***The Official Yorkfield Q9300/Q9450/Q9550 Overclocking Thread ***

[Deluxe1]

That's not to shabby Deluxe1! 3400MHz Quad Penryn is nothing to sniff at, look forward to some benchies once you finished your stability testing. I wonder if those temps are correct? Seems a little bit high for 1.232vCore but I guess it depends on your cooling?

Ive got a Zalman CNPS9700-NT and its always been great at cooling my previous chips.

Ive re-applied some thermal paste but my temps are more or less the same,37 37 38 38 idle and full load 63 63 65 65.Its a cool day today so not looking forward to my temps in the summer...

I think the problem may be that my IHS isnt as flat as it could be on the quad,heres my E6600 & Q9450 for comparison:

E6600

Q9450

 

[Banny]

I lapped mine today, its not totaly flat but very close (checked with razer blade) also did the heatsink.
Not made a great difference to temps, but it had bought the cores much closer together only one degree difference now, instead of 3.
I maybee could have got it flatter but was worried about taking too much off it.
They seem to bow in the middle.

 

[Bin Boy]

@BinBoy: are you overclocked at all ?

Not overclocked yet i like to let the cpu run in a abit first.

I read a review that the thermal paste that came with thermalright generic was actually better than AS5 and out of all the test products there was a +-2 between them all.

 

[Big.Wayne]

i like to let the cpu run in a abit first.

Yup I'm the same, sus out the stock temps and such like, run a few stock benches etc.

 

[VoG]

ok, given it a 12hour overnight prime ( small ffts ).
1.4v at idle, 1.35v under load.
tad underwhelming tbh, especially when you consider my G0 did 3850 stable on the same vcore.clicky clicky

temps on the Q9450 peaked at 53C last night with the heating on.

 

[psychas]

water or air ?

edit : nvm opens his eyes

 

[hardtarget]

Nice Review of the Intel Core 2 Quad Q9450.

 

[drak3]

So... on average these will do 3.5 for now?

 

[psychas]

does not look that good at 3.7 just a bit higher scores then q6600 on stock what about if q6600 on 3.6 vs q9450 on 3.6 or q6600 on 4 ghz

 

[Legion]

From everything I've seen and read.3.8 is the "top end" of Q9450 and xeon X3350.
The simple truth is you wont see the magic 4ghz from anything less than a QX9650,run 24/7.........
It's my sincere belief however that this "could" be solved by a platform change as everything in the equation points towards the platform itself rather than the sum of it's parts.
By that, I'm stating that current motherboards just cant and don't deliver power to quadcore chips in the same way they deliver it to dual core at high fsb's.
I'm not blaming mobo manufacturer's perse,I'm pointing the finger straight at intel,as there is a very good reason they label/multi-unlock QX chips.
I really cant see the way this will be changed until the release of Nehalem.
I.e. mem controller taken off-board and placed on chip,releases the way for point-point interconnects.........
No guarantee that oc-ing will be any better but we'll at least not be suffering from platform limitations which we are starting to see and are suffering from now.

 

[Hesky82]

By that, I'm stating that current motherboards just cant and don't deliver power to quadcore chips in the same way they deliver it to dual core at high fsb's.

I think its pretty specific to 45nm chips though, they really do seem to have an agressive FSB wall. Compare Kentfield quads to these new Yorkfield quads. on average you can obtain a higher max FSB from a G0 stepping Kentfield than you can with a Yorkie. Mine happy hits 500FSB with slightly Higher PLL and NB voltages. Worth noting that the Kentfield cores consume more power than the new quads and should stress the mb cpu power regulation circuit more.

Same can be said about those Wolfdale dual's, G0 stepping Conroes easily hit higher FSBs than current Wolfdale chips. I stuggled to get my E8400 to 500FSB stable, really needed a large vcore boost.

 

[drak3]

They could at least sell a higher multi for a premium instead of an arm and a leg for the unlocked QX

Wish I was earning 5 grand a month like the average city IT guy, but at the moment I dont :/

 

[weescott]

no good clocks on these yet.

 

[Hesky82]

They could at least sell a higher multi for a premium instead of an arm and a leg for the unlocked QX

Wish I was earning 5 grand a month like the average city IT guy, but at the moment I dont :/

Tell me about it! I really want a QX Yorkie! just could never ever ever justify its purchase, and the non extreme ones just aren’t at all tempting

 

[Banny]

The older 65nm chips overclock much higher FSBs, my 6300 did well in excess of 500fsb.

 

[Hesky82]

Some chap got 700fsb out of an E6850 on an MSI P35 board. He was forced to lower the multi to 6 just to max out the fsb lol

 

[fornowagain]

http://www.anandtech.com/weblog/showpost.aspx?i=428

High VTT sends our best QX9650 to an early grave...

We have been using this particular QX960 processor for an upcoming X48 comparison article and testing was going quite well (not blaming a particular board somehow seems appropriate). We thought we had made it past that mystical, magical, and more importantly, stable mark of 450FSB with a Quad Core. However, in order to get to that point on this particular board, we (Editor- Raja in case the police get involved) made the mistake of using a very high VTT termination voltage of 1.51V (VTT is used to terminate data lines between the MCH and CPU).

We should have known better really, especially after Kris posted up a recommended voltage-operating chart in the ASUS Striker II Extreme review. We do not know what is worse now, Kris saying I told you so (jokingly, well maybe not) or the thought of replacing this $1000 CPU. We thought it would be beneficial for others to learn from our mistakes so here goes.



This particular CPU was a great example of a QX9650 hitting 4GHz stable at a mere 1.29V on most motherboards (tears are flowing in the beer tonight). We know users are running VTT voltages even higher than ours on 45nm processors and probably have not had a problem yet. We will run high VTT voltages in short bursts to test the limits of the board and CPU. However, this is the first time we have tried anything over 1.45V on a 24+-hour basis to test application stability.

Let this be a warning – do not go over 1.4V maximum for 24/7 use! We are certain that the high VTT voltage and extended testing was the cause of death, as we made no other major or obvious changes within the BIOS that could have instigated a failure. Obviously, we tried to boot the processor in a number of other motherboards without success before we decided to post our results up.

This is our first 45nm Quad core processor we managed to kill outright during testing. We hope it is the last one too. The problem is that we also have a Q9300 that is on life support after experiencing a 36-hour run at 435FSB with VTT set to 1.45V. While our experiences might not represent results elsewhere, we thought our advice to just, “Say no to high VTT” is worth a quick post. We have had enough VTT/GTL adjusting in the last year to last us a lifetime – just give us Nehalem quickly, please.

 

[Hesky82]

Doh, I would feel sick after doing something like that, killing a great clocking processor is such a horrible thing to do.

VTT(PLL) can destroy a CPU quicker than Vcore thats for sure! why on earth I can set mine to 1.9v when running a 45nm processor I will never know

 

[Big.Wayne]

VTT(PLL) can destroy a CPU quicker than Vcore thats for sure! why on earth I can set mine to 1.9v when running a 45nm processor I will never know

I don't even know what the VTT/PLL adjustments are? It seems to be a new thing that came along when people started overclocking quads?

Pretty sure there was no option like that on my Old P5B-Deluxe but I can see it now on the P5K-E, I just leave it set on auto!

Do people use those VTT/PLL adjustments on dual-cores?

 

[Hesky82]

The mc/cpu phase locked loop clock signal circuit has always been implemented, only now we can adjust its phase amplitude. It is a ridiculously sensitive bit of circuitry and I’m surprised there is an option in many boards to over-volt it so significantly. I’ve definitely seen gains on my quad but fear to up it on my Wolfdale.