Overclocking Intel's New 45nm QX9650: The Rules Have Changed

@ 2007/12/19
Intel has once again re-energized the overclocking community with the recent release of the Core 2 Extreme QX9650 quad-core processor, the first of many new mobile, desktop, and server CPUs fabricated using their radically new 45nm process technology. Early results have highlighted Intel's extremely conservative approach in assigning processor power specifications and because of this, users looking to push their systems beyond rated speeds are finding they have a lot more headroom than normally expected. Our overclocking trials have met with great success, as exploiting this processor's hidden performance margin is easier than ever imagined.

Comment from Rutar @ 2007/12/20
really good article indeed, and I am a very critical reader


Those 150 to 200Ws for 4 Ghz however scare my 480W a make it want the Wolfdale 8400
Comment from Kougar @ 2007/12/20
Are you serious, he was worried about the technicality of it? This article should be stickied on every forum I know, its simply perfect. It fits with my observations, answers hard to find questions, and gives some much needed common sense in regards to overclocking.

For example only one other site that I recall has started mentioning that vdroop is actually done by design, but they didn't go into specifics as Kris has done. To think that ASUS actually listened to wrongly justified complaints and implemented Load Line Calibration as a setting in their boards instead of explaining what vdroop really is...

I am puzzled though, any Gigabyte board offers vCore options below those on that P5E3, even their P31-DS3L budget board goes as low as 0.51v. My DQ6 offers 0.500v as a setting, and actually will run my Q6600 stock frequency using 1.15v.

If X38 uses a 90nm process and a stock vMCH of 1.25v, is P35 the same in both regards? Intel's datasheet does not say. I am still only on page 9, but I have a funny feeling my random Q6600 400FSB bug may have been due to not enough vMCH...
Comment from Massman @ 2007/12/20
Like I've already mentioned to Thorgal, this is a world-class article
Comment from thorgal @ 2007/12/19
Quote:
Originally Posted by geoffrey View Post
Just got through it, the best technological article since years, I'm quite sure no one else then Kris could write something like this together. This is what TTR was all about, seems like his knowledge will finally meet the big crowd.

For many, some misteries are unveeled, but many are still unanswered. I bet the FSB wall on CPU's will be on Anand's list in the future, multiple people have been brainstorming lately about modding the Front Side Bus terminology, maybe one may find the right tricks soones or later.
Talked to him this afternoon about the article, you know he was afraid that people wouldn't like it because it was too technical ? Told him that if people are serious about overclocking they'd better get technical No, honestly, this is indeed one of the best OC articles ever to make it on Anand. Hope there's more to follow.
Comment from geoffrey @ 2007/12/19
Just got through it, the best technological article since years, I'm quite sure no one else then Kris could write something like this together. This is what TTR was all about, seems like his knowledge will finally meet the big crowd.

For many, some mysteries are unveiled, but many are still unanswered. I bet the FSB wall on CPU's will be on Anand's list in the future, multiple people have been brainstorming lately about modding the Front Side Bus terminology, maybe one may find the right tricks sooner or later.
Comment from Sidney @ 2007/12/19
Quote:
Originally Posted by Kougar View Post
Very interesting, and only halfway through this! Great to finally have some CPU-only power consumption figures to cross reference with!

Makes my stable 3.2GHz overclock of a Q6600 using stock 1.2v seem downright bad

Making your Q6600 @3.2 1.2v downright sensible
Comment from Kougar @ 2007/12/19
Very interesting, and only halfway through this! Great to finally have some CPU-only power consumption figures to cross reference with!

Makes my stable 3.2GHz overclock of a Q6600 using stock 1.2v seem downright bad
Comment from geoffrey @ 2007/12/19
Sweet spot?
Comment from Sidney @ 2007/12/19
Quote:
A good overclock is not always about maximum speed. Instead, we would like to advocate what we call the "intelligent overclock" - find that point where going higher just doesn't make any sense…then simply stop and tune from there. We have provided many of the tools for doing so in this article; now it's up to you to make it happen.
Well said, as to everything else; should have taken the major in Math or Econ.

The rules have not been changed since day one.
Comment from thorgal @ 2007/12/19
Aha, Kristopher Boughton has entered Anandtech Or did he fall for their charmes

As always, great article from the co-founder of TTR.