Overclocking Escapades: Intel socket LGA775 Core 2 Quad Q6600 SLACR

This one is just plain disappointing. Think I may have one of the worst Q6600's.

I got it to 3.2GHz and then it just absolutely stopped scaling, like hitting a brick wall. I got 3.3GHz stable, but it requires a ridiculous amount of voltage, far more than I think should be necessary. It's definitely more voltage than I would want to run for a 24/7 stable overclock.

Intel Core 2 Quad Q6600 G0 #1 (date code: L804)
VID: 1.325v
2.4GHz (266x9) (stock): 1.25-1.28V (1.325V)-not tested for lowest voltage
2.8GHz (312x9): 1.25-1.28V (1.325V)-not tested for lowest voltage
3GHz (333x9): 1.25-1.31V (1.35V)
3.1GHz (345x9): 1.31-1.38V (1.41250V)
3.2GHz (356x9): 1.38-1.44V (1.48750V)-G33M-DS2R 1.368-1.424V (1.46250V)-EVGA nForce 780i SLI
3.3GHz (413x8): 1.432-1.480V (1.51250V)- EVGA nForce 780i SLI

I'm so surprised, I read a lot of reviews and threads about people getting to 3.4-3.6GHz with little or no trouble. My first sample really doesn't seem to be capable of that with ambient cooling though (I tried both air and liquid cooling).

Intel Core 2 Quad Q6600 G0 #2 overclock settings (date code: L732)
VID: 1.25v
2.4GHz (266MHzx9) (stock): 0.976v (1.00v)-EP45-UD3P
2.5GHz (278x9): 0.992v (1.025v)-EP45-UD3P
2.6GHz (289x9): 1.024v (1.05v)-EP45-UD3P
2.7GHz (300x9): 1.04v (1.075v)-EP45-UD3P
2.8GHz (312x9): 1.104v (1.125v)-EP45-UD3P
2.9GHz (322x9): 1.12v (1.15v)-EP45-UD3P
3GHz (333x9): 1.152v (1.175v)-EP45-UD3P
3.1GHz (345x9): 1.152v (1.175v)-EP45-UD3P-51°C
3.2GHz (356x9): 1.168v (1.20v)-EP45-UD3P-50°C
3.3GHz (367x9): 1.20-1.216v (1.2375v)-EP45-UD3P-54°C
3.4GHz (378x9): 1.248v (1.275v)-EP45-UD3P-57°C/ 1.280-1.344v (1.39375v)-G33M-DS2R (Load: 60-59-57-55°C)-Liquid Cooled by Corsair H100i
3.5GHz (389x9): 1.296v (1.325v)-EP45-UD3P-62°C
3.6GHz (400x9): 1.328v (1.3625v)-EP45-UD3P-67°C (Idle: 33-35-33-33) (Load: 67-67-62-62)

Is there any chance you might have an intel board like a p5k deluxe? Right now I don't think I would be so quick to blame the CPU. The 780i is basically the same as a 680i right? I couldn't get my q6600 stable even at stock with the 680i.. I sold the board and got a p5k deluxe wifi and it was night and day. Those nvidia boards were decent with dual cores I think.

freeagent said:

Is there any chance you might have an intel board like a p5k deluxe? Right now I don't think I would be so quick to blame the CPU. The 780i is basically the same as a 680i right? I couldn't get my q6600 stable even at stock with the 680i.. I sold the board and got a p5k deluxe wifi and it was night and day. Those nvidia boards were decent with dual cores I think.

Click to expand...

Nope, don't have any working Asus boards on hand for 775.

I have a Gigabyte EP45-UD3P which is very unstable at anything past 400MHz FSB (some kind of hardware problem with the board), and the EVGA 780i SLI, that's all I've got that's working for benching at the moment. The 780i is doing a hell of a lot better than the EP45-UD3P does.

I hit this same wall with my Gigabyte G33M-DS2R. Different chipset entirely, so I doubt this an issue with the nForce 780i chipset.

i ran my q6600 @ 3.5 @ 1.4 volts for a 24/7 overclock on my 680i, i was able to get 4.2ghz validated on tap water, but not stable with anything else.

With these it's all about the motherboard. There's a thread about overclocking these on a P5N-E SLI and it was barely possible to hit 3.0GHz. I got that motherboard without researching first after I had my old one failed (which was a G32 or G33 but that's all I can remember) but the old motherboard could run it 24/7 at 3.6!

I remember my 680i would do great things with my E8500 but just couldn't do anything with my Q6600. I thought that got mostly fixed up with the 780i's and the FTW versions of the 780i's were pretty good. That being said I know i wound up switching to P5Q-Pro boards from Asus and had much better OCs with both chips. I have a Q6600 coming back to me when I pick it up this afternoon. If you are interested in it maybe we could work something out. .

Got a new Q6600 in, and it's able to run stock frequency at a much lower voltage than I would have thought possible.

It's running stable at 0.976v, I figured the lowest would have been around 1.1v. Pretty astonished by that. The drop in voltage (from the stock 1.25v) gave a nice temperature drop of around 6°C under load.

Testing is still ongoing.

I can definitely say that not all CPU's are created equal. I don't know if my new Q6600 is a "golden chip," or if the old one was a below-average sample, but they definitely are not the same. The new one is SO much better as far as as the voltage/power requirements for given clock speeds, the two don't even compare. It's like they aren't even from the same model line.

I'm running an 800MHz overclock on the new one at a voltage that MIGHT just have been sufficient to run at the stock frequency on the old CPU, and I'm not done yet.

Intel Core 2 Quad Q6600 G0 #1
VID: 1.325v
2.4GHz (266x9) (stock): 1.25-1.28V (1.325V)-not tested for lowest voltage
2.8GHz (312x9): 1.25-1.28V (1.325V)-not tested for lowest voltage
3GHz (333x9): 1.25-1.31V (1.35V)
3.1GHz (345x9): 1.31-1.38V (1.41250V)
3.2GHz (356x9): 1.38-1.44V (1.48750V)-G33M-DS2R 1.368-1.424V (1.46250V)-EVGA nForce 780i SLI
3.3GHz (413x8): 1.432-1.480V (1.51250V)- EVGA nForce 780i SLI

Intel Core 2 Quad Q6600 G0 #2 overclock settings
VID: 1.25v
2.4GHz (266MHzx9) (stock): 0.976v (1.00v)-EP45-UD3P
2.5GHz (278x9): 0.992v (1.025v)-EP45-UD3P
2.6GHz (289x9): 1.024v (1.05v)-EP45-UD3P
2.7GHz (300x9): 1.04v (1.075v)-EP45-UD3P
2.8GHz (312x9): 1.104v (1.125v)-EP45-UD3P
2.9GHz (322x9): 1.12v (1.15v)-EP45-UD3P
3GHz (333x9): 1.152v (1.175v)-EP45-UD3P
3.1GHz (345x9): 1.152v (1.175v)-EP45-UD3P-51°C
3.2GHz (356x9): 1.168v (1.20v)-EP45-UD3P-50°C
3.3GHz (367x9): 1.20-1.216v (1.2375v)-EP45-UD3P-54°C
3.4GHz (378x9): 1.248v (1.275v)-EP45-UD3P-57°C

Your good one seems very good so far TT! That EP45 board should be the one to use I think (unless maybe you have a rex)?

What're you using for cooling? Bios settings?

funsoul said:

Your good one seems very good so far TT! That EP45 board should be the one to use I think (unless maybe you have a rex)?

What're you using for cooling? Bios settings?

Click to expand...

Using my Thermalright Ultra 120 Extreme BLACK for cooling with a push-pull fan setup.

As for BIOS settings, CPU vCore is the setting listed on the right (what I actually see in the OS is on the left). DRAM voltage at 1.92v. Everything else is set to Auto, as I haven't had the need to adjust anything else yet with this CPU.

Will probably end up having to bump the MCH Core (NB voltage) up to 1.3v at some point (usually where I stop on NB voltage), and maybe bump up the CPU Termination voltage (that's the FSB voltage if memory serves me correctly).

To get the most out of the chip, you'll need to up NB and VTT. Also...if you're setting 'final' fsb in bios...you'll want to boot to a lower fsb then use setfsb to crank it up from inside winblowz.

Tech Tweaker said:

I can definitely say that not all CPU's are created equal. I don't know if my new Q6600 is a "golden chip," or if the old one was a below-average sample, but they definitely are not the same. The new one is SO much better as far as as the voltage/power requirements for given clock speeds, the two don't even compare. It's like they aren't even from the same model line.

I'm running an 800MHz overclock on the new one at a voltage that MIGHT just have been sufficient to run at the stock frequency on the old CPU, and I'm not done yet.

Intel Core 2 Quad Q6600 G0 #2 overclock settings
VID: 1.25v
3.2GHz (356x9): 1.168v (1.20v)-EP45-UD3P-50°C

Click to expand...

Looks like a pretty nice Kentsfield sample!

I wonder when it will start to hit its own specific voltage wall??

I'd like to see the voltage requirements for that chip when running stress stable at ~3.6/3.8/4.0GHz?

PolRoger said:

Looks like a pretty nice Kentsfield sample!

I wonder when it will start to hit its own specific voltage wall??

I'd like to see the voltage requirements for that chip when running stress stable at ~3.6/3.8/4.0GHz?

Click to expand...

Well, I found 3.6GHz.

3.6GHz (400x9): 1.328v (1.3625v)-EP45-UD3P-67°C (Idle: 33-35-33-33) (Load: 67-67-62-62)

It runs hotter than I'd want for 24/7 though, 67°C under load is more than I care to see on LGA775. I don't want to see much more than 60°C personally.

I think voltage-wise it can handle a higher overclock, but temperature-wise it is almost at it's limit.

I had to swap the TRUE heatsink out for my Corsair H80 liquid cooler to see if I could get the temperatures lower. It did get it a little lower, brought the maximum load temp down to 60°C, and the idle temps came down about 2-3°C.

And then I opened a window (it was 25°F outside last night), and brought the ambient temp down from 76°F/24.5°C to 68°F/20°C, and with that the load temp came down to around 55°C.

Low voltage/low leak samples run hotter... Back in the socket 775 heyday someone would have loved to run a sample like that under custom water... Running daily with ~1.40v/1.425v/1.45v (load).

Are you benching this setup or just setting it up as a 24/7 boxen (or benching first, then 24/7)? Your load temps are nice and low...plenty o' overclocking headroom left! I generally will let these chips get up around 85C under load (for benching purposes).

Smack that thing around a little more! hehehe

PolRoger said:

Low voltage/low leak samples run hotter... Back in the socket 775 heyday someone would have loved to run a sample like that under custom water... Running daily with ~1.40v/1.425v/1.45v (load).

Click to expand...

Yeah, that's pretty much what I'm seeing here. Wish I'd known that sooner.

It's weird though, on sockets 939, AM2, AM3, LGA1366, and LGA1155, I usually see a temperature reduction when I can lower the vCore. So far this is the only socket type where that doesn't hold true.

With the first sample (1.325v VID, date code L804) overclocked to 3GHz with 1.26v under load (1.31v at idle) I was seeing typical idle temps of 26-30-28-25 and load temps of 46-46-42-39 (that last core always ran cooler, no idea why). With the second sample (1.25v VID, date code L732) overclocked to 3.3GHz with 1.2v under load (1.26v at idle) I see typical idle temps of 32-34-32-30 and maximum load temps of 51-50-47-46 (at 3GHz it was about the same except the max load temp was 50°C).

funsoul said:

Are you benching this setup or just setting it up as a 24/7 boxen (or benching first, then 24/7)? Your load temps are nice and low...plenty o' overclocking headroom left! I generally will let these chips get up around 85C under load (for benching purposes).

Smack that thing around a little more! hehehe

Click to expand...

I was testing it for max overclock, and to give me an idea of what speed I could safely run at for 24/7 usage.

I put it in my main rig a couple of weeks ago to give me a bit more speed in daily tasks. That was the whole idea really, to find a quad that would run faster in my main PC than the one I had, so I didn't want to abuse it too much since I planned on using it in my main PC.

I was planning to set it at 3.4GHz, but the CPU wasn't stable at the settings I was putting in (not enough voltage). It was late at night and I didn't feel like testing to find what voltage it needed to be stable on the main system (main system is very different setup from benching rig, as it has 4x2GB sticks of DDR2 and benching rig only had 2x1GB sticks), so I just dropped it down to 3.3GHz and decreased the voltage accordingly. It's been running stable at 3.3GHz for a couple of weeks now, current uptime is 186 hours now according to Real Temp.