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 : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: Dan3 who wrote (29560)	9/15/1999 1:21:00 PM
From: Tenchusatsu	Read Replies (3) \| Respond to of 93625

Dan, <Video cards have local memory> ... which need to be filled using long burst transfers. In other words, bandwidth matters for modern AGP cards which run textures off local memory. (In contrast, AGP controllers that run textures off main memory, like Intel's i740 card and 810 chipset, depend on latency more than bandwidth.) <If your memory gets bogged down by the large burst transfers coming from AGP and PCI ... Isn't this more of an argument for Athlon's point to point bus over Intels bus?> No. Time and time again I've said that the memory is the bottleneck here. Athlon's 200 MHz bus is useless when running with PC100 SDRAM. The performance hit gets even worse during those times when AGP and PCI are hogging memory, locking out that well-publicized Athlon P2P interface. But if you think about it, Athlon's 200 MHz bus will take better advantage of RDRAM than Intel's 133 MHz bus. I think the only reason why AMD isn't admitting that right now is because they're afraid of tarnishing their "Anything But Intel" image. <VIA, SIS, and ALI are all supporting VC DRAM. Two chipsets have been sampling and will be in volume production in about the same time frame as DRDRAM chipsets. In one to two months.> Really? I didn't see this in EETimes. Nor do I hear Bert McComas, a very intelligent guy himself, sing the praises of VC DRAM. To me, the only source of VC publicity is you. <Sure do, that's basically the trick behind DRDRAM, interleaving on the chip itself, and it looks like that cost may kill it.> You didn't get my point. I'm saying that unless you plan on having gigabytes of memory on your system anyway, RDRAM will be cheaper than interleaving SDRAM. Tenchusatsu