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: Yougang Xiao who wrote (53748)	3/31/1999 1:19:00 PM
From: RDM	Read Replies (4) \| Respond to of 1572500

<The question I have for you is that while AMD attempts to achieve higher speed for K6-2 line, will this speed path problem happen again?> It is not unlikely that AMD will have further "speed path problem". Everyone has speed paths that are weak. They become a problem when the parameters affecting them fluctuate. AMD is the underdog. They are playing catchup and agressively go to the next generation. These generations are made available by ongoing linewidth and layer thickness reduction developments. Intel has the luxury of being currently ahead in performance. They have more time to produce an initial inventory of parts before announcing. While there is some sign that Intel is pushing closer to the line than in the past. One such example is the announcement of the 550 Mhz Pentium III for May introduction in February. <Additionally, will the speed path problem occur to K6-III line? > The K6-III is just as likely and perhaps more so since the die is large until it can be produced in .18 micron (perhaps in October). <K7 ?> The K7 has a fundamental design change that makes it different from the K6 and Pentium processors. It has a change in archictecture that makes it much easier to achieve a much higher clock rate(800 mhz vs 600 mhz). The tighest and worst timing choke points have been loosened which may make some operations slightly slower at a given clock rate. The belief is that the higher clock rates and the increased number of execution units that may operate simulateously in the K7 more than make up for the disadvantages. The K7 (IMHO) may leapfrog Intel in performance by all or most performance measures by this fall. It will be introduced with a initial clock rate of 500-600 Mhz, however this same .25 micron K7 part may be able to run at 800 Mhz before the end of the year. With .18 micron 1 Ghz and beyond may be possible as soon as next year if support chips and memory and cooling allow it. The question is will there be setbacks when a part is in limited availability. I believe that the answer may yes. This is a very natural result of high competition for the highest performance parts. With the K7 crossing the Intel performance barrier it may even happen that Intel will have to push the envelop more resulting in Intel having "speed path problems". I believe that competition, and process uncertainities play as big of a role in the "problems" as management judgment errors.