Hyperhorn

Wow, great stuff! I've seen you used 2.6 V for a validation at 772.6 MHz. Do you remember the Vdimm used for >800 MHz?

Maybe it's a hidden message to Christian Ney. I'm looking forward to more test results ...

These two settings cause >20s difference? That's impressive ... Did you test in which way it affects stability/OC potential?

Hah, that's great. Don't tell me that's an i7-990X that can run Super Pi at high clock frequencies and isn't crippled to low 6.x GHz numbers.

Did somebody test if you can use the 125 or 166 MHz strap and stay at ~100 MHz? I was interested in the performance loss and if it can affect max BCLK in the 100 MHz range, but it's hard to get a clean boot after leaving the UEFI menu. However I managed to run 100 MHz at 166 MHz according to the UEFI strap setting, but couldn't see the heavy performance loss I expected to see. In theory it should reduce e. g. the PCI-E-frequency to ~60 MHz this way, right? So I tend to think it's some kind of a bug and strap settings are bound to the BCLK (range) they're named after. So what's your experience/knowledge about this topic? I use a P9X79 Deluxe (UEFI 0650) + i7-3960X ES.

LOL, epic win!

Massman explained his point of view and given the role he has at Hwbot I think you have to accept it (or mention at least some new aspects to discuss about even if it is unlikely that this will turn around anything). However can somebody please rename the thread title? "Voting" looks a lot better.

Cold air, you know. j/k Unfortunately many LN2 results in the hwbot database are tagged with "stock cooling" even newer submissions like this one: http://hwbot.org/submission/2215934_yoko_cpu_frequency_fx_8150_7758.4_mhz

This can happen e.g. if you change the CPU multiplier with AMD Overdrive. Then CPU-Z also reads CPU-NB-multi as 4x.

Hard to tell, because AMD Overdrive, Core Temp and Aida64 were unable to get reliable temperature values with this particular CPU. With default clocks/voltages temperatures shown @ Idle were 9 to 18 °C and @ Load around 35 °C, so because AMD can't trick physics I didn't care about the values. Maybe I will try an Asus tool later. btw this result was achieved by booting @ 202 MHz reference clock (UEFI value undercuts around 0.7 MHz) and then raising the CPU multiplier with AMD Overdrive simultaneously for all cores. With two Integer-cores enabled I couldn't achieve a single additional MHz reference-clock-wise at 26x, so for a quicktest that's basically the maximum frequency with this cooling solution. 53xx MHz should be possible with fine-tuning, but you will need LN2/LHe to make this CPU shine MHz-wise anyway.

Congrats, great Cedar Mill overclocking!

Hm, if the latency bug is the only reason this benchmark isn't rewarded with points, why not let make a hwbot version with latency measured shown but without affecting the final score, so all that counts would be the copy/read/write results?

PCGH version: http://www.pcgameshardware.de/aid,844385/Overclocking-Weltrekord-mit-Bulldozer-FX-8429-MHz-von-AMD-erreicht-Intel-geschlagen-Jetzt-mit-Video/CPU/News/ Translation: http://translate.google.de/translate?sl=de&tl=en&js=n&prev=_t&hl=de&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.pcgameshardware.de%2Faid%2C844385%2FOverclocking-Weltrekord-mit-Bulldozer-FX-8429-MHz-von-AMD-erreicht-Intel-geschlagen-Jetzt-mit-Video%2FCPU%2FNews%2F

Congrats, really didn't expect that Bulldozer goes >8.4 GHz even before launch. When we look back at the Phenom II launch the clock frequency record back then with LHe was beaten by LN2 users later. So I'm really excited about the scores we will see when Bulldozer is finally available. (2D/3D-stable is different from CPU-Z, so that will be interesting, too )

My guess: "+1" --> Christian agrees with you regarding happy botday "btw for non italian guys, asd = lol" --> Christian explains the meaning of "asd" for dumba**es like me, who only know the classic "lol"

Congrats Sam, that's a fantastic result! I really didn't expect >8.3 GHz to be the next CPU-Z WR, rather 8.25-8.27 GHz. I knew you wouldn't stop before you get the WR, but I'm glad to see all the hard pretesting work, buying/selling and batch hunting worked out for you as it did for the older OC in % WR.

This thread leads to nothing. From an outsider's perspective it's a bit like [infinite loop]"Know what? You suck!" "I don't suck, you sucker! Tell me why I suck so I can prove I don't." "Forget it! I would tell you why you suck if you weren't such a sucker." "Doesn't matter, because I know I don't suck for sure![/infinite loop] I suggest to discuss this somewhere non-public (or not at all if there's no change of behaviour) and Hwbot forum moderators should prevent that the involved community members goad each other in the near future. However that's nothing I can decide, so I'm out of this thread already.

Hm, I can't agree with "the windows line up exactly". While the desktop scheme, placement of (different) icons and all CPU-Z/3DMark windows is identical, there are minor, a few pixel small differences regarding the positioning among the windows. From my personal experience I know I have dozens, if not hundreds of screenshots on my HDD which have the same minor differences as you tend to place the windows on the same spot everytime even after a restart, because sooner or later you will find out this or that way there are all necessary windows and there is no crucial data missing. I know we're not talking about a single user but two individuals here, but I think it's more likely you tend to place windows the same way if you're benching together a lot (compared to two random-picked guys, who don't know each other in real life). I'm not saying these results are fine in every aspect (not willing to evaluate), but for me it is excluded that the screens were done in a row without an inconvenient re-placement of all 3DMark 03/CPU-Z windows (why would you do it that way and not choose a whole different placement, if you wouldn't like to *get caught*?) or a restart as there are the small, a few pixel small differences.

Amazing, I didn't expect a new K10 MHz record these days! How does it feel to break your own record? Congratulations!

Great result! btw it seems to be a nice place for outdoor-benching.

You're doing it ... ... so wroooong What did the poor hardware do to you?

If you're loosing almost 40 seconds on Sandy Bridge that's not an effect of an usual software tweak you don't use but the others. It's likely the OS you use is a slow one or flawed and the hardware settings are way different from the settings other overclockers use. Minor things can have a big effect sometimes, a bugged BIOS/UEFI is also a possibilty. I think it would be helpful to show some screenshots of your results with the usual CPU-Z info (subtimings read-out with e.g. Aida64) so the community can try to spot where you loose performance. Maybe you have just overlooked something.

Bulldozer |= Llano

Nice result! I wasn't able to push my Crosshair III Formula to almost 400 MHz reference clock like you did with your Biostar board. With the right mainboard it should be very easy to get >100 % OC out of a X2 260u without subzero cooling (as you showed us almost with 97 % OC), which can't be said about much K10-CPUs.

Wow, very strong chip. Congrats! Lucky draw or very much cherry-picking (=my guess) involved? However I think it is possible to optimize the RAM-timings a bit without pushing too hard, e.g. 7-7-6.