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: kash johal who wrote (39420)	10/15/1998 7:05:00 PM
From: Tenchusatsu	Respond to of 1588650

Kash-poor (tm PREngle), <AMD get's higher performance with p2p.> Yeah, and AMD may have higher costs and limited expandability with P2P. You haven't given me details on why the K7's P2P paradigm is better than Xeon and Merced's multiprocessor bus. After all, two P2P connections going at 200 MHz is overkill if most of that bandwidth sits idle due to memory or chipset latency. <They also get higher performance with the bigger L2 cache - up to 8Mb.> You didn't listen before, Kash. The K7 is probably going to start with 512K of L2 cache. That's no better than Pentium II. The 8MB figure is a theoretical limit. <They also get higher performance with L1 128K.> You didn't listen before, Kash. An L1 cache which is 128K is bound to be slower than a 32K or 64K L1 cache. That's the accepted rule-of-thumb. Maybe AMD found a way to work around this slower speed in the L1 cache, I don't know. The point is that as a cache becomes larger, the speed goes down (or the cost goes up). <Are you getting the picture???> Obviously the picture you're getting is a four-way, 1 GHz AMD K7 system, with 8MB of full-speed L2 cache, 400 MHz front-side bus, and dual RDRAM support. This system in your mind will cost less than $1000, and AMD will still be able to push their ASP to $250 and their margins to 75%. I, on the other hand, prefer to be a realist. Tenchusatsu