KZNerd, <Tench, I asked for a clear statement why the evolutionary approach cannot keep up in the same way it has for the last 30 years.>
First of all, what's your definition of "evolutionary," and why is the transition to RDRAM so "revolutionary" compared to previous advancements? To me, things change so quickly in the computing world, so it's hard for me to define anything as a "revolution."
<The issue actually is: "Is it better in any substantial way that actually makes it worth all the disruption that it is forcing on the dram makers as well as the on-going higher cost of the drdram?">
Well, why do you think Intel is hedging its bets on PC133 SDRAM? Perhaps the transition to RDRAM is a lot more than the industry can handle at the moment, but the transition will take place.
From the technical side of things, all I can say at the moment is that Rambus is better. Forgive me for not saying more, but I feel like I can't due to the limited scope of this discussion forum. If you can understand stuff like bidirectional signaling, bank interleaving and precharging, the impact of pin count, turn-around cycles for SDRAM, the need for source-synchronous transmission, and performance profiles of chipsets, then I can give you some clearer statements. Otherwise, you'll just have to trust that my judgement is sound.
Tenchusatsu |
