Q6600, optimum overclocked temperature

Peter Tsang · 14 Aug 2011 at 18:06

Folks,

I'm still trying to find the maximum overclocked speed for my Q6600.

So far 3.4GHz seems to be quite stable with 30 minutes fail-safe Prime95 stress test.

I want to know if the following temperature is sort of acceptable Idle (40,39,38,35) and MaxLoad(56,57,54,51), what you reckon?

Thanks!

koooowweeee · 14 Aug 2011 at 18:29

You can go into the 70's with that chip. They say 1.4v is safe volts for that chip but as its old now if it was mine i would give it some welly and go to just under 1.5v if you ever needed to replace it these around50 quid used now.

The chances of breaking it with them volts are slim

benjo · 14 Aug 2011 at 18:30

Those temps are fine

Scougar · 14 Aug 2011 at 23:20

Good temps :-)

What cooler are you using?

Buchanan0204 · 14 Aug 2011 at 23:24

1.5V is more than acceptable. I've been running mine at 1.48V for a few months now and 1.5V for around a year before that. 3.6Ghz stable as a rock.

Ramp it up!

P.S. Temps can go into the high 70's without a hitch!

Peter Tsang · 15 Aug 2011 at 06:38

Folks,

MaxLoad(56,57,54,51) was when I ran some flight simulation like Strike Fighters and Rise of Flights. Prime95 gave me 71C for the first 20 minutes and then fell back to 69 for the remaining of the test which lasted for 45 minutes.

I used CoreTemp to monitor the temperature and it reports a 1.300v constantly. I am actually amazed that under that voltage it keeps stable running 3.4GHz.

I'm just wondering if I ramp up the voltage to let say 1.400v, that will give it a higher speed at 3.6GHz

I'm currently using the H60 from Corsair which is not very high end. The 600T Corsair Chassis provides pretty good air flow and both in take fans are running at 768 rpm. The one of the top of the case ventilate air at 1,780 rpm.

koooowweeee · 15 Aug 2011 at 08:01

You might not need them volts for 3.6 try 1.35 first

use cpu-z to monitor the volts

Peter Tsang · 15 Aug 2011 at 09:17

koooowweeee said:
You might not need them volts for 3.6 try 1.35 first

use cpu-z to monitor the volts

I really don't have the faintest idea as to how many volts will help making 3.6, thanks!
I hope it passes the Prime95 stress test.

koooowweeee · 15 Aug 2011 at 09:35

No one knows. People can guess but each chip is different.

What does cpuz say your volts are

j.col · 15 Aug 2011 at 14:03

i have had my Q6600 upto 80.C and then it cut out, so keep an eye on it
and dont forget more volts = more heat

Peter Tsang · 15 Aug 2011 at 14:28

Folks,

Seeing is better than thousand words.

1.35 won't make it to 3.6, it failed the test.

I up'd it 1.4 and it failed again.

Now it's at 1.45 and it passed the test, as to the temperature here ->

Emphacy · 15 Aug 2011 at 14:43

Temps are good, mine reaches the near 80's overclocked to 3.4GHz.

Voltage is 1.383V.

j.col · 15 Aug 2011 at 18:23

Peter Tsang said:
Folks,

Seeing is better than thousand words.

1.35 won't make it to 3.6, it failed the test.

I up'd it 1.4 and it failed again.

Now it's at 1.45 and it passed the test, as to the temperature here ->

temps are good, so i wouldn't raise it any higher.
Also you need to run prime95 for 12 hrs to be classed as stable and not a few mins

Ben Cole · 15 Aug 2011 at 18:36

j.col said:
Also you need to run prime95 for 12 hrs to be classed as stable and not a few mins

That's a "how long is a piece of string" statement.

Did you just make that number up or did the "official board of stable computer certification" make it law?

pob9 · 15 Aug 2011 at 20:11

Ben Cole said:
That's a "how long is a piece of string" statement.

Did you just make that number up or did the "official board of stable computer certification" make it law?

no he probably took it as a rough estimate as prime is known to run for hours and then fail just when u think its all good. :rolleyes:

Emphacy · 15 Aug 2011 at 20:11

j.col said:
temps are good, so i wouldn't raise it any higher.
Also you need to run prime95 for 12 hrs to be classed as stable and not a few mins

SERIOUS? Okay, maybe a few minutes isn't any good, but if the thing is stable after an hour, it's likely to run fine for another 11!
Unless you're running some kind of rendering workstation, no ones system is going to be running at 100% whack for 12 hours straight, no breaks...

LOL, Run Prime95 for 12 hours, makes me laugh everytime I see it, it's just pointless. Run it for about 45 mins OP, if it survives that, it'll be fine.
Also, give it a run through some demanding games, such as Crysis, I've noticed sometimes it can pass Prime but still come up with a BSOD in game!

koooowweeee · 15 Aug 2011 at 20:37

Ben Cole said:
That's a "how long is a piece of string" statement.

Did you just make that number up or did the "official board of stable computer certification" make it law?

lol

think that was judge dread who said it

as above i only give it a 30 min run and then dame if it crashes up the volts a little

j.col · 15 Aug 2011 at 21:13

Emphacy said:
SERIOUS? Okay, maybe a few minutes isn't any good, but if the thing is stable after an hour, it's likely to run fine for another 11!
Unless you're running some kind of rendering workstation, no ones system is going to be running at 100% whack for 12 hours straight, no breaks...

LOL, Run Prime95 for 12 hours, makes me laugh everytime I see it, it's just pointless. Run it for about 45 mins OP, if it survives that, it'll be fine.
Also, give it a run through some demanding games, such as Crysis, I've noticed sometimes it can pass Prime but still come up with a BSOD in game!

i have had prime95 fail after 10hrs, but as you said its personal preference but i prefer to do an overnight run.

i usually also use Intel burn test, but this is usually runs the cpu hotter than prime95, so i didn't recommend it in this case, as he is the high 70's already.

but like i said, its personal preference,

here is a quote from jokester

"Prime blend is a mix of Large and then Small FFT, you need to run it 8+ hours though to ensure that it's run complete sequences of both, otherwise you may as well have just run Large FFT which it runs for the first few hours." http://forums.overclockers.co.uk/showthread.php?t=18241202&highlight=long+run+prime95

or rjkoneill who i belive does the overclocking at ocuk

"again, another poor assumption.
8 hours prime is required.
either that or 20k on LinX or a good solid 500 runs on IBT
a comuter is not stable because it does a bit of stress.
extended stress is required.
my current cpu will do 4.6 all the time.
it does 4.8 with the memory loosened off and 5ghz with a lot of messing.
but the latter two will run for six hours max before giving up the ghost."
http://forums.overclockers.co.uk/showthread.php?t=18239014&highlight=long+run+prime95

Ben Cole · 15 Aug 2011 at 21:45

j.col, I don't necessarily disagree with your statement. I was just having a bit of a laugh at the "12hrs" as if it was the exact amount.

I'm a sarky tw*t, I think we all knew what you really meant.

setter · 15 Aug 2011 at 21:50

When the quad core database was on the go a while back on here, 8hrs of prime was the requisite for entries to qualify, i guess it just kinda stuck. Thread here.

http://forums.overclockers.co.uk/showthread.php?p=10096083

Some great information in it as well.