Pete, I think you are stretching hope with your last post:
<Within three months, the 760MP comes with 2GB PC2100, and dual Tbirds at somewhere over 1.2Ghz each (possibly Mustangs at 1.46GHz) at a price still less than the SMP Cmine board. It may even have a third party optimizing compiler and triple the performance.>
You've got to be kidding me. First of all, the motherboard itself isn't going to be trivial to manufacture. The north bridge itself will require over 1,000 pins (two EV6 ports plus two DDR channels plus whatever connection to the south bridge). Then you have the enormous power requirements of each Athlon. Finally, it's not even clear that the single Athlon 760 chipset will be here in three months, much less this dual Athlon platform.
Second, I don't think it's reasonable to hope for "triple the performance" out of an optimizing compiler. Even AMD is expecting 50% improvement in SPECfp performance (and less in SPECint), and only part of that is due to compiler enhancements.
And third, it's most probable that when dual Athlon platforms on 760 DDR chipset appear, so will dual Willamette platforms on Colusa dual-RDRAM chipset. No way will a dual-Athlon be able to compete performance-wise with dual-Willamette. The only hope is to compete in a price/performance sense once again, but that matters less when you're talking about high-end workstations.
The only advantage this 760 platform will have is cheaper memory thanks to DDR. That ought to offset the higher platform costs associated with dual-DDR channels.
Tenchusatsu |
