Digital BW, The Print

> > Here's the results of a quick & dirty PS Guassian blur test: dual G4 -
25
> > seconds, dual Athlon MP1800 - 15 seconds, dual 2.2 Xeon - 8 seconds.
>
> What's the memory architecture/speed of the two x86 systems?  The issue is
> the cache of the Xeon isn't near large enough to hold the imaging file, so
> it resides in memory...and I'd bet you most anything, that memory speed
> "helps" more than the fact that it's a Xeon vs a non-Xeon.

DDR 333 vs 800mhz Rambus. Rambus is faster ram, no doubt, but in practice
the difference in performance isn't nearly as great as the difference in
memory bandwidth between the two systems would lead you to expect.

Would you be willing to bet your Hassie gear? The Xeon has almost twice the
floating point performance of the Athlon system (used in this comparison),
and the hyperthreading under XP shows four virtual CPU's.  The Gaussian blur
filter uses both CPU's, and with the Xeon/XP setup, Windows task manager
shows all four being used. So yes, it is a Xeon vs non-Xeon thing.

Anyway, this is starting to veer OT, so I'd like to bow out of this
discussion. I'll send your the address where you can ship the Hassie gear
off-list.