Need help describing Moore's law and semiconductors

[smp]

I'm trying to describe how microprocessors and semiconductors work, and how the smaller the stepping (?) and the more transistors you can fit into a given slice of silicon equals faster processors. Please help! I'm writing a paper and I'm having some serious mind/writers block. Thank you!

 

[jcmkk]

You might also want to discuss the purity of the silicon.

 

[pm]

Here are a few papers to kickstart your muse:

Here's the original paper written by Intel's Gordan Moore in 1965 that formed the basis of "Moore's Law" (acrobat required).

Here's a paper from the IEEE Spectrum (an electrical engineering magazine) in 1989 regarding the history of CPU's and where the industry is headed.

And here's the keynote speech that Gordan Moore gave at the 1997 Intel Developer Forum on "An Update to Moore's Law".

I found these mostly by typing "Moore's Law" (quotes included) into Google. I found the first and the last articles to be particulary interesting.

 

[Chipset]

<< You might also want to discuss the purity of the silicon. >>



Ask Britney Spears.

 

[jcmkk]

good one chipset

 

[smp]

pm to the rescue!!!

Thanks pm. Hey you don't happen to have any old papers regarding this stuff laying around that you want to donate to me? I'm writing a paper about the future of microprocessors and Moore's law. My essay is supposed to deal with a scientific paradigm shift, or a scientific fix. Has moore's law been disproved yet? Did he say that a processor could only reach 3ghz?


hah @ Chipset!!!

 

[pm]

<< Thanks pm. Hey you don't happen to have any old papers regarding this stuff laying around that you want to donate to me? >>

A copy in acrobat format isn't enough?


<< Has moore's law been disproved yet? Did he say that a processor could only reach 3ghz? >>

Moore never made any claims about performance or frequency at all. His statement was purely about cost per transistor and transistors per chip. His 'law' (and I put it quotes primarily for his benefit since I hear he has never liked the words "Moore's Law" and doesn't consider the results of his paper to be a "Law" at all) is on page 2 of the first link above under the section "Costs and Curves":

"The complexity for minimum component costs has increased at a rate of roughly a factor of 2 per year."

That's Moore's Law. The number of transistors in a chip doubles per year. And it's been disproven several times. Briefly in the 70's, it underestimated the trend (it was closer to 2.5X per year), and over the long haul it's been optimistic since the current trend from the 60's forward is for a doubling every 18 months.