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) -- Ignore unavailable to you. Want to Upgrade?

To: Elmer Phud who wrote (123998)6/17/2004 11:39:37 AM
From: Joe NYCRead Replies (1) | Respond to of 275872
 
Elmer,

AMD's gotten everything out of this process they're going to get. They're at the end of the line and they need 90nm.

Do you think so? I think there is a good chance of 2.6 GHz 130nm parts, maye late Q3, at least on FX. The roadmap has the first 90nm chip on the FX line in 2005 and they need another speed grade, so 130nm has to do it.

The current shipping FX53s can apparently do 2.6 GHz on air cooling even without voltage hikes, AMD can accumulate enough to do at least a low volume FX launch at 2.6 GHz.

Joe



To: Elmer Phud who wrote (123998)6/17/2004 12:55:22 PM
From: TGPTNDRRespond to of 275872
 
Elmer, Re: <AMD's gotten everything out of this process they're going to get. They're at the end of the line and they need 90nm.>

What do you expect needs to change other than the process name for AMD to be at 90 Nm?

Binsplits go up but the yields go down due to the inability to reliably resolve features at those dimensions.

To me resolution would seem to be equipment related. Do you think we're looking at major equipment replacement to get to "90 Nm" or do you think it'll be a shrink of all but channel lengths? That would be an interesting shrink, wouldn't it?

Both companies had a long history of first a photographic shrink, then an optimization stepping on new process. Perhaps that needs to be reversed.

Maybe by September the situation will be more clear.

The same can be said of the market, I guess.

Looks like Jabil doesn't like prospects for next Q. Anybody know what socket JBL mostly builds for in MOBOS?

-tgp



To: Elmer Phud who wrote (123998)6/17/2004 1:43:33 PM
From: fyodor_Read Replies (2) | Respond to of 275872
 
Elmer: I would think this would not look like a normal distribution on the high end of the curve.

Something like what you suggest is what I would assume that any "high-end process" bin split distribution would look like. I.e. a convolution of a Gaussian and some other, possibly discontinuous, function.


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