**Official DFI LanParty UT 790FX-M2R Review/Overclock/Guide Thread**

[KTE]

Someone here or elsewhere posted a +15k 9800GX2 2k6 score and 44xx CPU score with a 2.9xG Phenom and 2.3xG NB/HT.
For Phenom, that CPU score is pretty damn good. I'm starting to see a trend here... ATI cards are "eating" the CPU score more than nV cards on this platform.

click for full size including memset...
http://www.3oh6.com/forum_posted/0803/350htt_4gb-1.png

Hey 3oh6, did you really require that high vCore for that setting? By my experience, it's very high, I only required 1.3V to get to 3.35/6G stable with the stock HSF using two of them CPUs.
Also, how accurate is the DRAM voltage given there? If true, looks very good.

By the way, I checked the stepping on my phenom 9600BE, it is CAAWB AA 0747MPMW - any info on that as to whether it appears to be a decent clocker or not, or is it too early to really say yet what steppings are good?

Unfortunately, its not a stepping issue anymore, more luck based. I tried this by having two 9600 both of the same week and batch and both were different oc'ers, one didn't make it above 2.55G screeny and above 2.5G stable and the other hit 3G screenies and 2.8G stable with 1.3V, stock HSF. TRUE on them didn't help nor did a little subzero. That batch I've also had on one and it reached max 2.68G stable, but that doesn't prove anything about yours.
High speed bin yield issues show up like this exactly, large oc variation between same wafer cut dies, as we see with Phenom. Hence why no 3G/2.8G/2.7G officially planned Phenom yet and late 2.6G.

should have gone darkside then if I want to OC?

They oc far better, yep, there is no doubt about that. But the Q6600 G0 is the key on the Intel side IMHO. If you've seen what mine benches on air, then it outdoes Penryn Yorkfields by a good margin apart from the 4-4.5x priced QX9650/QX9770.

I was hoping for close to 2.8GHz, that way I'd be at the same speed as my old opty 165 just with 2 more cores basically.

I would say you can expect this, from what we've seen so far, by only the top 30% or so, but I would say, a realistic aim with Phenoms is 2.65G stable, ~60% chance. Looks like in-between the average, the poor and the good to me, as 2.6G is fairly average ignoring Vista 64b users.
I've owned 6 Phenoms now; 4 have benched 2.95-2.7G, 3 stable plus 2.7G, 2 stable plus 2.65G and 1 stable only at 2.5G. My new 9600BE I've sold this week, might pick up another this week (not looking forward to opening the box) and just lost picking up a 0739 from my uncle 2days ago, damn, really wanted to test it. It was supposed to be the guaranteed high clocking early batch which some early reviewers were getting, i.e. Kyle.

Sorry Mate but unless you get exceptionally lucky (which most don't) reality is that maybe 200mhz above rated clock frequency for 24/7 use is being very realistic ....

Camon Camp, you've been reading the Vista threads too much and ignoring mine!
No messing, honestly, if your sample can't get even 2.6G stable, then it really is bad luck and you need to get someone else to touch the CPU first off the shelf because you're obviously extremely jinxed.

What's the max you've hit fully stable with Phenom yet (you'll need a week for this at least)?
Thought of trying a different board?

@KTE

I didn't forget to post some DFI results, but as you can see I can't replicate the same settings as before; tested a 3-way 3870 CF, fresh Vista x64 install, updates, Cat 8.3 drivers 3GPU CF active,etc... 3DMark06 results lower then just one 3870 , had no time the debug, since 1 GFX wasn't mine, but isn't all bad news, I can hotflash the bios now.

Great, yeah, you have one of the better Phenoms around. What was max HT on the DFI with low NB/CPU Multi?

I wonder whats happening with the 3GPU setup and the lower oc

Been browsing the web, playing music in WinAmp and running Prime95 for 1.5 hours now by just changing from 14x251 to 16x219.

If you test P95, a cheat people use is to carry on doing other tasks while it runs in the background.
As I stated when I first got a 9500, the way to test P95 is to leave the window as active, in-focus, or Windows prioritizes for the Window in focus, and so the results you're getting is not the actual stress test, so you later end up getting instability out of the blue, but it really was nothing but what you would've caught if stress tested right.

If it's 'just' that my 14x multi isn't liked anymore, Im still not very happy.

Could easily be, my 5000+ didn't like the 16x multi.

Here's the two BSOD's I just got at a restart:
BCCode : 1e BCP1 : FFFFFFFFC0000005 BCP2 : FFFFFADF8F80C202
BCP3 : 0000000000000000 BCP4 : FFFFFFFFFFFFFFFF OSVer : 5_2_3790
SP : 2_0 Product : 256_1

BCCode : d1 BCP1 : FFFFFADB90BD00D2 BCP2 : 0000000000000002
BCP3 : 0000000000000008 BCP4 : FFFFFADB90BD00D2 OSVer : 5_2_3790
SP : 2_0 Product : 256_1

Any Minidumps?