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.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Scumbria who wrote (80414)	5/4/1999 2:19:00 PM
From: Tenchusatsu	Respond to of 186894

<The cache line size of CPUs being only 32 bytes, what benefit is there for connecting them to high bandwidth dram? Bursts in and out of the CPU are of extremely short duration for the vast majority of applications. Latency is far more important than bandwidth on the desktop.> Cacheline sizes can change (though probably not for Camino). Also, one of the benefits of RDRAM is that it's better suited toward handling concurrent traffic from multiple sources, such as processor, AGP, and PCI. Even Bert McComas, an "anti-Rambus" guy, agrees with this. As for latency vs. bandwidth, one solution to the latency problem is to integrate the memory controller onto the CPU die. Compaq/Digital will do exactly this when they put four RDRAM channels onto the die of the upcoming Alpha 21364. You may be seeing other companies integrating memory controllers as well. And I guarantee you that you won't be seeing DDR SDRAM controllers being integrated onto the die. The low pin count of RDRAM makes it perfect for such integration. Got any other solutions to the latency problem? The whole computing industry is open to suggestions. Tenchusatsu