Moore's Law

[hw help]

Can someone give me an example of Moore's Law used in real life?

 

Thanks

[EverCurious]

man, thats a dangerous question to ask around here, i asked for some help about a theoretical physics paper, and some guy said i would never get ahead in life and etc...go read the forum its pretty sad its the one started by me called "Any Ideas?" and its in the theoretical physics section

[swansont]

Can someone give me an example of Moore's Law used in real life?

 

Thanks

 

Moore's law states that the number or transistors on a computer chip (or processing power or speed) will double every 18 months or so. It has been observed to hold. So, what do you mean by "used in real life"? It applies to computers and only computers.

[the tree]

It has been observed to hold.

It's soon to stop holding, as silicon can only get so small.

[ecoli]

It's kind of a ridicuous thing to label as a law, but it does seem to be holding true.

[Aardvark]

It's soon to stop holding, as silicon can only get so small.

 

Any chance of other materials being developed?

[swansont]

It's kind of a ridicuous thing to label as a law, but it does seem to be holding true.

 

Actually it's exactly what you'd expect to be labeled as a law - a simple mathematical relationship that's been observed to hold true under some set of circumstances. That's what a law is.

[m4rc]

I see Moore's law more as a self fufilling prophesy. After Moore noticed the trend and published it, the computer chip manufacturers had something to compare to their progress. If they could not keep up with Moore's law then they were not at the cutting edge. So they would all spend enough on reseach until they either reached the level predicted by Moore's law or went bankrupt. So the existance of Moore's law (along with the fear of not keeping up with it) is the reason that it is true.

[matt grime]

That makes no sense. Which computer company would not want to be the first to break Moore's law?

[swansont]

And falling behind does not mean that throwing more money at the problem will necessarily speed you up. There's a research adage: It takes a woman nine months to produce a baby, but that does not mean that nine women can produce a baby in a month.

[matt grime]

I should have course have said 'which company would not want to break moore's law by making improvements significantly faster than it predicts'. No company would want to break moore's law by failing to double at least at the proposed rate.

[ecoli]

Actually it's exactly what you'd expect to be labeled as a law - a simple mathematical relationship that's been observed to hold true under some set of circumstances. That's what a law is.

 

But, by definition, a law cannot be broken. If people decided to stop making computer chips, or everybody suddenly died, Moore's law would be broken.

[swansont]

But, by definition, a law cannot be broken.

 

No, not generally. Only true if you recognize that laws don't always apply universally. This is a common misconception often held by the same people that think that a law is superior to a theory. A law often times has limited application and is not universal, failing to hold under some set(s) of circumstances. One gets into trouble in trying to apply laws under circumstances where they do not apply.

 

A law can never be broken under circumstances where it was valid, and laws are only valid under circumstances where they aren't broken. But tautologies like that aren't especially helpful.

[ecoli]

A law can never be broken under circumstances where it was valid, and laws are only valid under circumstances where they aren't broken. But tautologies like that aren't especially helpful.

 

if that's the case, then anything can be descirbed as a law... you can just limit the circumstances to any parameters you want.

[bascule]

Moore's Law is one particular observation of a trend of exponentially increasing progress which comes about through the additive feedback of improving technologies in all sorts of fields and disciplines. While in Moore's hayday an IC's die was painstakingly laid out by hand, today not only is the layout done on computers, but computers are doing the layout, with the CPU itself written in a description language much like a computer program, and computer algorithms calculating the optimal layout of the components from the description. This software is constantly improving. CPU design concepts are constantly improving. The computers used to design new CPUs are constantly improving. The materials used to construct new CPUs are constantly improving. All of these things feed back off of each other and actually increase the rate of exponential change.

 

Kurzweil covers hundreds of these trends and generalizes them into something he named The Law of Accelerating Returns

 

See the chart on Wikipedia

[matt grime]

if that's the case, then anything can be descirbed as a law...

of course not, i cannot call my dining table a law because of that description

 

you can just limit the circumstances to any parameters you want.

 

yep, that's how it works. Moore's law would certainly fail if we suddenly ran out of silicon, and other materials, but it is a tacit assumption that we won't.

 

the point is that there is a ***non-vacuous*** set of circumstances under which the law appears to hold very well.

 

 

However, I would personally not say that Moore's law is actually a law in the scientific sense. It is a law like murphy's law or parkinson's law is a law. An aphorism that holds true. It is surprising that it still remains true despite many people thinking that it would start to fail as we hit smaller and smaller dimensions. I don't think anyone is saying that there is some causal reason for it to be true like newton's laws of gravitational attraction, or kepler's laws of planetary motion.