KZ, <In any case, these apps will run faster on DDR-Sdram which will achieve 2100 MBytes/second and be available by the time these apps ramp to any wide spread usage.>
First of all, how do we know that the apps we're talking about will run faster on DDR? Sure, DDR can brag of a 2.1 GB/sec peak bandwidth, but will actual throughput come close to this, even in the best-case scenarios? We've seen a lot of "benchmarks" regarding RDRAM performance, even though some may be way off the mark, but we've seen absolutely no DDR "benchmarks," good or bad. Second, we know that DDR will be available by the time these apps become widespread, but so will RDRAM, which already has a headstart anyway with Camino's launch next week. And third, chipset competitors like Via have traditionally lagged behind Intel in the performance department. Heck, they make PC133 look no faster than PC100 on an Intel chipset! There's a good chance that they'll also make DDR look sorry compared to RDRAM.
The problem with anti-Rambus people is that they never apply the same harsh standards of judgement to alternatives like DDR as they do to Rambus. I've noticed this in Bert McComas' presentations, where he focuses strongly on Rambus' negatives, but only sticks to cheerleading when it comes to PC133 and DDR SDRAM. Many other Rambus opponents follow this sort of pattern. There is a wonderful noun that describes this sort of behavior, and it's called hypocrisy.
Tenchusatsu |
