Q6600 can it go further?

phekdra · 7 Jan 2008 at 18:11

Hi All,

I've got a new Q6600 and a Gigabyte P35C-DS3R, and it's quite happily running at 3.2Ghz (400*8) at 1.35V core (1.312V according to CPU-Z). I'm trying to get it up to 3.4Ghz but simply upping the bus to 425Mhz causes Windows to freeze after a few minutes of prime95. I've tried setting the core voltage to 1.4V, 1.425V and 1.4375V, but on each occasion Windows freezes after 2 to 3 hours. The last time, according to Speedfan the core temperature was only 50C, so I don't think the temperature is the problem (I have a Thermalright Ultra-120 Extreme).

I also have 2Gb of memory (OCZ PC2-8500C5 running at 2.5X) which I have set +0.4V overvoltage (presumably 1.8v+0.4V = 2.2V!), although Speedfan says it's only getting 1.92V. There's not much to go on, but can anyone from experience tell me whether a system freeze like that is more likely due to the memory, the core voltage, or possibly the north bridge voltage? I don't really want to up the core voltage any further if it seems to produce no gain in stability.

Phekdra

sablabra · 7 Jan 2008 at 18:34

Try and achieve the same speed using 9x multiplier. (380x9 = 3.42 GHz)

phekdra · 8 Jan 2008 at 06:10

sablabra said:
Try and achieve the same speed using 9x multiplier. (380x9 = 3.42 GHz)

Thanks for the hint. 11 hours of prime95 later suggests it was the bus speed that was causing the problem! I can feel 3.5Ghz coming on...

Phekdra

WJA96 · 8 Jan 2008 at 07:13

All my Q6600's except one have needed over 1.5V to go through the 3.6GHz barrier. You also need to take Vdroop into account although the latest Gigabyte BIOS should have a very effective anti-Vdroop system (Loadline Calibration).

phekdra · 8 Jan 2008 at 07:53

WJA96 said:
All my Q6600's except one have needed over 1.5V to go through the 3.6GHz barrier. You also need to take Vdroop into account although the latest Gigabyte BIOS should have a very effective anti-Vdroop system (Loadline Calibration).

I don't think I want to give it more than 1.5V, so if I can manage 3.5Ghz with, say, 1.45V or less then I'll be reasonably happy. 3.42Ghz was achieved with 1.425V and that was a random value I picked out of the air as being a bit more than the 1.35V I need at 3.2Ghz, so I might try it again with less than that!

I might look into Loadline Calibration - I presume you mean F9 as being the latest BIOS. I was briefly using F9 when I was trying to solve my bus speed problems and didn't notice this option, but then I wasn't looking for it. I'm back on F8 at the moment. I was quite worried that despite 1.425V in the BIOS, under load the processor was only getting 1.31-1.33V for long periods. I was also surprised that despite setting the memory voltage to 2.2V, CPU-Z was telling me I was getting less than 1.9V.

Anyway, it's stable so I'm happy!

Phekdra

pastymuncher · 8 Jan 2008 at 11:31

WJA96 said:
All my Q6600's except one have needed over 1.5V to go through the 3.6GHz barrier. You also need to take Vdroop into account although the latest Gigabyte BIOS should have a very effective anti-Vdroop system (Loadline Calibration).

Yep, mine does as well. Then again, this board has terrible Vdroop as well. I would say that most Q600's need more than 1.45v Vcore for 3.6ghz+.

phekdra · 11 Jan 2008 at 18:30

WJA96 said:
You also need to take Vdroop into account although the latest Gigabyte BIOS should have a very effective anti-Vdroop system (Loadline Calibration).

Could somebody with this board please point out where Loadline Calibration is enabled in the BIOS. I've tried F8, F9, and now F10 but can't see it anywhere. :confused:

Unfortunately, the number of people looking for pencil mods with this board suggests that it doesn't actually exist...

Phekdra