Upgrading on a budget??

Ok cool . . . obviously one hour of Prime95 "blend" isn't total proof that the vCore is enough to keep that Intel® Core™2 Quad Q6600 100% stable @ 3.0GHz but its good enough for the moment! :p

Theres two things to do which I think will drops the temps a bit more, the first is obviously dial down the vCore a bit more . . . if you go too low though the system will fail to boot windows or indeed reboot the system as soon as you load Prime95 . . .

The second thing is to go back into BIOS to this page again and change

CPU PLL Voltage from [Auto] to [Minimum/1.500v?]

If the board is giving the chip too much CPU PLL under [Auto] setting that will indeed be pushing the load temps up so its worth a try . . . if [1.500v] is too low the same thing will happen as when you UnderVolt the vCore too much (Failed load of Windows/Reboot Under Prime95 stress) . . .

In your shoes right now I would first whack CPU PLL down and run a quick test just to see if it lowers the temps and if indeed the system remains stable . . . If the temps don't drop further from that then I guess you may as well put CPU PLL back on [Auto] and see how much further you can reduce the vCore! :cool:
 
PLL set to 1.5v, boots and runs prime but so far temps still at 67/70

Thanks so far, I think its best to leave prime running over night.
 
Last edited:
About the SpeedStep, I think you are referring to me. Sorry for not being clearer as I was just trying to explain what SpeedStep is in simple term. What I meant was most people will disable Speedstep when they are 'overclocking', as it would be one less variable to concern with when testing the voltage and stability during overclocking.

After they are done overclocking and found their stable speed and voltage, they can put SpeedStep back on. But sometimes when people push the vcore voltage to the lowest possible value for their stability clock, SpeedStep could cause crashes/BSOD because cutting too close (may be the split-second of not enough vcore voltage), and if that happens, give the vcore voltage the tiniest bump should solve the problem.
 
Hello Marine-RX179 :)

About the SpeedStep, I think you are referring to me. Sorry for not being clearer as I was just trying to explain what SpeedStep is in simple term.
I'm not sure any explanation was needed here actually . . . it seems there are a lot of "cookie-cutter" LGA775 OverClockers that always say the same thing when they go to help someone overclocking . . . "Disable speedstep, disable CIE" . . . Its like they are reading from a script! :confused:

sometimes when people push the vcore voltage to the lowest possible value for their stability clock, SpeedStep could cause crashes/BSOD because cutting too close (may be the split-second of not enough vcore voltage), and if that happens, give the vcore voltage the tiniest bump should solve the problem.
Is this based on actual experience or is this something you read in an overclock guide somewhere?

I ask because I don't believe this is true? . . . what facts is this based on please? :D
 
Is this based on actual experience or is this something you read in an overclock guide somewhere?

I ask because I don't believe this is true? . . . what facts is this based on please? :D
I actually had the experience myself. I had overclocked my Q6600 at a lower vcore voltage in the pass and prime stabled for 12 hours. No problem after a few days of use, and I put SpeedStep back on. After that, I get random BSOD, and by disabling SpeedStep no more BSOD.

However, I don't think the SpeedStep is the sole contributeor to the cause of the instability...I suspect it is more like SpeedStep combine with Vdroop of the motherboard that's what cause the BSOD. I'm not saying all socket 775 motherboard will have Vdroop problem, but for those that does, together with SpeedStep plus bare minimum vcore voltage could results in instability.
 
PLL set to 1.5v, boots and runs prime but so far temps still at 67/70
Ok so adjusting the CPU PLL Voltage voltage down didn't do anything to lower the temps? . . . stick it back on [Auto] then . . .

Thanks so far, I think its best to leave prime running over night.
Ni! :p

Drop the vCore again from [1.300] to [1.2750] ;) . . . the long prime test can wait! :cool:
 
I actually had the experience myself. I had overclocked my Q6600 at a lower vcore voltage in the pass and prime stabled for 12 hours. No problem after a few days of use, and I put SpeedStep back on. After that, I get random BSOD, and by disabling SpeedStep no more BSOD.
Can you remember the "exact" clock setting please? . . . What CPU MHz? . . . what FSB was you running etc?

When did the crash happen? . . . during Stress or during Idle?
 
Ok so adjusting the CPU PLL Voltage voltage down didn't do anything to lower the temps? . . . stick it back on [Auto] then . . .


Ni! :p

Drop the vCore again from [1.300] to [1.2750] ;) . . . the long prime test can wait! :cool:

Thanks Big.Wayne

It ran prime all night with a max temp of 73, I will drop the vCore again this evening when I am back in.
 
Ok so adjusting the CPU PLL Voltage voltage down didn't do anything to lower the temps? . . . stick it back on [Auto] then . . .


Ni! :p

Drop the vCore again from [1.300] to [1.2750] ;) . . . the long prime test can wait! :cool:

Evening vCore dropped to 1.2750 and PLL back on auto, boots and running prime. Still 70degs under load I think a new cooler is in order.
 
Last edited:
Sensible, your losing nothing by purchasing some better cooling, you can use it when you upgrade in the future, nothing wasted. As for performance, what are you using to determine any improvements?
 
Sensible, your losing nothing by purchasing some better cooling, you can use it when you upgrade in the future, nothing wasted. As for performance, what are you using to determine any improvements?

Just my usual work process, 2D, 3D CAD and Rendering, oh and some heavy photoshop work.
 
Last edited:
Back
Top Bottom