Issue No.01 - January/February (2003 vol.5)
Published by the IEEE Computer Society
Mike Foster , DARPA
Information technology has profoundly transformed the world we live in and will inevitably continue to do so in the future in ways we probably can't even imagine. It could be the most revolutionary technology we've experienced since the steam engine touched off the Industrial Revolution. Underlying this revolution, though, is the silicon microchip.
The real power of the silicon microchip as a critical driver of the information revolution comes from its adherence to Moore's law. Although this principle has become shorthand for today's rapid speed of technological change, Moore's law, technically speaking, states simply that the number of transistors on a silicon chip will double every 12 to 24 months.
Intel's cofounder, Gordon Moore, made this observation in 1965, and it has proved remarkably accurate for almost four decades. Moore has jokingly stated that if similar progress were made in transportation, a modern-day commercial aircraft would cost US$500, circle the Earth in 20 minutes, and use just five gallons of fuel—but it might only be the size of a shoebox.
The semiconductor industry has used Moore's law as a benchmark for innovation and progress as well as for setting the pace for technological change—not just in semiconductors but also in applications of semiconductors to information processing and technology in general.
Moore's lesser-known second law declares that the capital cost of a semiconductor manufacturing facility will continue to rise each year, becoming so astronomical that a chip-manufacturing facility could cost up to US$40 billion or more by 2010. This means that it will become increasingly difficult for any one company to afford to produce its own chips, likely giving rise to joint ventures or consortia and more outsourcing of chip fabrication.
Moore's second law could in fact predict the demise of Moore's first law: capital costs will eventually reduce the number of efficient producers to the point where monopoly supply conditions prevail and costs will fall below the Moore's law curve. The other consequence of great interest is whether those plants will ultimately be in the US or in Southeast Asia.
Many economists say the development and deployment of semiconductors has been the foundation for current American economic growth resurgence. This isn't so surprising given the ubiquity of microchips today. Key components exist in all computing and communications equipment as well as in aircraft, automobiles, household appliances, consumer electronics, cell phones, electronic toys, scientific instruments, and myriad other products. Semiconductor manufacturing adds almost 1 percent to the US gross domestic product, in terms of value added, and is key in the US industrial base.
However, the International Technology Roadmap for Semiconductors ( http://public.itrs.net), a technical blueprint for future chip developments from the global Semiconductor Industry Association, states that we will reach the technological limits somewhere during the second half of the next decade, as microchip feature size approaches molecular dimensions.
Given the strategic role of Moore's law in information technology advances, US economic productivity, and the US military, we asked several experts to discuss the issue of Moore's law and the possible repercussions of its demise.
We hope you enjoy reading them, and we welcome any comments you might have on the intriguing possibilities inherent in contemplating the end of Moore's law.
Nancy Forbes is a freelance science writer in Washington, DC, and a contributing editor for CiSE and the American Institute of Physics' The Industrial Physicist. She has served as a consultant to the US government on emerging technologies of interest to the military. She has a BS and MS in physics from Columbia University and is a member of the American Physical Society and the Association for Women in Science. Contact her at firstname.lastname@example.org.
Mike Foster is a program manager at the Defense Advanced Research Projects Agency, with interests in computer architectures, quantum information processing, and tools for decision support. He has a PhD in computer science from Carnegie Mellon University. He is a member of the IEEE, ACM, and Sigma Xi. Contact him at DARPA, 3701 North Fairfax Dr., Arlington, VA 22203; email@example.com.