Safe Voltage for Core 2 Processors

[MikeR397]

Does anyone have any input on what is a safe voltage for the Core 2 Duo Processors? I've got a E6600 at 3.2ghz at 4.625 vCore and was wondering if that is ok or could i push it a little more. I've heard a lot of places anything under 1.50 is ok with good cooling (I have a Zalman 9500 fan). My goal is for this processor to last about 3-5 years.

 

[Mogadon]

I take it you mean 1.4625V.

It really depends on who you ask, i'd say anything under 1.5V is fine, though of course the lower the better.

 

[Thor86]

What kind of temperatures are you reading at full load?

 

[execine]

3.2ghz@1.46volts ??? mine is runing stable 3.4@1.375

 

[cmdrdredd]

Originally posted by: execine
3.2ghz@1.46volts ??? mine is runing stable 3.4@1.375

every CPU, mobo, & Memory config is unique.

 

[MikeR397]

Sorry, yes i mean 1.4625 Vcore. Also, I cannot do anything less than 1.4625v and achieve 3.2ghz stable either. I am on bios 1305 with Asus P5w deluxe adn have tried other bios' as well.

Execine, do you have the stepping 6 b2 version of the E6600? That is what I have and it seems everyone with those chips cannot get past 3.2-3.3 with them without huge voltages. Also, when you say you are stable, is that Prime stable, pass 3dmark cpu test, or what? I think that I can get 3.4ghz on 1.4625 that has had no problems doign everyday things, but it fails cpu tests. I have not had time to test it on games yet with those settings, but since most things don't use 100% cpu for more than a few seconds, it hasn't given me a problem yet.

 

[Avalon]

I usually use the 10% rule, but I'm more conservative than a lot of the other big overclocker's here. Hence why they usually whip my butt when the next big OC'ing chip comes around

 

[RichUK]

I would say <1.5v. But i would be happy to run 1.55 24/7 temp permitting. Then again i don?t tend to keep my chips too long anyway.

 

[Baked]

Read the box. Max 1.35V.

 

[Some1ne]

That wasn't the question though, the question was "how much voltage can I give it, and still have it last 3 to 5 years". With the number stated on the box, it could probably last some 15 to 20 years.

If you're shooting for 3 to 5 years, I'm in agreement with the general consensus...stay within about 10% of the specced voltage, at most, so <= ~1.48V.

 

[Furen]

I'd stick to stock voltage and less than 3.0GHz for clock speed. Yes, I know this may be a bit conservative but crazy overclocking (and overvolting) is usually not a good way to have high longevity. It's not like you can perceive the difference between 3GHz and, let's say, 3.5GHz unless you are doing things that are timed.

 

[n7]

Originally posted by: execine
3.2ghz@1.46volts ??? mine is runing stable 3.4@1.375

Yeah?

Care to post some screenies of Orthos Blend running for more than a few minutes ?

I'm not saying i don't believe you, but i kinda wanna see those screenies first, since that's damn good for barely over stock vcore.

And to answer the OP's question, i'd say <1.5V providing temps are reasonable.

 

[execine]

yeah... orthos gives me error, but prime95 works, strange... well, anyway, since i tried gaming, encoding etc... and i didn't experience any problems, i'm not gonna bump up the voltage just for this test.

 

[akshayt]

I doubt if anybody can tell with certainity whether your C2D will last 3yrs at anything a little over stock voltage. Remember the mass dieing of 9800Pro cards, who knew?

 

[execine]

so what, you can always return it and get a new one... no one will ever prove that you overclocked...

 

[RichUK]

Originally posted by: execine
so what, you can always return it and get a new one... no one will ever prove that you overclocked...

But that is just unethical. You need to be responsible for your own wrong doings. Its not Intel?s fault if you burn a chip is it, so why should they be responsible.

 

[BonzaiDuck]

Throwin' in my two-cents-worth, here.

First observation: the experience I've had with VCORE on earlier Intel products, suggests that the voltage fluctuates -- even at idle. There is a voltage droop under use.

Second: Intel's "Maximum voltage" spec that appears on the retail box is their estimate of a maximum "safe" voltage that will assure that they will keep their "returned CPUs under warranty" to some known level. This is all about statistics and probability. And money.

Third: If you push the voltage as set in the motherboard to the maximum spec level, it will still fluctuate above and below that level.

From what I gathered based on my own questions on this forum more than a year or two earlier, it is the "load" voltage of greatest concern to someone worrying about CPU longevity with higher VCORE settings.

I just ran a virus-scan on my Prescott -- with motherboard VCORE set at 1.4V -- the Intel "limit." At idle, the Prescott rides up at a a fluctuating voltage averaging around 1.435V. Under load, the average pegs to about 1.375V. How this behavior has changed with the newer CPUs and the Core 2 Duo, I will only find out firsthand when I move my own project from the drawing board, to the workbench, to actual use and tuning.

Here I offer a mix of "likely observation" and speculation. I seemed to observe that PSU voltage rails that were slightly out of spec contribute to a variation between the motherboard VCORE setting and the actual measured VCORE. But that could also just be my imagination. Anyway, for the record, all my voltage rails run within 1% of the spec requirements -- at least now they do . . .

 

[themusgrat]

Originally posted by: Furen
I'd stick to stock voltage and less than 3.0GHz for clock speed. Yes, I know this may be a bit conservative but crazy overclocking (and overvolting) is usually not a good way to have high longevity. It's not like you can perceive the difference between 3GHz and, let's say, 3.5GHz unless you are doing things that are timed.

Well, not really at all. I could tell a huge and definite difference in loading times and BF2 performance by moving my opty 165 from 2.6 to stock 1.8, so yes, these things do make a difference. You have to remember that with the level of software we have today, there is a minimum hardware spec that is pretty high... So small changes above that minimum are very diescernable. Maybe not when you are GPU limited, but in CPU limited apps and situations, because we are working from a minimum, small changes will help.

 

[MikeR397]

Thanks for the insight. Here are a few of my findings....

I can run prime stable at 3.0 ghz with 1.375v.
I can run prime stable at 3.1 ghz with 1.40v.
I can run prime stable at 3.2 ghz with 1.4625v.
I can run prime stable at 3.3 ghz with 1.50v.
I can run prime stable at 3.4 ghz with 1.575v.

I'm not trying anything higher than 1.575v even for testing, so I am assuming anything higher than 3.4ghz is unavailable with my processor.

Also, I've tried everything under the sun as far as other motherboard settings like video card power to heavy, adjusting ram timings, ram voltages, mch/ich/fsb voltages, ect. Vcore is the only thing that effects my overclocks. Which is what it should be until i get to above 400fsb anyway.

These are the minimum voltages I can use to be able to achieve these speeds. I have tried BIOS 801, 903, and 1350 for the P5W Deluxe. On a side note, I have had the best results with the 903 bios and have even been able to use a slightly lower voltage to get those speeds occasionally. It also seems to boot a lot faster (yes Qfan's are disabled).

Anyway, I hope this information is helpful to others trying to overclock the E6600 Stepping 6. If anyone has any suggestions, I'd still like to hear them in case I missed something; but for now, I think i'm goign to stick at 3.1ghz with 1.40v since that appears to be the point where the necessary voltage slope skyrockets. Maybe later in it's life I might push the processor up higher, but it's probalby not worth the risk on a brand new one for that extra .1 ghz.

 

[Geekwannab]

I have been running 350FSB(3.1 GHZ) , Vcore @ auto with my E6600. Orthos stable comming up on 3 hours now. But then I have the week 27 chip which are much better OCers . Check this out . 3.6GHZ(400FSB) with 1.4875 Vcore is stunning with the E6600.

 

[Baked]

I have a week 24 E6400. Vcore voltage for different clock speeds are as follow -

Stock - 1.18V (C1E on) / 1.265V (C1E off)
3200 - 1.275V
3400 - 1.350V

All speeds are Orthos stable for 12 hr plus. Only settings I change in the BIOS is Vcore and Vdimm. Vdimm is always kept at 1.9V or 2.1V based on CPU clock speed.

 

[Furen]

Originally posted by: themusgrat

Originally posted by: Furen
I'd stick to stock voltage and less than 3.0GHz for clock speed. Yes, I know this may be a bit conservative but crazy overclocking (and overvolting) is usually not a good way to have high longevity. It's not like you can perceive the difference between 3GHz and, let's say, 3.5GHz unless you are doing things that are timed.

Well, not really at all. I could tell a huge and definite difference in loading times and BF2 performance by moving my opty 165 from 2.6 to stock 1.8, so yes, these things do make a difference. You have to remember that with the level of software we have today, there is a minimum hardware spec that is pretty high... So small changes above that minimum are very diescernable. Maybe not when you are GPU limited, but in CPU limited apps and situations, because we are working from a minimum, small changes will help.

I didn't suggest sticking to stock clock. Having a 3GHz Core 2 is probably fast enough to shift the onus of most applications onto other things (like video in games, or storage devices when loading). My point was that 3GHz or so would require stock voltage (or a little above stock) while trying to go higher tends to require lots and lots of voltage. Also remember that the CLOCKSPEED also affects the chip's longevity.

TO OP: I'd stick to 3.1GHz. The voltage hike is small enough to shrug off and 3.2 requires quite a bit more voltage (and it gets even worse after that).

 

[Madellga]

Originally posted by: Furen

Originally posted by: themusgrat

Originally posted by: Furen
I'd stick to stock voltage and less than 3.0GHz for clock speed. Yes, I know this may be a bit conservative but crazy overclocking (and overvolting) is usually not a good way to have high longevity. It's not like you can perceive the difference between 3GHz and, let's say, 3.5GHz unless you are doing things that are timed.

Well, not really at all. I could tell a huge and definite difference in loading times and BF2 performance by moving my opty 165 from 2.6 to stock 1.8, so yes, these things do make a difference. You have to remember that with the level of software we have today, there is a minimum hardware spec that is pretty high... So small changes above that minimum are very diescernable. Maybe not when you are GPU limited, but in CPU limited apps and situations, because we are working from a minimum, small changes will help.

I didn't suggest sticking to stock clock. Having a 3GHz Core 2 is probably fast enough to shift the onus of most applications onto other things (like video in games, or storage devices when loading). My point was that 3GHz or so would require stock voltage (or a little above stock) while trying to go higher tends to require lots and lots of voltage. Also remember that the CLOCKSPEED also affects the chip's longevity.

TO OP: I'd stick to 3.1GHz. The voltage hike is small enough to shrug off and 3.2 requires quite a bit more voltage (and it gets even worse after that).

I agree with you. I'm also around 3.2GHz with 1.37V, but to go higher it requires a step increase in VCore.

The benefit/risk factor quickly gets on the no no side.

In my case
1.375V to go from 2.1GHz to 3.2GHz (Orthos Stable, 55 to 59C under load).
1.4125V to go from 3.2GHz to 3.4GHz (Orthos Stable, 57 to 62C under load)
1.45V to go from 3.4GHz to 3.6GHz (not yet stable, 60 to 65 under load).

Core Temp also gets much higher once you pump up the VCore.
Power output increases at a power of 2 factor if you increase VCore.

Stay under 1.40V VCore if you can.

 

[Golgatha]

Originally posted by: MikeR397
Thanks for the insight. Here are a few of my findings....

I can run prime stable at 3.0 ghz with 1.375v.
I can run prime stable at 3.1 ghz with 1.40v.
I can run prime stable at 3.2 ghz with 1.4625v.
I can run prime stable at 3.3 ghz with 1.50v.
I can run prime stable at 3.4 ghz with 1.575v.

I'm not trying anything higher than 1.575v even for testing, so I am assuming anything higher than 3.4ghz is unavailable with my processor.

Also, I've tried everything under the sun as far as other motherboard settings like video card power to heavy, adjusting ram timings, ram voltages, mch/ich/fsb voltages, ect. Vcore is the only thing that effects my overclocks. Which is what it should be until i get to above 400fsb anyway.

These are the minimum voltages I can use to be able to achieve these speeds. I have tried BIOS 801, 903, and 1350 for the P5W Deluxe. On a side note, I have had the best results with the 903 bios and have even been able to use a slightly lower voltage to get those speeds occasionally. It also seems to boot a lot faster (yes Qfan's are disabled).

Anyway, I hope this information is helpful to others trying to overclock the E6600 Stepping 6. If anyone has any suggestions, I'd still like to hear them in case I missed something; but for now, I think i'm goign to stick at 3.1ghz with 1.40v since that appears to be the point where the necessary voltage slope skyrockets. Maybe later in it's life I might push the processor up higher, but it's probalby not worth the risk on a brand new one for that extra .1 ghz.

I stuck at 3.0Ghz with 1.40V myself too. Your values are almost verbatim what I found with my E6600. I just wanted to see what the limits were before I actually started using it for my primary rig. I figure if I can Orthos + Folding@Home 2x instance stabilize it at 3.4Ghz, then it ought to run all day and night long without issues at 3.0Ghz for many years to come. I'm using a Zalman 7700Cu series cooler.

I'm GPU limited at stock (2.4Ghz) anyway in games even with an X1900XTX, and my video editing isn't going to incredibly improve with the extra 0.4Ghz I could squeeze out of the CPU. So I figured why bother stressing the CPU needlessly?

 

[Polish3d]

Originally posted by: Geekwannab
I have been running 350FSB(3.1 GHZ) , Vcore @ auto with my E6600. Orthos stable comming up on 3 hours now. But then I have the week 27 chip which are much better OCers . Check this out . 3.6GHZ(400FSB) with 1.4875 Vcore is stunning with the E6600.

Wow, I must have a damn good chip then. And yes, it's dual prime stable for ~30 min (long enough for me, it didnt fail, I just stopped the test)