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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: THE WATSONYOUTH who wrote (67116)	7/31/1999 12:44:00 AM
From: Elmer	Read Replies (1) \| Respond to of 1575626

Re: ", if you have specific questions about any particular point, I will try to elaborate if possible." OK: Re: ". I believe the original COPPERMINE delay was way overblown. I think INTEL only had problems with the L2 cache at the higher speed sorts at .18um (above 667MHz) The fix might well have been two new block level masks to raise the threshold voltages in the L2 cache ONLY. This would fit all the rumors I've seen on this thread.I believe the fix was minor (perhaps 60 millivolts ) with minimal or no loss in performance. This means INTEL should be able to produce all but perhaps the highest speed sort COPPERMINES in quantity NOW. " This is rather specific conjecture and it certainly doesn't seem like just an opinion to me. What is your basis for this statement? Re: "I expect a 25% improvement at .18um over the highest stable .25um part for the same design. I think the .25um part is limited to perhaps 650MHz but I doubt you will see a .25um speed sort there unless Intel accepts significant circuit limited yield loss " The term "yield loss" is usually associated with defects, not speed path limitations. There would be no "yield loss" because the parts could still be sold at a lower speed bin. There would simply be a lower "binsplit". Also what is a "speed sort"? the term "sort" is usually assorted with wafer sort where no speed determination can be reliably made. Why are you mixing incompatible terms? Re: "Where does that leave AMD? I see no response at .25um that could match what INTEL could do NOW at .18um. " If you have been following this thread, much has been discussed about the K7 adding a clock to their L1 cache accesses. If L1 cache access is in fact the limiting speed path, then simply looking at process capability may not give the full perspective here. Witness the high frequencies attained by the Alpha processors on what look like inferior process technology. So why do you say you see no response from AMD when architecture could be as significant as process technology? Re: "My guess is that INTEL will remain the MHz leader with COPPERMINE to at least the middle of next year. " I hope you are right but please provide a basis for your belief. Re: "I guess what I am saying is this. Line tailoring can add 40% in performance (nominal to minimum channel length in the same generation))" Please give us an example that will demonstrate this. EP