Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: jspeed who wrote (217307)	11/21/2006 3:50:45 PM
From: dougSF30	Read Replies (3) \| Respond to of 275872

Of Intel PR-department strove however to play down it as the publication of hypothetical numbers.

To: jspeed who wrote (217307)	11/21/2006 5:27:37 PM
From: combjelly	Respond to of 275872

"which can be combined to 160 Quad cores" Hmm, it looks as if Intel's costs aren't nearly as low as people claim. It looks as if AMD pays somewhere around $5k per processed 300mm wafer from Chartered. I base that on Chartered's information on their 300mm wafer ASP's and the adders for extra metal layers. Which is roughly where Intel's costs are according to this. I am not sure what manufacturing overhead covers, and I think that 'Core Yield Loss' covers packaged chips that fail test. So if the manufacturing overhead is included in the $5k that Chartered charges, then Intel's costs are higher.

To: jspeed who wrote (217307)	11/30/2006 9:27:30 PM
From: THE WATSONYOUTH	Read Replies (2) \| Respond to of 275872

Well, that's all very interesting. Based on 143mm^2 die size and 16.5% edge/kerf loss for that size chip, I would estimate 426 total possible die per 300mm wafer. 75% yield gives 319 good die. How about that.... same as Otellini's hypothetical case. Now for the much smaller 81mm^2 Cedar Mill, I estimate 765 possible die using a 15% edge/kerf loss for the smaller die. Of course the smaller die will yield higher given the same defect density. In fact up to about 84%. That's 642 good die or 321 dual cores. Gee....almost exactly the same number of dual cores from one Cedar Mill wafer as CWM. So here is a real tough question. Which same yielding chip would Intel choose to make? CWM or dual core Cedar Mill? Come on guys, the CWM yields can not be any where near 75%. In fact, I think 75% just happens to be where the yields are in fact the same. Now, we need to find a combination of yields that are significantly lower and cause the dual core Cedar Mill route to have maybe 23% better yield than the CWM route as implied in the article. Lets see . I estimated 44% for CWM. That's 187 good CWM per wafer. I estimated 60% for Cedar Mill. That's 459 single core or 229 dual core. 229/187 is 1.225 I think maybe this was the point at which Otellini ripped up the charts. Now why do you suppose that was. The world needs to know. THE WATSONYOUTH