Tell me more ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.
up vote 6 down vote favorite
1

I see different sources that say the number of transistors will double every 24 months or every 18 months.

Does anyone have a credible source?

share|improve this question
1  
What's your problem? Both sources state that the number of transistors will double approximately every two years. As Moore is a founder of Intel, why not trust Intel the most? Also, if you read the sources of wikipedia, you'll find that Moore himself adjusted his prediction. – Falcon Sep 2 '11 at 16:58
The CEO of a company I once worked for said that More's law was that "Your profits should double every 18 months" and he was DEAD serious. – aceinthehole Sep 2 '11 at 19:34

4 Answers

up vote 10 down vote accepted

The original source ftp://download.intel.com/museum/Moores_Law/Articles-Press_Releases/Gordon_Moore_1965_Article.pdf .

Note that Moore's law said nothing about performance, it was simply that the optimal number of components on a single IC would increase exponentially with time (from a cost/function point of view).

enter image description here

Quote from the 1965 paper: "This allows at least 500 components per linear inch or a quarter million per square inch"

The current core i7 is approximately 0.4 sq inches and has 750M transistors so 2billion per square inch!

share|improve this answer
+1 for a link to the article. Note that "optimal" was based on the manufacturing cost per component. There are fixed costs for running a wafer through a FAB, so the more components you can fit per wafer (unit area) the less expensive your per-component cost is. If you pack them too tightly, however, you increase the risk of creating defective units, which start driving your per-unit cost back up again. – Jay Elston Sep 2 '11 at 17:24
@Jay and then there are all sorts of extra power laws for; component size vs wafer size, edge losses, kerf losses, stepper time - it's an interesting topic! – Martin Beckett Sep 2 '11 at 17:25
up vote 7 down vote

The wikipedia article explains it;

Moore slightly altered the formulation of the law over time, in retrospect bolstering the perceived accuracy of his law.[17] Most notably, in 1975, Moore altered his projection to a doubling every two years.[18] Despite popular misconception, he is adamant that he did not predict a doubling "every 18 months". However, David House, an Intel colleague, had factored in the increasing performance of transistors to conclude that integrated circuits would double in performance every 18 months.[note 1]

share|improve this answer
up vote 5 down vote

As usual, Ars Technica has an excellent article on the meaning and history of Moore's Law. It cite's Gordon Moore's 1965 paper containing the original prediction(s). Whether the period is 12, 18, or 24 months really doesn't matter all that much in most cases -- the most important element is that component density has continued to grow geometrically for much longer than anyone thought it would. I'd recommend starting with the Ars article and then reading Moore's paper if you're still interested in knowing more.

share|improve this answer
2  
actually Moore's Law is not so much of a law as it is a target; each year, the CPU industry uses Moore's Law to calculate the density/cost that they will be aiming for their next fab plant that they will be building. The law is used to determine hiring strategy, R&D expenditure, etc so that the "Law" is fulfilled year by year. – Lie Ryan Sep 2 '11 at 20:45
up vote 0 down vote

Check here: sound right to me:

http://www.computerhistory.org/semiconductor/timeline/1965-Moore.html

It started at 12 months, then changed to 24 months...

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.