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: Epinephrine who wrote (105022)	4/13/2000 10:49:00 AM
From: Joe NYC	Respond to of 1574001

An interesting post on Ace BBS: aceshardware.com Typically, new processors are poorly laid out at introduction, and therefore not particularly impressive compared to their best competition of the day. Look at Athlon, launching at only 600 MHz versus 500 MHz for Katmai. However, as new versions are introduced, the layout is improved and relative speed grades increase. It is estimated that T-Bird will be able to reach 1267 MHz at 180nm design rules. And the shrink to 150nm design rules could take T-Bird to 1.5 GHz, while P-III might, at best, get 1067 MHz (shippable) out of 150nm design rules. Even more impressive, it is estimated that Mustang might be able to get 1.8 GHz from 150nm design rules. Of course, with the Athlon family this s using copper interconnects. Adjusting for aluminum interconnects, Mustang might still hit 1.6 GHz at 150nm. This translates to an increase from 20% higher speed grade at introduction to a 50% higher speed grade by the second derivative variant (not counting Spitfire). Actually, this is even more impressive when it is realized that Athlon classic was launched using six metal-layers while P-III was using just five metal-layers at the launch of the Athlon. This means that K7 would have been only about 11% faster in equivalent process technology at launch. One thing that most observers tend to ignore is that Mustang will gain its speed grade advantage over T-Bird primarily through a transition from a 95% standard cell design to one using much more fully custom logic. This move may actually lead to Mustang having a smaller core size than T-Bird. It is conceivable that a Mustang variant with 512K of L2 cache would be about the same size as a T-Bird with 256K of L2 cache. This could lead to AMD replacing their entire Athlon product line with Mustang derivatives, right down to a Mustang based Spitfire replacing T-Bird based Spitfire in the same way that Cu-Mine based Celeron II is replacing the P-II derivative Celeron. Imagine the impact if Athlon Pro (T-Bird) is replaced by a Mustang based Athlon Pro-II with 512K of 16-way set associative L2 cache and Athlon Select (Spitfire) is replaced by a Mustang based Athlon Select-II with 128K of exclusive architecture 16-way set associative L2 cache. Especially if Athlon Select-II has a smaller die size than Athlon Select classic, this could spell real trouble for Cu-Mine. With Athlon Pro-II shipping at a peak Speed grade of 1.5 GHz (or would that be 1.533 GHz) AMD could safely match Cu-Mines peak speed grade with Athlon Select-II's peak speed grade. Since such an Athlon Select-II processor would stomp all over Cu-Mine at the same speed grade, especially with Athlon Select-II being serviced by PC2100 DDR SDRAM, even if AMD were to charge $199 for the 1.0 GHz variant, Intel's peak speed grade Cu-Mine price would be severely depressed. But it is possible Intel will be able to engage in payback with Willamette in about two years.