*Official Retail G0 Q6600 Overclocking Thread*

[Orbitech]

LOL

It will only go down as the thermal paste wears in also.

what cooler do you have for your C2Q? I used a left over thermtake retention bracket for hardmounting my FuZioN because I didn't like to just have screws hrough the motherboard without backing.

When I saw SPI running @3.6 (1.26v) I couldn't stop laughing with my friend..
My E6600 needs 1.525 to run @ 3.6GHz.. Yes I know incomparable things in many aspects but still way to funny to see a quad being so efficient..

As for the cooler ,TR Ultra 120 extreme.. For 24/7 3.6GHz is a piece of cake for air now with low temps most importantly... Yes it can only go lower with the Ceramique the next 2-3 days ..

Haven't tested higher than that with air, I think I'll wait for my new w/c setup and then I'll see what it can do further.. Very promising though for now..
L733B (vid 1.2250) of course..

[ZenEffect]

i reconfigured to a slower 3.6 Stable...
and its actually making me re-think my previous settings for 3.94...

[Counter CS]

Hi, there is some test FSB with my DFi DK P35

[Counter CS]

ZenEffect we posted at the same time

[ZenEffect]

ZenEffect we posted at the same time

ya!

i was also thinking... it took less voltage to get 3.6 stable than the actual vid of my previous processor 1.325!

[GripS]

I keep seeing high FSB clocks with low multi's on the quads with the DK board. Let's see an 8 hour run of prime of at least 3.6ghz on a quad.

[arfett]

Would I really get any noticable performance gains using 450x8 than I am with 400x9?

[kberguy1]

I spent the last two or three months struggling with a hot and not very speedy B3 Q6600 CPU. I was eventually able to get to 3200 MHz with it, but it took 1.4 volts in BIOS and an hr of lapping IHS to 2000 grit finish to do so. My Prime95 temps at this speed were between 68 and 72 C. At one point, I thought my Patriot Pc-8500 memory sticks with Micron D9 GKX IC's were the problem, so I spent $400+ to get high end Corsair memory with Micron D9 GMH IC's (P35 chipset based motherboards seem to like the GMH Ic's better).

Late last week, I replaced it with a new Q6600 G0 stepping processor. I had planned on getting a new E8500 processor, but it was on back order. This gave me an opportunity to do a comparison between my old B3 processor and the new G0 stepping in the same system, so I set my new CPU to 3200 MHz and worked with it to obtain its minimum Prime95 stable settings.

Here is what I found: My old CPU needed 1.400 vCore to get to 3200 MHz, while my new one needed only 1.2175 volts. My old CPU had load temps of 72, 70, 70, 72 C, while my new CPU had load temps of 43, 39,39,43 C. Note that the old CPU's IHS had to be lapped to try to equalize core temps, so its mating surface with the lapped Thermalright Ultra 120 Extreme HSF was definitely an advantage for heat transfer.

My new G0 stepping CPU has a 8hr Prime95 stable OC of 3752 with 1.3750 vCore in BIOS, but I have my everyday setting at 3600 MHz (9 x 400 MHz FSB) with a vCore of 1.2875 in BIOS (1.296 as shown in CPU-Z). Under Prime95 with these everyday settings, my load temps never exceed 56 C and are usually between 51 and 54 C.

In the end, the difference between a B3 and a G0 processor for me was this: -.1825v (for G0 at max OC speed of B3 compared to B3), -.0250v (for G0 at max 8hr Prime 95 stable OC compared to B3 at its maximum 8hr Prime 95 stable OC), -29-31 C (for G0 temperatures at max OC speed point of B3 compared to B3), and +552 MHz (the current difference between the 8 hr Prime95 stable OC points of my G0 and B3 CPU's).

[jonny_ftm]

If you can do 3.6ghz on 1.4v, then you certainly do NOT need 1.35 for 3.2ghz, voltage is NOT linear correlation to frequency. Therfore my point was that he would be using unnecessary volts, and unnecessary wear to his CPU, and if it were stable it would require closer to stock voltage for 3.2ghz.
So what exactly is it your baiting over?

Mines pretty new batch, and I have ~1.43v set in BIOS which is only 1.35v real.
So 1.4v just seems normal for 3.6ghz, mine needs a lot more voltage to go over 3.7ghz stable so I dont bother. 3.6ghz stable and lower volts than my E6600 is good enough for me.

Those affirmations are what made me
The words "always", "never", "certainly" shouldn't be used by any real overclocker, as every chip is different. It is well known, and even was lastly published with nice curves at anandtech, that there's a linear relation between vcore and frequency for most CPUs. But, this stops till a point where the curve becomes exponential. That is, a minimal frequency increase will need much more volts than before. The more you claimb in frequency, even higher will be the vcore needs. For most G0 Q6600 chips, that limit is between 3.6-3.7 GHz (it is also your case). For many it is near 3.4 GHz, and for a few of us it is between 3-3.2 GHz. Just read through the forums (XS is not the only one, heh...) and you'll realize it.

I didn't mean in anyway to be rude, but if someone has an unstable OC on small FFT, suerely it is not due to a high vcore, unless it is too hot. And yes, that's at least something for sure. So, yes, he can be stable at 1.4v for 3.6GHz and unstable at 3.2GHz with 1.35v. Lowering his vcore while he's unstable is not a good answer from you, right?

My two cents, take it or leave it and no need to comment on this thingy for longer.

[The0men]

Well I think you better look up an exponenial graph then. Exponention can start almost running dead flat, and end up at a complete right angle, but the change becomes more promenant at higher frequency. Obviously. There's nothing linear about it, just that notice it more because the curve becomes more promonant.

I would put money that to get 3.7ghz for example you would need alot more than 1.4v, thats called an exponential graph.

[spajdr]

Talking about 3.7Ghz .. while for 3.6Ghz i need 1.4V with my Q6600, when i run 1.425V and 3.7Ghz, prime runs for few minutes, then black screen, BSOD and restart, where could be the problem? NB and SB cooling?

[The0men]

My CPU similarly needs 1.4v for 3.6ghz but need 1.5v for 3.7ghz.

Exponential curve. What I was trying to explain above.

So I would say, simply not enough CPU voltage.

[M-XXXX]

What are your max FSB on air/water?

[P e r p l e x]

My CPU similarly needs 1.4v for 3.6ghz but need 1.5v for 3.7ghz.

Exponential curve. What I was trying to explain above.

So I would say, simply not enough CPU voltage.

You can tell where my chip just wont get any higher no matter how many volts I give it.


Core Voltage - 1.4275
Memory - 5-5-5-15 1:1
Frequency - 378.00 X 9.0 (3402.00)
Load Temp – Under 60
Stress Test: PASS

Core Voltage - 1.4275
Memory - 5-5-5-15 1:1
Frequency - 380.00 X 9.0 (3420.00)
Load Temp – Under 65
Stress Test: PASS

Core Voltage - 1.4275
Memory - 5-5-5-15 1:1
Frequency - 382.00 X 9.0 (3438.00)
Load Temp – Under 65
Stress Test: PASS

Core Voltage - 1.4275
Memory - 5-5-5-15 1:1
Frequency - 384.00 X 9.0 (3456.00)
Load Temp – Under 65
Stress Test: FAIL

Core Voltage - 1.4375 (bump .01 since I failed)
Memory - 5-5-5-15 1:1
Frequency - 384.00 X 9.0 (3456.00)
Load Temp – Under 65
Stress Test: PASS

Core Voltage - 1.4375
Memory - 5-5-5-15 1:1
Frequency - 386.00 X 9.0 (3474.00)
Load Temp – Under 65
Stress Test: FAIL

Core Voltage - 1.4475 (bump .01 since I failed)
Memory - 5-5-5-15 1:1
Frequency - 386.00 X 9.0 (3474.00)
Load Temp – Under 65
Stress Test: FAIL


Core Voltage - 1.4575 (bump .01 since I failed)
Memory - 5-5-5-15 1:1
Frequency - 386.00 X 9.0 (3474.00)
Load Temp – Under 70
Stress Test: PASS

[sesdave]

Not a lot of information to assertain if its your CPU reaching it max capability or not. Whats your setup and BIOS setting.

[The0men]

Lol they should continue to increase in speed as you increase the voltage. You only need to look through previous results. Who is to say something else like NB voltage is not limiting you? Or other temps. Nor was your last increase very much as they need more voltage for each increase in frequency after a certain point from stock.

There is hardly any information, not to mention your last run was a pass, so increase speed again and voltage untill it fails. Though of coarse not many people want to flog their hardware. Of course I can't say anything for certain because I tend to get flamed.

[TheOnlyBaD]

having problems with my quad. it seems that i'm using only 2 cores. as u can see, even in idle temps are realy strage what should i do?
configuration:
CPU RATION CONTROL- [manual]
RATION CMOS SETTINGS- [9]
C1E SUPPORT- [disabled]
MAX CPUID VALUE LIMIT- [Disabled]
VANDERPOOL TECHNOLOGY- [Disabled]
CPU TM FUNCTION - [Enabled]
EXECUTE DISABLE BIT - [ Enabled]

Ai OVERCLOKING - [manual]
CPU RATIO CONTRO - [manual]
RATIO CMOS SETTINGS - [9]
FSB STRAP TO NORTH BRIDGE- [auto]
FSB FREQUENCY- [333]
PCI-E FREQUENCY- [auto]
DRAM FREQUENCY- [auto]
DRAM COMMAND RATE- [2T]
DRAM TIMING CONTROL- [auto]
DRAM STATIC READ CONTROL- [auto]
TRANSACTION BOOSTER- [auto]
CLOCK OVER CHARGING MODE- [auto]
CPU SPREAD SPECTRUM- [disable]
PCI-E SPREAD SPECTRUM- [disable]
CPU VOLTAGE- 1.28
CPU VOLTAGE REFERENCE- [auto]
CPU VOLTAGE DAMPER - [Enabled]
CPU PLL VOLTAGE-[auto]
DRAM VOLTAGE-[auto]
FSB TERMINATION VOLTAGE- [auto]
NORTH BRIDGE VOLTAGE- [auto]
NORTH BRIDGE VOLTAGE REFERENCE- [auto]
SOUTH BRIDGE VOLTAGE - [auto]

[Jupiler]

Try Prime v25.5 or 25.6.
When you start it, it will automatically launch the 4 cores.

[TheOnlyBaD]

Try Prime v25.5 or 25.6.
When you start it, it will automatically launch the 4 cores.

downloaded it , installed it, opened it, got blu screen

[Schlotkins]

I posted this over in the support area, but given this is the Q6600 G0 thread I thought I would ask here.

The original bad axe does not have 'official' support for ANY quad cores. However, people have indicated the Q6600 B3 stepping worked on the board perfectly. Does anyone here know if the Q6600 G0 stepping works on the original Bad Axe? If so, any reason to be worried without official support?

Thank you,
Chris

[KTE]

Well;

Bench means nothing for 24/7 -> Kentys run hot.

I can fully bench 3.8G 1.3V all loads, maybe less, never tried.
3.6G needs 1.264V to complete all benches.
BUT
It requires 1.264V idle real for 3.4G stable -> normal air cooling though.
3.6G @ 1.32V idle real will only fail P95 small FFT/large FFT on core 3. Passes blend, which is easier and less core stress.
3.6G @ 1.332V fails again on core 3.

Haven't tried more P95 but Linpack 10mins failed 1.352V idle/1.344V load real too. Failed in the calculation outputs. What does that mean? It means while the system runs at max loads perfectly fine, to the naked eye apparently, the actual computations are all mostly corrupted and slower than 3.6G should give. So you get less than 3.4G performance.

[Psycho_eddie]

Currently

400 x 8 (3200Mhz) @ 1.296V Loaded

Mobo asus P5E




Temps 55C loaded on True 120EX


Vid 1.3150



Want 450x8 for 3600Mhz what volt would be max recomended?

[sesdave]

Currently

400 x 8 (3200Mhz) @ 1.296V Loaded

Want 450x8 for 3600Mhz what volt would be max recomended?

I need 1.415(BIOS) for mine ,spme need more,some cant get there whatever but the latest ones seem to be achieving 3.6 on less than 1.4v. As long as your temps look good give it what it needs to achieve prime stable.

[KTE]

Psycho_eddie have you tested your CPU can do 450FSB at low multi and the volts needed? That's your starting base.

[Kondik]

2 Days ago i decided to "refurbish my PC" so i bought an Q6600, funny part was it's G0 (5) revision and from 19.1.2008 factory date. Put it on 333 FSB so its 3.00 GHz now and im happy with that but i pu GeminiII with Noctia P12 on it [ one] and the temps are 34 - 38 idle and load 51 - 53. Isn't that too high with 1.15 V ?