E6600 Overclocking

[Kralnor]

I'm attempting to overclock my E6600. My motherboard is a GA-965P-DS4 and I'm using 4GB Corsair XMS2 PC5400.

So far I am up to 2.79GHz (9*310FSB) and Prime95 has been running FFT for 45 mins without a hitch. SpeedFan is showing a max temp of 57C while CoreTemp reads 62C. I am not using auto vcore but have left it at default.

Which temp reading should I trust and how far can I go before temps get too high?

 

[saaiello]

I use speedfan and its always within a couple degrees of my bios so i trust speedfan. I wouldn't let the temps get higher then 68c. You should really use a good after market cooler if your serious about overclocking. Which E6600 you using the core 2 duo or the Pentium dual core?

 

[Kralnor]

This is a Conroe C2D. I am using the stock cooler for now since I wanted to experiment a little with CPU OC'ing before investing in a third party cooler. This is my first time ever OC'ing a CPU so I am taking small steps.

I was recommended the Cooler Master Hyper 212+ in another thread. How does its noise level measure up to the stock cooler? My rig is very quiet and I'd hate to change that.

Core Temp just touched 64C so I'll probably stop here until I get some better cooling.

 

[Kralnor]

Another question that came to my mind was the limiting factor of the RAM. Obviously, I wouldn't be able to go higher than 333MHz FSB without OC'ing the RAM since it's only PC5400.

If I can't get the E6600 above 3GHz even with an aftermarket cooler, it seems a bit wasteful to buy one now that it is completely stable at 2.8GHz.

 

[saaiello]

Depending on your ram you can probably get it to run at 400 mhz The Hyper 212 will be a little louder at full speed but will bring your core temps down if your hitting lower 60s at 2.8 i wouldnt go any higher til you get an aftermarket cooler. 400mhz on your ram shouldn't be too big of a deal you might have to relax your timings a little but should be achievable.

 

[coozie7]

Used that cooler on my C2D E6600 and it's very quiet in PWM (auto) mode, even under load at 3.2GHz.
BTW, you'll soon start to need some extra voltage or the system will not restart, even though the overclock seems stable while it is running.

 

[Kralnor]

400MHz on the RAM? Sweet, I had no expectations for going above 333MHz. The RAM is Corsair XMS2 PC5400.

I'm also aware that the voltage needs to be upped as I go higher, which is what will really ramp the temperature up. So far, there are no problems with restarting.

I guess I'll check whether the cooler fits in my case and then order it. It's not that expensive anyway.

 

[Kralnor]

According to this review the Hyper 212+ goes up to 50 dBA under full load

http://www.frostytech.com/articleview.cfm?articleid=2419&page=5

I'm afraid that's unacceptable. I'm really looking for something more quiet, even if it costs more. The stock cooler only runs at ~37 dBA under full load so I probably want to stay below 40 dBA.

I understand it will probably limit my OC, but to me, noise is more important than speed.

Is there any reliable software for controlling the CPU fan speed? I'm pretty sure my Motherboard's PWM is defective as it used to shut off the CPU fan entirely even when under full load...

 

[Kralnor]

I took the chance and got the Hyper 212+. Unfortunately, it is too loud for me at full speed, so I need some way to control the fan speed.

For now I'm going to run the risk and use my motherboard's smart fan control in full auto mode. I'm not sure which exact settings were causing it to halt in the past, as there are various smart fan settings in the BIOS.

The drop in temps is nice though. Mildly OC'ed at 2.8GHz, the stock cooler gave 65C during Prime95 stress testing, but now it is down to 48C (and only at 500 rpm!), although that is with the sides of the case removed, so it will probably go up by a few degrees. Still, a ~15C drop is very welcome and should allow for more OC headroom.

Does anyone have experience with the Gigabyte 965P-DS4 and its auto fan speed settings?

 

[Kralnor]

So far I have run 333FSB (333*9 = 3GHz) and have had no problems. Prime95 ran for 25 hours without producing any errors. The auto fan speed settings have not caused trouble so far either *knocks on wood*. I think I had some problems when I used legacy fan speed settings, but on full auto it seems fine.

For testing stability, is it necessary to perform a full 24h Prime95 run each time one ups the FSB a little? (Say, 10-15FSB) Or is a quick one hour run enough before proceeding to up the FSB? I of course understand that a longer test is needed to ensure that the system really is stable at given settings, but are shorter tests adequate to gauge how high you can go?

 

[saaiello]

I took the chance and got the Hyper 212+. Unfortunately, it is too loud for me at full speed, so I need some way to control the fan speed.

For now I'm going to run the risk and use my motherboard's smart fan control in full auto mode. I'm not sure which exact settings were causing it to halt in the past, as there are various smart fan settings in the BIOS.

The drop in temps is nice though. Mildly OC'ed at 2.8GHz, the stock cooler gave 65C during Prime95 stress testing, but now it is down to 48C (and only at 500 rpm!), although that is with the sides of the case removed, so it will probably go up by a few degrees. Still, a ~15C drop is very welcome and should allow for more OC headroom.

Does anyone have experience with the Gigabyte 965P-DS4 and its auto fan speed settings?

I warned you it was loud LOL.

 

[saaiello]

So far I have run 333FSB (333*9 = 3GHz) and have had no problems. Prime95 ran for 25 hours without producing any errors. The auto fan speed settings have not caused trouble so far either *knocks on wood*. I think I had some problems when I used legacy fan speed settings, but on full auto it seems fine.

For testing stability, is it necessary to perform a full 24h Prime95 run each time one ups the FSB a little? (Say, 10-15FSB) Or is a quick one hour run enough before proceeding to up the FSB? I of course understand that a longer test is needed to ensure that the system really is stable at given settings, but are shorter tests adequate to gauge how high you can go?

What I do is I will let it run for about 20 minutes on small ffts then i will let the blend run for about another 20 minutes if all is good after that you can proceed. But once you find a clock speed you want to keep run the blend overnight atleast.

 

[Kralnor]

What I do is I will let it run for about 20 minutes on small ffts then i will let the blend run for about another 20 minutes if all is good after that you can proceed. But once you find a clock speed you want to keep run the blend overnight atleast.

Great, thanks for the tip! Time to try 355 FSB

 

[jsc]

Right. 15 minutes is all you need to drive the temps to max.

If you find that your memory is becoming the limiting factor, relax the timings from 4-4-4-12 to 5's-15. If that doesn't help, you can increase your memory voltage to a max of 2.2 volts.

 

[Kralnor]

Yeah thanks, I'm going to try relaxing timings. Just got my first error in Prime95 at 345 FSB. Strange since I ran it for 12h without any errors but just now it gave an error in only 4 minutes.

 

[Kralnor]

It seems like I have hit a wall at 3.1GHz (345 FSB). I have run Prime95 using both Small FFTs for 12h and blend for 12h without any errors. All I had to do was up the cpu voltage one notch from 1.3500V to 1.35625V compared to at 3GHz.

At 3.2GHz I can still boot into Windows fine and use applications like normal, but Prime95 instantly fails on one core. I have tried 1.36250V, 1.36875V and 1.37500V for the cpu but none had any effect. I also tried giving the RAM 2.0 volts rather than 1.8 volts and lowering timings to 5-5-5-15 but that didn't help either.

Since I use four DIMMs I am wondering if the MCH voltage should be increased? Also, according to this http://www.corsair.com/_datasheets/CM2X1024-5400C4.pdf the RAM has only been tested at 2.0 volts. Is it really safe to push it beyond that?