Gig4moller:

MY CHIP DOES 4 GHZ ON THE P5W !!

It's the absurd FSB limitations that are screwing up everything!
400 * 10 did NOT post. I don't even want to DREAM About how bad the chip would clock on its default 9 multiplier...probably no higher than 9*370 or something...

370 * 10 was perfectly stable at 1.3v.
380 * 10 was unstable even at 1.38v ! Kept getting either random errors in one of the prime threads or roundoff errors or something after awhile.
And 1.3v to 1.325v simply *locked up* the system hard after a few iterations.

Now I was sitting here thinking I may have a worse chip than evga2000 that won't do 4 ghz at ANY voltage...then I suddenly remembered why I paid $1300 for a chip...

I cranked up the multiplier to 12 and the FSB DOWN to 334.
Yielding 4.01 ghz.
System POSTED (!!), went into windows, ran Prime and laughed at my incompetence. It was the freaking FSB that was holding everything back -.- No idea if more VMCH on the northbridge would have helped but who cares...I have an XE chip....

The temps started getting too hot though, at 1.3875v on the Tuniq heatsink. One of the cores started pushing 72C, then I figured ok that's enough. I have a 4 ghz capable chip, that can probably do a helluva lot more on a board without stupid FSB problems. I'm currently attempting prime at 1.35v at the same speed (4013 mhz). I'm not after 24 hour prime stability anyway, just game stability. If one of the cores doesn't error out in 30 minutes with prime, then I'm good to go. (I will have to find 3dmark 2006 somewhere and run that just to be sure).

So, in summary, YES it works FINE on the P5W! Just be prepared to keep the FSB low and the MP high if you expect it to clock. Unlocked multipliers FTW.

I think it goes without saying that people with locked Multis will want to look at a P35 or X38/48 board, since you WILL be held back. Don't the non extreme quads allow limited multiplier adjustment, BTW?