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: Bill Jackson who wrote (66441)	12/27/2001 9:53:37 PM
From: Elmer	Read Replies (2) \| Respond to of 275872

Elmer, a 50-60% advantage, ?? Hard to assess this since we have no way of truly comparing the two statements. On the face of it twice the dice make each one cost half the price. let us say that AMD dice cost $1 and Intel dice cost $2. Then they both add a $5 organic package, Now AMD costs $6 and Intel costs $7. That means that other costs dominate. First off AMD said they had a 10-20% die cost advantage but regardless, that's not quite how it works. A die twice the size of another die will yield less than half. It's not a linear thing. Think of it this way (BTW this is just a first approximation): You have a process with a defect density of about 50 defects per wafer. You have a wafer with 100 die and about 50 defects. You will lose about 50 die, actually less than 50 because some die will catch more than 1 defect. You have a yield of ~50% You compare that to a wafer with 200 die and the same 50 defects. You will get about 150 good die or a yield of about 75%. Notice how you can't use the percentage yield as a valid comparison because of the different die sizes, even though the process has the same number of defects for each device, and you now get 150 good die per wafer or 3X the number of good die when there's a 2/1 size difference. That was only a first approximation and other effects come into play. The ratio of complete/incomplete edge die changes and other factors as well such as cache size verses total die area etc. EP