SI
SI
discoversearch

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)
AMD 231.83+1.7%Jan 16 9:30 AM EST
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: pirasa2 who wrote (238440)8/9/2007 10:08:42 PM
From: misenRead Replies (2) of 275872
 
Forget Windsor entirely for the moment. Assume 4000 300 mm wafer starts per week. Conservative number, so late into the 65 ramp. Times 13 = 52000 wafers per quarter. 70685sq mm wafer area, 126 sq mm die size for Brisbane. Reduce wafer area 10% for edge effects. Assume 90% yields.

Voila, you produced 23.4 million KGDs in a quarter.

Your math above works out to 450 good die per wafer (assuming 100% line yield -- 4000 WSPW = 4000 wafer outs/week).

To achieve 450 die per wafer on a 126mm2 die requires defect densities in the range of .05 to .10 defects/cm2. World class is considered to be in the range of .20 to .25 defects / cm2. AMD has recently indicated that Barcelona yields have defect density < 0.50 / cm2. So you are grossly overestimating die yields. Using a more reasonable estimate of 0.25/cm2 for defect density and a line yield of 95%, you get 340 die per wafer started.

If the defect`density is actually closer to 0.5/cm2, then good die per wafer start is ~ 250.

You may reply that you were convervative in your wafer start capacity -- perhaps. But Barcelona will consume a huge quantity of silicon for any reasonable output. Each Barcelona die will be equivalent in wafer requirements to at least 5 Brisbanes (assuming 50 Barcelonas versus 250 Brisbanes).

You're also assuming there is a market for 23M Brisbanes per quarter which I think is dubious.

I would agree with previous poster's viewpoints that 23M processors/quarter from AMD is far from realistic.

Misen
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2026 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News