[Marco Polo]

Specs in profile, pulled this Q6600 out of a new Dell Inspiron Desktop and don't think its oem chip. VID is 1.3250

CPU-Z Current OC

http://valid.canardpc.com/show_oc.php?id=624060

OSDT crashed only minutes in, restarted PC. PRIME works for hours and seems to be stable/fine w/ temps. But with OSDT and a few random general pc usage I have witnessed BSOD and restart, so clearly something is not stable. Need to figure this out.

Desktop w/ CPUZ, CORE TEMP, HMONITOR

Currently I have BIOS settings on AUTO mostly including VCORE, who wants to help push me in the right direction to getting this OC stable and finding the appropriate clock speeds for my CPU.

I'm not sure what volts to start at.

 

[Sickened1]

Could you give us a screen shot of temps underload?

 

[Marco Polo]

Will do... so far I have continued to have BSOD randomly during use... i'll be multitasking blasting music, firefox, image editting, then get a hardware BSOD.

Core Temp has shown max temp around 75-78C @ current OC... not sure what hardware issue im having guess its not temps but volts/auto

in process of loading prime here is ss from test 6 just a few minutes in. Will upload photo of completed prime test if needed.

 

[doat]

seems like you need voltage every time i bsod i press the + for voltage once and then test and repeat until i don't bsod.

 

[gallocs]

Quote:

Originally Posted by Marco Polo Will do... so far I have continued to have BSOD randomly during use... i'll be multitasking blasting music, firefox, image editting, then get a hardware BSOD. Core Temp has shown max temp around 75-78C @ current OC... not sure what hardware issue im having guess its not temps but volts/auto

First of all, the MAX temp you should EVER run a Q6600 at is 72 degrees Celsius. I would STRONGLY suggest you either get a better heatsink/fan or invest in a water cooling setup to continue your OC.

That aside, the MAX voltage you can use on a Q6600 before electron migration begins killing your chip is 1.55V, so your current voltage is safe. I would suggest trying to increase your FSB voltage to ~1.45V or 1.4V, try putting the NB to 1.3V, set your CPU to ~1.4V and test it with LinX, or OCCT.

Also, if you have the GSKILL Light blue RAM sticks that run @ DDR2 1066 5-5-5-15, you need to set their voltage to 2.1V. That could be the issue. The bios may be defaulting your ram to 1.8V. Try all of the aforementioned stuff, but feel free to play with the voltages a bit (up some until stable, then slowly lower other voltages step by step until your still stable.)

 

[That_guy3]

Have you tied 9x multi yet? Im about to revert back to 8

 

[SgtSpike]

Lower your OC by 100mhz, then test again. Also, take your vcore voltage off of auto, and set it manually to 1.44v.

 

[ljason8eg]

Quote:

Originally Posted by gallocs First of all, the MAX temp you should EVER run a Q6600 at is 72 degrees Celsius. I would STRONGLY suggest you either get a better heatsink/fan or invest in a water cooling setup to continue your OC. That aside, the MAX voltage you can use on a Q6600 before electron migration begins killing your chip is 1.55V, so your current voltage is safe. I would suggest trying to increase your FSB voltage to ~1.45V or 1.4V, try putting the NB to 1.3V, set your CPU to ~1.4V and test it with LinX, or OCCT.

Anything over 72C is not a death sentence...not even close. The chip doesn't even start to throttle until it hits 100C. Yeah, running it 24/7 at high temps might hurt the lifespan a bit, but I doubt OP will be stress testing 24/7.

Same with the voltage. You can go over 1.55v and not kill the CPU. 65nm Intel CPUs are more durable than you think.

 

[Marco Polo]

Would my old XP-90C+120MMblue led fan cooler be better then the current Zalman 9700LED? I just got this cooler, and I got it used off ebay lol My old heatsink is about 4 years old and if you look it up the XP-90C is 100% copper.

Ram was already set to 2.1, and PCI-E FRQ to 100 (is 101 better?), gonna restart soon and try the recommendations, thanks

 

[stevenma188]

Quote:

Originally Posted by gallocs First of all, the MAX temp you should EVER run a Q6600 at is 72 degrees Celsius.

However I would like to add that the Max temp for the Q6600 is based on a TJMax of 90*C. Since your program is reporting everything back based on a TJMax of 100*C, temps are reported 10*C higher than actual. You can manually set the TJMax in Coretemp to give you accurate temp readings.

 

[Marco Polo]

Quote:

Originally Posted by gallocs First of all, the MAX temp you should EVER run a Q6600 at is 72 degrees Celsius. I would STRONGLY suggest you either get a better heatsink/fan or invest in a water cooling setup to continue your OC. That aside, the MAX voltage you can use on a Q6600 before electron migration begins killing your chip is 1.55V, so your current voltage is safe. I would suggest trying to increase your FSB voltage to ~1.45V or 1.4V, try putting the NB to 1.3V, set your CPU to ~1.4V and test it with LinX, or OCCT. Also, if you have the GSKILL Light blue RAM sticks that run @ DDR2 1066 5-5-5-15, you need to set their voltage to 2.1V. That could be the issue. The bios may be defaulting your ram to 1.8V. Try all of the aforementioned stuff, but feel free to play with the voltages a bit (up some until stable, then slowly lower other voltages step by step until your still stable.)

And yes I have the same RAM as you, the blue set @ 2.1 (can't go higher can it? Or Timings tighter without sacrifice?)

 

[Marco Polo]

Quote:

Originally Posted by stevenma188 However I would like to add that the Max temp for the Q6600 is based on a TJMax of 90*C. Since your program is reporting everything back based on a TJMax of 100*C, temps are reported 10*C higher than actual. You can manually set the TJMax in Coretemp to give you accurate temp readings.

THANKS for this good bit of info. I just adjusted and its stress testing at 68C max now lol all my temp programs read TCMJ @ 100C, max is really 90C? This is very helpful.

 

[gallocs]

Quote:

Originally Posted by ljason8eg Anything over 72C is not a death sentence...not even close. The chip doesn't even start to throttle until it hits 100C. Yeah, running it 24/7 at high temps might hurt the lifespan a bit, but I doubt OP will be stress testing 24/7. Same with the voltage. You can go over 1.55v and not kill the CPU. 65nm Intel CPUs are more durable than you think.

Intel's website says max temp of 72 Celsius and MAX voltage 1.55V.

I didn't know about the inaccurate Tjmax settings under coretemp though. Nice to know.

 

[!!GoNe^BuCk^W!lD!!]

Quote:

Originally Posted by gallocs The bios may be defaulting your ram to 1.8V.

i know my P5K-E WiFi puts the ram at default 1.8v when my mushkins i had that were 2.25v - 2.35v

in your BIOS around the vcore i believe there's the vdroop thing called Load Line Calibration or LLC, enable it
set your voltages to:
vcore 1.36
vdimm 2.1

if it dont boot or you get BSOD
bump up the vcore 1 notch

i know when i was OCing my x3210(Q6400) vid 1.21 with my P5K, to get it to 500FSB(4.0GHz) i had to max out my voltages except Vcore(1.43v) vdimm(2.3v) and southbridge(1.2v, there was only 2 options 1.2 and 1.5 i believe)
i ended up slowly killing my ram after a year but im pritty sure it was cuz the vdimm at 2.3v

Edit: i just seen you already have the vdimm to 2.1
northbridge is auto 1.3v? bump it up a notch

im about to pull my P5K down off the wall to to play with it....... i know i can get 3.9GHz w/ my fans on low
with my UD3P funky:turd: be happening after 3.7GHz but i can boot into windows at 4.15GHz and bench
but then it acts like all my parts are

 

[ljason8eg]

Quote:

Originally Posted by gallocs Intel's website says max temp of 72 Celsius and MAX voltage 1.55V. I didn't know about the inaccurate Tjmax settings under coretemp though. Nice to know.

Ok so the website says that. Also shows max temp of the i7's at around 67C, yet everyone goes over that, and no one seems to be killing them.

TJmax of 90C makes for some really, really low load temps. Impossibly low IMO. I always used 100C for TJmax. Take a look at this pic. Could it really be running 10C cooler, on air, with a 20C ambient? I don't think so.

 

[SgtSpike]

What you set the TJMax to isn't really important... what IS important, is that you don't bring your temperatures too close to it. Don't go over the TJMax when benching, and don't get within 20c of the TJMax in daily use, and you'll be fine.

 

[Marco Polo]

Is 1.45 VCORE a good number to start at for over 3GHZ oc ?

Just booting up again now with 1.45vc, nb 1.4, ram 2.1, most of rest auto

update
OCCT crashed half way through test

 

[Marco Polo]

I can't get OCCT to be successful it always says an error on a core i've tried auto volts, and 1.45vc, 1.5vc, nb 1.3/1.4

Is OCCT good to use on vista 64 for accuracy? Should I be using Prime first or OCCT first? I prefer prime but i like the OCCT built in protection/error detection, might decrease likelihood of BSOD or crash while running

 

[SgtSpike]

Quote:

Originally Posted by Marco Polo Is 1.45 VCORE a good number to start at for over 3GHZ oc ? Just booting up again now with 1.45vc, nb 1.4, ram 2.1, most of rest auto update OCCT crashed half way through test

Now that you have set the voltage manually, lower your overclock by 100mhz.

I wouldn't increase voltage any more than you already have for stability, so the next best thing is to lower the OC. The revision you have isn't particularly good at overclocking.

 

[Marco Polo]

im trying again, auto volts @ 399fsb x 8 = 3192mhz