*Official Retail G0 Q6600 Overclocking Thread*

[Ol'Baditude]

I was gonna get a Q9550 but they're still too expensive around here...

I'm thinking of hunting for a 1.2000/1.2125/1.2250 VID Q6600 and try to run it at 4Ghz for daily use instead of paying loads for a Q9550...

I'm just wondering if it's worth the effort.. Since i'm running mine at 3.6 without any probs..

From my experience with taking a 1.3250, 1.2625, and 1.2000 to 4.0 GHz, they all scale about the same after ~3.6GHz with respect to voltage. You need a good motherboard more than anything.

What I will say is that I got my new x48 chipset at the same time I got my 1.2000VID so they were not all tested on the same mobo. The 1.2625 and 1.3250VID chips both made it to 3.6 at 1.375-1.4v(actual) on the 680i. So far I need the same with the 1.2000VID on my x48.

I notice this chip takes 1.65v BIOS to run 4.0GHz stable. Although it is less than the 1.2625VID, it's only by .05v. The 1.325VID made it to 4.0GHz, but not at a voltage I felt like sustaining for as long as it took to do some benchmarks. I think it was something like 1.75 or 1.8v. I really don't know if it was stable.

What I would recommend is to save your money and buy a 45nm that will do 4GHz at a lower voltage for 24/7.

So, to answer your question, IMHO I don't think it would be worth it for you. Besides running benchmarks, there is no noticeable performance gain for me with a q6600 from 3.6 to 4.0GHz.

That being said, if you watch the classifieds close enough over at OCF(or anywhere else) you can pick up a 1.2000VID for $160.00 like I did. Hehe.

[jcool]

Holy , what kind of volts are u guys feedin ur quads
Mine never even saw 1,5V and that's under phase.. got an ultra-low-Vid-hot-as-hell CPU though.. it's at 260W with 1,45V already

[ihatbs]

Holy , what kind of volts are u guys feedin ur quads
Mine never even saw 1,5V and that's under phase.. got an ultra-low-Vid-hot-as-hell CPU though.. it's at 260W with 1,45V already

Your are lucky man

Mine needs 1.5v to reach 3.5Ghz

[Ol'Baditude]

Well Jcool, perhaps you can share some of your settings etc. It may well be possible, but not with my Board/BIOS as far as I've been able to tweak it. My system won't run prime for more than a second or two at 3.8GHz with 1.45v, although it will boot up and game. 24Hrs stable needs 1.525v. Haven't even tried Rosetta or F@H. So I can forget about 4.0GHz at that voltage.


I must have 1.65v with this 1.2000VID chip to run 4.0GHz stable overnight. Ambient climbed to 32deg the next day, so I didn't bother going for 24hrs.

[BeastNotro]

He is under Phase Change cooling.

"""""""Well Jcool, perhaps you can share some of your settings, Mine never even saw 1,5V and that's under phase"""""

OR NOW

Case & Cooling: hybrid SS/water cooling in LianLi PC201

[Ol'Baditude]

Does it really matter what Kind of cooling he has? It shouldn't make a difference should it? A chip is going to take a certain amount of voltage to reach a certain Overclock regardless of what's cooling it. The difference is going to be how hot it runs per voltage added.

There is going to come a point, if you go high enough, where the chip would temp out without the extreme cooling to get it there. But, to say that one person can do 4GHz at a lower voltage because he's using phase change as opposed to Liquid or air, I just don't know about that. He certainly may run cooler, but to change thermal and electron transfer effects to the point where a noticably lower voltage could be applied to stabilize the same overclock, I must say I've never experienced that.

No offense intended here guys, but by that rationale, I should be able to get far more per volt out of this chip under my water loop than I was ever able to get out of it with my TRUE. That is just not the case at all.

Maybe I misunderstood your posts JCool and Beast, and you'll have to enlighten me. Does the voltage required go down exponentially with each 10 degrees you go below zero or something?

[BeastNotro]

Cooling does matter as the cooler the CPU is the less voltage is usually needed to hit same GHZ as others with poorer cooling.

Many times chips under chilling and phase change can use less voltage maybe 1.45 as opposed to 1.55-1.60 for some others with similar VID chips.

HEAT is a killer in OCing.

[Ol'Baditude]

Ok, I think I was just confused in the wording.

Honestly though, back to my previous point. I have experienced absolutely no difference in the amount of voltage it takes to OC this Q66, when changing from an air setup to a quality water setup. The only thing I notice is that my 3.6GHz overclock runs alot cooler now. You must have to go pretty damn low into sub zero territory to reap those kinds of reduced voltage rewards.

[BeastNotro]

Cetain chips the lucky ones with same VID some can use less volts then others in same builds and cooling.

Example: I had a Opteron 146 @ 3.2ghz with only 1.46v Air Cooled and a friend bought from same batch and same cooling, same mobo and ram and PSU as me, but he could never do 3.0ghz no matter what voltage he used.

That was just an example a true one at that for similar batches and builds.

You will see many ppl using less volts then other of similar batches and on better cooling.

Or to word it better would be HIGHER ghz at Same Voltage as Some at 1.50v can only get say 4.0ghz where as Colder Cooling = 1.50v = 4.5ghz .. example is all and truthful at that in many cases.

[Khalmyr]

My god!

what kind of chip can do 4.1 with only 1.456v? o_o

That's just insane

[Bail_w]

My Q6600 1.2 vid @ 3825 with 1.45v (in bios)



It can run Prime95 all night with ambient temp ~19C, but during the day it will crash after 15 mins with ambient temp of 25C.

[Khalmyr]

From my experience with taking a 1.3250, 1.2625, and 1.2000 to 4.0 GHz, they all scale about the same after ~3.6GHz with respect to voltage. You need a good motherboard more than anything.

What I will say is that I got my new x48 chipset at the same time I got my 1.2000VID so they were not all tested on the same mobo. The 1.2625 and 1.3250VID chips both made it to 3.6 at 1.375-1.4v(actual) on the 680i. So far I need the same with the 1.2000VID on my x48.

I notice this chip takes 1.65v BIOS to run 4.0GHz stable. Although it is less than the 1.2625VID, it's only by .05v. The 1.325VID made it to 4.0GHz, but not at a voltage I felt like sustaining for as long as it took to do some benchmarks. I think it was something like 1.75 or 1.8v. I really don't know if it was stable.

What I would recommend is to save your money and buy a 45nm that will do 4GHz at a lower voltage for 24/7.

So, to answer your question, IMHO I don't think it would be worth it for you. Besides running benchmarks, there is no noticeable performance gain for me with a q6600 from 3.6 to 4.0GHz.

That being said, if you watch the classifieds close enough over at OCF(or anywhere else) you can pick up a 1.2000VID for $160.00 like I did. Hehe.

Unfortunately I don't live in the US =(

Another reason to change the CPU would be that even though I can OC to 3.6Ghz at 1.4v, at 26ºC ambient the processor still goes from 31 on idle to 70 on full load... that's awfuly hot considering i have a TRUE with a reasonably good fan and my voltages aren't even considered high...

[Ol'Baditude]

Unfortunately I don't live in the US =(

Another reason to change the CPU would be that even though I can OC to 3.6Ghz at 1.4v, at 26ºC ambient the processor still goes from 31 on idle to 70 on full load... that's awfuly hot considering i have a TRUE with a reasonably good fan and my voltages aren't even considered high...

Your Delta CPU Temp from idle to Load is way too high. It should be 15-20degC at the most. Mine's ~12degC. Either you have an obstruction that's not allowing your HS to make full contact, or one either HS base or CPU IHS is warped. My bet would be the TRUE, since they're notorious for having un"TRUE" bases. I know both of mine sure did. A lap job fixed that right up. Another, less likely, culprit may be that there is too much TIM and it's acting as an insulator rather than conductor. But I think you would have to have a bad mount in the first place for that much TIM to remain between the HS and IHS. I don't want to sound patronizing, but that ain't right.
Symptoms: The idle temperature will seem a bit high, but nothing too out of the ordinary. It's when you put a load on it that it becomes really apparent. I first noticed this after trying to mount a DD MC-TDX on my 680i. It was the oddest thing since I thought I had clearance. When I took it off, I noticed I had actually put a mark in the top of one of the caps. That suk'd. I remounted in a different orientation, and everything smartened up. The next time it happened to me, I was mounting a TRUE before selling the machine and it turned out to be a bad Base. The store I buy from must have got a bad batch of these a while back. Anyhow, lapping the HS base worked well.

My Q6600 1.2 vid @ 3825 with 1.45v (in bios)



It can run Prime95 all night with ambient temp ~19C, but during the day it will crash after 15 mins with ambient temp of 25C.

And WTH??? Those Asus boards always seem to do well with the quads.

[jcool]

Ok, I think I was just confused in the wording.

Honestly though, back to my previous point. I have experienced absolutely no difference in the amount of voltage it takes to OC this Q66, when changing from an air setup to a quality water setup. The only thing I notice is that my 3.6GHz overclock runs alot cooler now. You must have to go pretty damn low into sub zero territory to reap those kinds of reduced voltage rewards.

You are right, the lower temperature doesn't do too much for the voltage needs of a CPU at a certain overclock. For example, mine can run 4Ghz at 1,45V on straight water as well.
But there is a temp threshold at which the CPU becomes unstable, that threshold gets lower the higher your OC is (and it's different and unique for each). For example, I need to stay below 60C coretemp if I want to prime 4Ghz or more. Which I can't hold 24/7 on water, and I don't get how you people manage to cool a quad at >1,6V and keep it under 60C. I always had a hard time doing that with even 1,45V, and I did have some badass wc setup.

Again, the reason must lie in the difference of CPUs. Some draw more amperes than others, with the low Vid chips usually being the ones with more leak currents, hence more heat dump than the high vid chips. Meaning that my Q6600 may well be eating the same amount of power at 1,45V as some high vid CPUs at 1,55-1,6V which is why you can run that kind of voltage in the first place.

If I would feed mine 1,6V my phase unit would go down and the PSU'd probably explode

Which is why I won't do it, problem is FSB anyway (stupid mobo won't go higher)

My god!

what kind of chip can do 4.1 with only 1.456v? o_o

That's just insane

Old ones, L8XX sucks

Just remember guys, if you want 24/7 settings the old rule still applies.. max 1,45/1,5/1,55V air/water/phase.
Some might go beyond that, but I wouldn't give more than 1,6V 24/7 no matter what.

[Khalmyr]

Your Delta CPU Temp from idle to Load is way too high. It should be 15-20degC at the most. Mine's ~12degC. Either you have an obstruction that's not allowing your HS to make full contact, or one either HS base or CPU IHS is warped. My bet would be the TRUE, since they're notorious for having un"TRUE" bases. I know both of mine sure did. A lap job fixed that right up. Another, less likely, culprit may be that there is too much TIM and it's acting as an insulator rather than conductor. But I think you would have to have a bad mount in the first place for that much TIM to remain between the HS and IHS. I don't want to sound patronizing, but that ain't right.
Symptoms: The idle temperature will seem a bit high, but nothing too out of the ordinary. It's when you put a load on it that it becomes really apparent. I first noticed this after trying to mount a DD MC-TDX on my 680i. It was the oddest thing since I thought I had clearance. When I took it off, I noticed I had actually put a mark in the top of one of the caps. That suk'd. I remounted in a different orientation, and everything smartened up. The next time it happened to me, I was mounting a TRUE before selling the machine and it turned out to be a bad Base. The store I buy from must have got a bad batch of these a while back. Anyhow, lapping the HS base worked well.




And WTH??? Those Asus boards always seem to do well with the quads.

I definitely need to learn how to lap that monster... It really might be what you're saying.. Temps are way too high =(

If you noticed the weather app in the sidebar, outside my bedroom the temp was 21ºC.. It's ridiculous to think that my cpu should be 70ºC while only being fed 1.4v

[Bail_w]

You are right, the lower temperature doesn't do too much for the voltage needs of a CPU at a certain overclock. For example, mine can run 4Ghz at 1,45V on straight water as well.
But there is a temp threshold at which the CPU becomes unstable, that threshold gets lower the higher your OC is (and it's different and unique for each). For example, I need to stay below 60C coretemp if I want to prime 4Ghz or more. Which I can't hold 24/7 on water, and I don't get how you people manage to cool a quad at >1,6V and keep it under 60C. I always had a hard time doing that with even 1,45V, and I did have some badass wc setup.

Exactly, same thing with my q6600 (1.2 vid). I must keep it under ~60C loaded with real temp or otherwise it will crash. Ex) i can run OCCT and prime95 all night long but when i try the IntelBurnTest, it will crash my system after 2 minutes because temp were at 65s loaded.

[Hans Gruber]

[QUOTE=Ol'Baditude;3266136]Your Delta CPU Temp from idle to Load is way too high. It should be 15-20degC at the most. Mine's ~12degC. Either you have an obstruction that's not allowing your HS to make full contact, or one either HS base or CPU IHS is warped. My bet would be the TRUE, since they're notorious for having un"TRUE" bases. I know both of mine sure did. A lap job fixed that right up. Another, less likely, culprit may be that there is too much TIM and it's acting as an insulator rather than conductor. But I think you would have to have a bad mount in the first place for that much TIM to remain between the HS and IHS. I don't want to sound patronizing, but that ain't right.
Symptoms: The idle temperature will seem a bit high, but nothing too out of the ordinary. It's when you put a load on it that it becomes really apparent. I first noticed this after trying to mount a DD MC-TDX on my 680i. It was the oddest thing since I thought I had clearance. When I took it off, I noticed I had actually put a mark in the top of one of the caps. That suk'd. I remounted in a different orientation, and everything smartened up. The next time it happened to me, I was mounting a TRUE before selling the machine and it turned out to be a bad Base. The store I buy from must have got a bad batch of these a while back. Anyhow, lapping the HS base worked well.

Far too many people believe the TRUE is better than the TRU aka Thermalright Ultra 120. In theory the TRUE should be better with the addition of two additional heatpipes (6) over the older TRU with four. However, Frostytech.com has said for a long time that TRUE fails to maintain full contact with the CPU because of the bowing that occurrs with the mounting of two additional heatpipes. Still the TRUE performs quite well even with on 25% contact on the CPU's in many cases. Here are the cooling numbers from Frosty Tech. http://www.frostytech.com/articlevie...id=2303&page=5

I have two Thermalright Ultra 120's myself. One runs on my Q6600 and the wall is 3.6Ghz on this chip. You need brute force (very high voltage 1.5v++) to climb well above 3.6Ghz. I certainly am impressed by some of the overclocks many of you have achieved.

My only question is wheather there will be a new 45nm CPU Quad introduced on the 775 chipset before the Nehalem chips and new chipsets appear. I'm hopeful Intel has one final 45nm Quad core yet to be released that can achieve 4.0Ghz on modest voltage 1.45v or lower.

[Ruslan]

Finally stable at 3.6 on air


How I keep it cool

[highspeed]

You are right, the lower temperature doesn't do too much for the voltage needs of a CPU at a certain overclock. For example, mine can run 4Ghz at 1,45V on straight water as well.
But there is a temp threshold at which the CPU becomes unstable, that threshold gets lower the higher your OC is (and it's different and unique for each). For example, I need to stay below 60C coretemp if I want to prime 4Ghz or more. Which I can't hold 24/7 on water, and I don't get how you people manage to cool a quad at >1,6V and keep it under 60C. I always had a hard time doing that with even 1,45V, and I did have some badass wc setup.

Again, the reason must lie in the difference of CPUs. Some draw more amperes than others, with the low Vid chips usually being the ones with more leak currents, hence more heat dump than the high vid chips. Meaning that my Q6600 may well be eating the same amount of power at 1,45V as some high vid CPUs at 1,55-1,6V which is why you can run that kind of voltage in the first place.

If I would feed mine 1,6V my phase unit would go down and the PSU'd probably explode

Which is why I won't do it, problem is FSB anyway (stupid mobo won't go higher)



Old ones, L8XX sucks

Just remember guys, if you want 24/7 settings the old rule still applies.. max 1,45/1,5/1,55V air/water/phase.
Some might go beyond that, but I wouldn't give more than 1,6V 24/7 no matter what.

Some nice information there, thanks. Killer q6600 you have btw

[BeastNotro]

Ok i stated before the cooler the temps the less voltage maybe needed and in some cases it is true if your really chilling it. It does depend on chip from chip for volts, but if you can freeze a cpu you should be bale to use less volts then if not freezing it or chilling it.

I have read it al over the next about such a topic and it has been debated many times and has been proven in instances more then just say 1 that really cold temps can allow for a lessor voltage needed then on AIR cooling.

The only way to know for 100% is LNC your CPU to find out....

[Ol'Baditude]

Exactly, same thing with my q6600 (1.2 vid). I must keep it under ~60C loaded with real temp or otherwise it will crash. Ex) i can run OCCT and prime95 all night long but when i try the IntelBurnTest, it will crash my system after 2 minutes because temp were at 65s loaded.

I'd have to agree with this. Can't prove beyond a reasonable doubt it was the temperature, but last time I was at 4GHz(which is the only time my CPU has ever been anywhere near 60Deg) I did have alot of BSOD's. Whereas, with only slightly less voltage, 3.8GHz was stable as anything at mid 50deg core temperatures on RealTemp. Mine is also a 1.2000VID.

[drizzt5]

IDK why! But I can get my q6600 stable at 3.9ghz perfectly fine with under 1.5v's. The moment I go over to 4ghz though... it takes so much voltage...

And when I mean stable, I don't mean intelburntest stable :P I'm loving this program its crazy.

[Alcibiades]

This thread is looooong, very informative and interesting.

Here's my experience
I have a true q6600 vid 1.275
I can prime95 9x400 @ 3,600 all day,
but 8x450 @ 3,600 will fail on 1 core but keep going and going on the other 3
so I say, well what does that really mean?
and I guess, well we haven't crashed,
it's just for 4 cores to hit 100% full load I need more vcore
but with a vid of 1.275 it's more vcore than i can cool, a bad vcore/temp ratio
450 FSB gives me much better DDR2 bandwidth than 400
so i choose 458 fsb as my 24/7 'cause i don't bsod or crash
whatever works i guess
some have even said they uninstalled prime after 1 day
'cause everything else was stable and would pass, only prime would fail
well some, most actually are now saying they've uninstalled or disregard their failed IntelBurnTest runs

haha well what's now different from me and others disregarding our failed prime95 runs?
I can happily say I have been running like this for at least 10 months now and I'm glad I chose to disregard the unrealistic and over the top demands placed on the cpu by prime95.

[drizzt5]

I rather have stability for sure then a few mhz faster. You want over the top? IBT is over the top compared to OCCT and prime 95.

[Alcibiades]

but what is stability or what does it mean to you?
to me stability is running all my apps and games without any ANY problems to report EVER, thank you.

IBT and Prime95 are over the top imho