LGA 775 Overclocking Questions

[Roark]

Actually this is probably a general OC question, but here it goes:

What is more of a limiting factor in OC'ing a CPU, motherboard or the chip?

Here is the story.
I'm getting a 2nd Q6600 (G0), the first one I can't OC as well as I would like.
#1 is @ 3.2GHz and requires 1.475v in BIOS (1.432v CPU-Z) to remain stable using an Asus P5K-E, goes to 71°C on IntelBurn.
#2 hasn't been tested yet, but will do so on the board it comes with, an ASRock G31M-S.

I don't really want to be yanking #1 to test #2 in the P5K-E unless I know I'll get 3.2 or higher since its kind of a hassle.
And I don't know if its #1 that is limiting to 3.2GHz, or the board just isn't up to the task (.043 vDroop seems like a lot?).
Is there any way to tell if its worth swapping chips in the P5K-E without actually swapping out chips, like which can undervolt better @ stock or something?

 

[patonb]

I think its board. I can barly get past 3.0Ghz on my q6600 G0

 

[Roark]

Its funny, I ran IntelBurn 2 nights ago and was hitting 71°C.
Just ran it again and was only hitting 61°C, so I guess I have a little more headroom.

Kind of strange though, ambient has only dropped about 2-3°C, so not sure what happened.
Unless my AS5 suddenly cured and decided to drop temps 10°C which I've never heard of happening before.

 

[patonb]

Yh tats AS5.. it needs run in time.

 

[Roark]

Yeah, I just thought it was a gradual curing so temps would slowly drop over time.
And all the reviews/testimonials I read were generally 5°C (+/- 2°C).

So 10°C after ~36hrs is a bit surprising to me, though maybe I just underestimated AS5.

 

[PP Mguire]

Actually, it usually is the chip bin if you have a decent board to begin with. Say, comparing P45 to P45 chipset boards the CPU and heatsink (and the RAM) will be the limiting factor. With my experience most P45 boards can hit a high amount of FSB before crapping out meaning the limiting factor is how high the RAM can go, how high the CPU can clock, and how far you can run the RAM divider on the chip. With these chips most of the time clock is based on how cool you can keep it.

 

[patonb]

My 680i did not like over3.0Ghz without some Vcore.

Most, if not all G0 chips would oc like a champ.

 

[Roark]

Alright, another question:

At what point am I going to do irreparable damage to my chip, and how can I tell.

My E8400 runs stable and within temp specs for normal tasks and whatnot.
However when I IntelBurn it hits 75°C (Tcase is 72.4°C).
I honestly burn it every other day or so since I heard this was the best way to quick-cure AS5.
Just not sure if those extra 2.6°C are doing any major damage to the chip or not.

 

[PP Mguire]

My 680i did not like over3.0Ghz without some Vcore.

Most, if not all G0 chips would oc like a champ.

The 680i was one of the worst boards ever lol.

 

[Ste]

Well, if you don't want to buy a new board you can always Lap the processor, buy a better heatsink and put a tornado or other high CFM fan on the beast.

When I had my q6600 at 3.7 ghz on an abit quad GT, I had two tornados on the heatsink after I lapped the proc with 2500 grit sand paper. One heatsink pushing the air through the tower part and another tornado pointed at the base, where the Q6600 made contact, I also used MX-2 thermal paste. That setup topped out at 72C running prime 95 for 7 hours.

But that setup did not last long, killed two sets of memory and the abit quad GT died after 5 months of running 3.7 Ghz almost non-stop, folding and seti programs. But the Q6600 still lives, just at stock settings.

But everyone else above me is right, its probably just the board. But have you played around with memory timmings and voltages as well? Also make sure all that speed stepping and energy saving features are turned off.