Moore's Law has become one of the most reliable constants in an increasingly technological world. According to Forbes.com, there are companies that gamble their entire futures on Moore's Law, simply because it promises better performance per dollar per year, at least in theory. Originally, Moore's Law stated that the number of electronic components that fit on a silicon chip will double every 12 months. But it's been revised several times over its 40-year existence, so that now it is associated primarily with the number of transistors that fit on a chip, and the timeframe has crept up to 24 months. Moore himself said that his Law can't continue forever, and that eventually the exponential growth rate will end in disaster.
In an interview with Forbes.com, Bernard Meyerson, a chief technologist for IBM, said that people have been misinterpreting Moore's Law for years, and will continue to do so well into the future. According to Meyerson, the core tenet of Moore's Law is that you can double the amount of stuff on a chip in a certain amount of time, which has been interpreted as doubling the amount of stuff in a unit area. He believes that this view is short-sighted, as it doesn't account for vertical integration and chip stacks, which means that while the physical area remains the same, the size actually doubles.
In fact, Meyerson envisages a world far more complex than vertical stacks. When the Forbes interviewer asked where he saw the industry in five years time, he painted a picture of planes of super high-density memory above planes of logic, of multiple cores in a single level, and reconfiguring the wiring between chips stacked on one another. In other words, he saw the future in 3-D, which would add density but reduce costs, all while optimising the chips' performance at a fraction of the energy that is currently required.
He also predicted major advancements in integrated optics and the use of optical signals, as well as light. Optics and light both move at a significantly greater speed than electricity, and are not hindered by factors such as resistive capacity delay. Light would also be able to communicate data very quickly while using much less power. Transmission by light is also far simpler and less prone to distortion than current signals, which makes it ideal for use in data centres where people need to transmit data over large distances relatively quickly.
Meyerson cautioned, however, that none of these possibilities makes the development and manufacture of chips any easier. He said that integrated optics and 3-D stacking are as technically challenging as any systems used today. He also added that stacking chips vertically is merely a convenient "fix" that gets the job done, but is not a permanent solution to the problems inherent in the evolution of microchips and computing. Each new "fix" is more difficult to come up with than the one before it, as problems become increasingly complex and begin to defy the laws of physics. It's only a matter of time before "fixes" cease to work, and we have to turn to technology that's completely new and different. Meyerson said that if we haven't started investigating these new technologies by now, it may already be too late.
Recommended sites:
http://www.forbes.com/technology/2008/06/09/ibm-moores-law-tech-cionetwork-cx_es_0609ibm.html?feed=rss_technology
http://www.techworld.com/opsys/news/index.cfm?newsid=3477
Sandra wrote this article for the online marketers Star Business Internet internet service provider and website hosting one of the leading Internet service companies specialising in business website hosting in the UK
Article Source: http://EzineArticles.com/?expert=Sandy_Cosser
สมัครสมาชิก:
ส่งความคิดเห็น (Atom)
ไม่มีความคิดเห็น:
แสดงความคิดเห็น