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: Elmer who wrote (83686)	12/20/1999 4:15:00 PM
From: kapkan4u	Read Replies (1) \| Respond to of 1572208

<Re: "Looks like spitfires will have 64KB of L2 cache on chip." Doesn't that seem very very odd? > Ever heard of victim caches?

To: Elmer who wrote (83686)	12/20/1999 4:45:00 PM
From: Tenchusatsu	Read Replies (2) \| Respond to of 1572208

Elmer, 64K of L2 cache on a Spitfire is indeed odd, especially if the L1 cache remains 128K. But if the AMD design guys do it right, they can make it so that 64K effectively becomes an extension of the L1 cache. (For example, Kap suggested that the L2 will be a "victim cache.") In effect, you can have a processor with 192K of cache. Of course, you won't have the same access latencies across the entire 192K of cache space, but as long as they can reduce the cache miss rate, there will still be a benefit. The only problem now is validation. It's not easy to rearchitect an L2 cache in the limited time frame that AMD has, especially if they decided to implement a victim cache. Cache coherency is never a fun thing to validate. But then again, who said competing against Intel was easy? Anyway, Spitfire's 64K L2 cache sounds very interesting. I wonder how it will compare to Intel's upcoming Coppermine-128. Tenchusatsu