You're making multiple assumptions here. First going 45nm to 32nm didn't "nearly double" the amount of chips per wafer. Second you're assuming that when moving from the 45nm node to the 32nm node that the processor transistor count stayed the same. And most importantly you're also assuming that the yield per wafer stayed the same going from 45nm to 32nm. Even with all of that it still ignores the fact that as the advanced nodes hit 28nm and below the complexities and problems skyrocketed (and therefore costs increased). If the # of CPU's yielded per wafer had stayed the same going from SB to IB you would be seeing a huge increase in price. The only saving grace was that because it shrunk to 22nm the # of CPU's yielded is "supposed" to increase (although with the fab problems they're having I would love to see if that's really the case).
As for the profit margins. What I am saying is that Intel is in a much better position to absorb a decrease in profit margins by shifting to 22nm than companies like AMD. The single worst decision that came out of AMD in the last decade *cough* Hector Ruiz *cough* was spinning off the fabrication facilities into Global Foundries. Now AMD has to rely on a 3rd party to manufacture their dies who are factoring in a profit before AMD even gets their hands on the dies. No surprise that with all of that in mind they're at 45nm moving to 32nm with Piledriver.
First I will addres your dual points that cover 4 areas. Counting is not your strong suit.
1) math shows yes I am right
2) yield values which you bring up are another argument, not as large an issue elsewhere not an issue here.
3) I never assumed the transistor count stayed the same, who told you that I will punch them in the eyehole.
4) "The only saving grace was that because it shrunk to 22nm the # of CPU's yielded is "supposed" to increase" Do you have some information otherwise except some stupid "news" post and a familiarity with nvidias incapability to produce a solid 45nm product on TSMC foundry?
A1) Math shows that 32nm produces nearly double the amount of 45nm chips.
A2) never said transistor count stayed the same. WTF you talking about.
A3) By eyehole I mean something worse and by punch I mean not using any part of my body eww. Defintiley don't want my fist in anyones buttocks.
A4) Per Anything, sorry that is just a ridiculous assumption. Come one. I asked their PR firm about some SSD facts, things knownw on the internet and got called and insider. whatever the **** that means. Dunno, I dont Intel.
A5) to the question you didnt ask. Yes it is cheaper to produce a smaller size than continue to produce the ame thing repeatedly.
Your not thinking correctly when they shrink transistors they add more transistors that's the hole point from the beginning, power saving speed was a extra in time. Have you seen the size of the SB die or BD.
They have increased the size of the wafers to try and lower cost however they have problems there also.
You are thinking wrong. Adding transistors has nothing to do with physically adding anything. It is a design etched in acid. Believe it or not, you have no choice on whether it is a cpu or not. Just like you have no choice but to be wrong, because you are. IT isa lithography process and designed before you entered the equation.
EDIT: Got caught up in the "fury" of posting. sorry. the posts are ridiculous no denying that. My responses are uncalled for I apologize. Not deleting them cuz they are funny. If you dont find them funny, report them and me. I will snicker the whole time