10 replies to this topic

[forever freedom]

Simple question.


As for me, i can't help but come to the conclusion that it's a self-fulfilling prophecy; i can't think of any other explanation. But if that's so, it means that if we tried hard enough, we could make even faster advancements, and that would have huge implications.

[RighteousReason]

Market forces naturally drive technological innovation. I think it's just a coincidence that the number of transistors that can be fit on an integrated circuit happens to double every two years.

[Cyberbrain]

Well technically Moore's Law originally applied to processors where the number of transistors would double on an integrated circuit every two years. And eventually Moore's Law will end since there's only a finite amount of space on which to build a transistor. But according to Ray Kurzweil new technologies will begin to take place such as light or quantum computing

These are very exciting times to be alive and young!

[RighteousReason]

This implications are huge even without faster-than-moore's-law advancements.

[Mind]

If you read or listen to some of Kurzweil's stuff ( this video probably does a good job) you will find that he sees Moore's law as part of a bigger trend that is driven by the law of accelerating returns. So he claims...as do many others that technological advancement/progress will continue even after Moore's law reaches its limit (just in a different form).

[forever freedom]

My idea is that we should advance computer power faster than we already do. It appears that instead of being an incentive to advance, Moore's Law is holding us back, because we believe that by advancing at the rate according to moore, we already did our job and we can rest until the next deadline comes in.


Maybe the market dictates Moore's law, as Savage suggested? So, companies like Intel have no interest in spending more in R&D to give us a chip 4x better than the current chip, if they can charge the same price for a chip only 2x better?

[Cyberbrain]

Moore's Law was an observation made over 40 years ago that to some extant is still seen today, it is not an actual law by which companies oblige to. In reality the trend is slightly different, it demands mostly on economics and marketing. Intel will continue to provide the best possible chips regardless of Moore's Law.

[RighteousReason]

Moore's Law was an observation made over 40 years ago that to some extant is still seen today, it is not an actual law by which companies oblige to. In reality the trend is slightly different, it demands mostly on economics and marketing. Intel will continue to provide the best possible chips regardless of Moore's Law.

Agreed, if Intel could make a chip 4x better, they would (that's more market share from AMD!). I don't think Moore's Law is holding anybody back.

[Lazarus Long]

I was going to put this article in the physics section but it is also relevant to this discussion.

http://www.eurekaler...u-mpf102108.php

Public release date: 21-Oct-2008
Contact: Mark Shainblum
mark.shainblum@mcgill.ca
514-398-2189
McGill University

McGill physicists find a new state of matter in a 'transistor'
Could previously unknown type of electron crystal help the future of electronics?

This press release is available in French.

McGill University researchers have discovered a new state of matter, a quasi-three- dimensional electron crystal, in a material very much like those used in the fabrication of modern transistors. This discovery could have momentous implications for the development of new electronic devices. Currently, the number of transistors that can be inexpensively crammed onto a single computer chip increases exponentially, doubling approximately every two years, a trend known as Moore's Law. But there are limits, experts say. As chips get smaller and smaller, scientists expect that the bizarre laws and behaviours of quantum physics will take over, making ever-smaller chips impossible.

This discovery, and other similar efforts, could help the electronics industry once traditional manufacturing techniques approach these quantum limits over the next decade or so, the researchers said. Working with one of the purest semiconductor materials ever made, they discovered the quasi-three-dimensional electron crystal in a device cooled at ultra-low temperatures roughly 100 times colder than intergalactic space. The material was then exposed to the most powerful continuous magnetic fields generated on Earth. Their results were published in the October issue of the journal Nature Physics.

Two-dimensional electron crystals were discovered in the laboratory in the 1990s, and were predicted as far back as 1934 by renowned Hungarian physicist Eugene Wigner.
(excerpt)

[Ghostrider]

Moore's Law was an observation made over 40 years ago that to some extant is still seen today, it is not an actual law by which companies oblige to. In reality the trend is slightly different, it demands mostly on economics and marketing. Intel will continue to provide the best possible chips regardless of Moore's Law.

It's mostly driven by competition. I think if AMD were gone, then Intel would still invest money in research, however, the goal might not explicitly be to make the chips faster. Moore's Law is really just a side effect of the process technology. Currently, the transistor size gets cut in half roughly every two years. Back in the 90s, it used to be every 3, but because of competition or other factors which I am unaware, they now shrink it about every 2. Which is pretty amazing considering that going from 90 to 45 nm is a lot harder than going from 2 um to 1 um...well, maybe not because more technology is available when shrinking from 90 to 45 nm. Anyway, the smaller transistors have 2 effects with regards to performance. One is that the physics of the transistor improve and the other is that since the transistors are smaller, more can be crammed in the same area as before. The architecture designers also perform some optimization and tricks to get more performance out of transistors, but only so much can be done here and it's really the process technology which most influences the power and cost of the processor (and the profit of its manufacturer). The reason I think Intel would still invest in R and D even without competition is because it still benefits them to make smaller transistors and improve yield. Ignoring competition, they would want to produce as many functional die per wafer as possible. So they would still try to shrink the transistor and improve yield, but this might happen at a slower rate if there were no competition.

[platypus]

If you read or listen to some of Kurzweil's stuff ( this video probably does a good job) you will find that he sees Moore's law as part of a bigger trend that is driven by the law of accelerating returns. So he claims...as do many others that technological advancement/progress will continue even after Moore's law reaches its limit (just in a different form).

The number of technological breakthroughs seems to be a lot smaller now than 100 years ago. Computing is doing great though.