 1,700,000,000 2006 Intel
