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: Brian Hutcheson who wrote (40918)	11/5/1998 9:04:00 PM
From: Tenchusatsu	Read Replies (2) \| Respond to of 1573421

<Paul , since independent tests rated the original (without L2) Celeron 266mhz as equivalent to the K6 200mhz in performance and adding 128K L2 cache resulted in Celeron A's increased performance then adding twice the amount of L2 cache (256K) to K6-2 should not only restore the performance gap mentioned above but increase that gap> Face it, AMD can't choose an easy-to-implement 128K on-die cache like Intel could. They have to go for 256K for two reasons: 1) The K6-2's L1 cache is 64K, compared to the P6's 32K L1 cache. (Amazing how the Celeron matches the performance of K6-2 even with a smaller L1 cache.) It's useless to make your L2 cache only twice the size of your L1; i.e. you have to go at least quadruple. 2) The K6 and K6-2 already benefit from an L2 cache on the motherboard. When you go from one to two levels of cache (Celeron), the performance increase is significant. But when you go from two to three levels of cache (K6-2 to K6-3), you'll suffer from dimished returns. So you see, AMD has no choice but to go to a larger 256K of on-die cache. Of course, the difficulty of implementing this could be a factor in the K6-3's schedule slip. Tenchusatsu