Q6600 to 3.4GHz help

willy

willy

willy

Associate
Joined
27 Nov 2008
Posts
1,107
Hi all,
At the moment running 3.3GHz (P5Q Deluxe and Corsair XMS 4gb 6400) with BIOS setting;

FSB 366
PCIE 100
FSB Strap 333
Timing 5-5-5-18
vCore 1.4 V
Dram Voltage 1.80 V
NB Voltage 1.26
FSB Term 1.3 V
CPU PLL 1.5

Anything else should i do if i want achieve 3.4GHz as i got BSOD when running Prime 95 blend for nearly 1 hour with FSB set to 377 ?
Temp at the moment 51 C the highest with Prime 95

Thanks a lot
 
Last edited:
What speed is the memory running at on the 333 strap?

A slight bump on the NB, FSB Term or PLL may be all it needs. Otherwise throw some more vcore at it.
 
Yeah just go through the 4 voltages one at a time to see which one or which combination makes a difference.

It's all about trial and error. Oh and a shed-load of patience.
 
Ok, will try it at the weekend :)
Will it be the same if i put the FSB up to 400 as well, just play with the voltage and watch my temperature :D
 
Yeah it's just the same, tho personally I tend to jump just 100mhz on the CPU speed at a time once I get over 3.3ghz on a quad-core. You'd think that it might take longer to find your max clock, but it doesn't really as you're making less adjustments each time, and thus there's less testing. Plus you will have stable settings for that speed.
 
Right, will try to set my FSB to 380 at the weekend and up vCore first a bit and see if it k OCCT and Prime 95 after that.

What voltage do we usually put up, vCore, NB, FSB and PLL?
 
Could be any or a combination.

Usually I would stick to the vcore until it stops having an effect at that speed, and then try the other voltages a notch or 2 at a time eg - if upping the vcore still doesn't get Prime95 past 1 hour, then try the NB voltage. If that has an effect then happy days. If not then try FSB Term etc. Often you'll find that you may need to up more than one of them to get to the next level eg - upping the NB only gets Prime95 to three hours, but also upping FSB Term gets you further. But don't up the voltages until you need to. Once you get stable at that speed, try the next level. Again it could be any of or a combination of the 4 voltages that will be needed to be raised again. Once you get over 3.4ghz you'll probably find that you need to increase the voltages in bigger steps, as opposed to the single notch of the lower clock speeds.

As I say, it's all trial and error.
 
Last edited:
Cob said:
Could be any or a combination.

Usually I would stick to the vcore until it stops having an effect, and then try the other voltages a notch or 2 at a time eg - if upping the vcore still doesn't get Prime95 past 1 hour, then try the NB voltage. If that has an effect then happy days. If not then try FSB Term etc. Often you'll find that you may need to up more than one of them to get to the next level eg - upping the NB only gets Prime95 to three hours, but also upping FSB Term gets you further. But don't up the voltages until you need to.

As I say, it's all trial and error.
Click to expand...
Found this myself, albeit more error than trial, 3.6ghz stable was easy enough to achieve with +0.02v on the chip voltages and 1.4 vcore. 3.8ghz on the other hand, required a lot of messing about with the chip voltages.
 
So i just try up vCore and test it, if it doesn't past 1 hour, up NB and try again, if it doesn't past 1 hour up the vCore again?
I'll try it :D

Thanks mate
 
setter said:
Found this myself, albeit more error than trial, 3.6ghz stable was easy enough to achieve with +0.02v on the chip voltages and 1.4 vcore. 3.8ghz on the other hand, required a lot of messing about with the chip voltages.
Click to expand...

Is the chip voltage not the same as vCore :confused:
 
willy said:
So i just try up vCore and test it, if it doesn't past 1 hour, up NB and try again, if it doesn't past 1 hour up the vCore again?
I'll try it :D

Thanks mate
Click to expand...

Not quite. Once the vcore stops having an effect then concentrate on the chipset voltages. Don't up the vcore again til you've tested the other voltages and combinations of them.
 
So how do i know that the vCore stops having an effect? BSOD with Prime 95 Blend?
Also some people saying that use OCCT to test how much voltage i need? Can i do that?

Thanks
 
I use OCCT for quick tests as passing an hour of it usually indicates that you'll pass a few hours of Prime95 (tho it doesn't seem to be quite as good an indicator for Prime95 Blend).

Vcore initially stops having an effect when upping it alone no longer lets you boot Windows and pass the stress tests. At that stage you have to start considering the chipset voltages as well. High overclocks are a combination of enough vcore, enough NB voltage, enough FSB Term, enough PLL and correct memory settings. Mix that with a lot of luck and you'll eventually find your chips limit, whether that be the chips max speed, the max speed it will go in your motherboard, or when temps start to get a bit silly.
 
Oh... so when i put it the other day with FSB 377 and vCore 1.4 V it went fine to Vista and then got BSOD when i did Prime 95 Blend for nearly an hour, so that might not be vCore so might be chipset voltage that i need to increase then? Am i right?

So when i increase NB, FSB and PLL, i just try it 1 by 1 and bit by bit until i get it stable?
 
Yip. Try 0.025v steps.

But always try upping the vcore first. It sometimes needs a decent boost to get to the next level once you get to 3.3-3.4ghz (often 0.05v at a time).
 
if you want to speed up the time you spend testing your system and really want to find out if your overclocks are truly stable then grab yourself a copy on intel burntest. dont waste your time with the likes of price and orthos, you can run those for days and that wont mean crap compared to a 1 hour run of intel burntest.

get the details here: http://www.xtremesystems.org/forums/showthread.php?t=197835
 
You must log in or register to reply here.
Back
Top Bottom