Well.... Got it done with a dremmel tool and the Megahalem is now attached. Gotta say it's not as good as I thought it would be.

Is it the temps or the failed factory specs?

The temps.... I was able to lower the volts a bit and get them down some but it's not much better than the stock cooler I had on there. I could only get one fan on it because it would take out the first two ram slots if I tried to use two fans LOL. I don't know why these people don't think of that and make these things a little taller, I mean I know they gotta fit into a case for most applications.... but make a taller one for us guys that aren't using a case! :D


I'm sitting at 3.8ghz now... got it to boot at 3.9 but it wasn't stable at all so i'm still playing around with settings I know nothing about. Looks kinda grim for anything higher though.

Don't lose faith brother, she's just playing hard to get right now, take ur time, get to know her a little bit more, slip something in her drink then bang her at 4.2ghz :rofl

I'm trying but until I know what all the settings do and understand them it's futile.

Only thing I can suggest is scouring google, reviews, forums for clues to the AMD bios. Beyond that its old fashioned science test and log.

Yeah....just not much out there yet. And I can change settings but as far as what they actually do sometimes it's not that obvious to be able to log anything, and sometimes I can see what it does either in performance or CPUz. It's a crap shoot LOL!! :D

When I started on the Rex I was lost too. Granted I understood the bios better than the pics you posted but still... I used prime95, super pi, and 2k6 for testing. I would only change one thing at a time. For example with clock skews for the nb. When I was getting close to 600fsb, I was using super pi and prime for testing. I would change the skews one level at a time, reboot and test. First thing I did was run super pi 1m since it's very sensitive to your stability. Sometimes you will be a little faster by upping vcore and making the cpu mroe stable. Of course run it over 5 or 6 times to get a clear result.

Then I would run prime95 small fft for a few minutes and then blend for a few. This is really where you find stability. Super pi 1m was just a quick speed/stability test. If I failed a thread in prime, I would write down how long it took to fail and what core. If it lasted longer than the last time I ran prime, I changed the clock skews more. If it failed at the same place, I would go back to where I started before trying skews. Make sense?

Yeppers.... that's pretty much what i've been doing, but I fear that until I get some extreme cooling on this bad boy i've gone about as far as I can go.

Pics Have been added to the first post in the Members Rigs forum!:D