Moore's law

From Lesswrongwiki
Revision as of 04:54, 4 June 2012 by Daniel Trenor (talk | contribs) (The End of Moore's Law?)
Jump to: navigation, search
Smallwikipedialogo.png
Wikipedia has an article about

Moore's Law is a term attributed to Intel founder Gordon E. Moore who observed in 1965 that the number of transistors that could be purchased inexpensively and placed on an integrated-circuit doubles every year. He later revised this figure to every 2 years (1975). The doubling period is often mistakenly reported as "18 months" or shorter, this is due to being confused with the overall processing power of computers doubling more quickly because of factors such as increases in clock speed, increases in cache memory or improvements in chip design.

Importantly, the ability to increase the number of transistors available is due to reducing transistor size rather than increasing the size of the integrated circuit. As of 2012 the smallest commercial available transistors in a microprocessor are 22 nanometers, their 1965 counter-parts were 100 micrometers (100,000 nanometers). Intel have revealed they expect to release a 5nm in the near future which would continue the projections of Moore's law.

The End of Moore's Law?

For almost as long as there has been Moore's Law there has been commentators predicting its demise due to the limitations of engineering being reached. However, it does appear that component size does have a limiting factor in the form of the Laws of Physics. On the smallest of scales electrons become "smeared" across time and space meaning that they can no longer remain contained within the boundaries of very small components which results in disastrous contamination of a processing signal.

Carbon nanotubes and other engineering solutions would appear to address this problem, but even these cannot be shrunk any further due to the fixed physical limits of their structure.

As shrinking components any further would be impossible a new form of chip will be required to extend Moore's Law. These chips will be "stacked" 3-dimensional chips rather than the 2-dimensional designs of today. There are a number of engineering issues that need to be addressed in order for 3D chips to be viable, the most significant of which is the issue of heat dispersal.

Blog Post

External Links