Is it possible to give a CPU too much voltage

[Crescent13]

??

 

[Unkno]

Yes, voltage can also kill the CPU but not instantly if your getting good temps.

 

[Somniferum]

Yes. Over time, excessive voltage will destroy an integrated circuit, regardless of heat.

How much voltage is excessive? How much time will it take?

Google "electromigration" if you really want to know....

 

[Crescent13]

Would an Opteron 144 @ 1.525V be too much?

 

[Avalon]

Originally posted by: Crescent13
Would an Opteron 144 @ 1.525V be too much?

Should be fine so long as the temps are good. If you upgrade CPUs often, I'd worry even less.

 

[Crescent13]

Temps are 35C max idle, 40C max load. (I say max as in that is as high as it has ever gotten during idle/load)

 

[diapickle]

You are probably good with those temps.

 

[A554SS1N]

Originally posted by: Crescent13
Would an Opteron 144 @ 1.525V be too much?

That's fine really.

 

[Crescent13]

Well I tried to lower the voltage, and I got it to run safely overnight @ 2808MHz with ~1.485-1.500V. Temps dropped by 4C. I think this is the highest overclock I can get, I tried raising the voltage to like 1.575, and I couldn't get an extra MHZ on the FSB (with 9x multi, thats from 2808 to 2835 mhz). I'm pretty happy with this OC now.

 

[Somniferum]

The point to remember is that even with low temps, the higher you push the voltage, the higher the likelihood that your chip will fail early due to electromigration. If you upgrade your CPU fairly often (yearly, say) then who cares. But if you are hoping to use the chip for several years, it's worth factoring that into your decision of what voltage to use for everyday operation. In other words, if you have to raise the voltage dramatically to squeeze out 30 extra MHz, it's probably not worth the risk, all other things being equal.

 

[VStrom]

What I follow (temps not withstanding):
0-10% increase over stock voltage is prety safe
11-15% is moderately safe
16-20% is borderline...long term life (3+ years) is questionable
21%+ will most likely cause damage sooner or later (the higher the sooner).

 

[Unkno]

one question, would increasing the voltage by like 10% still be able to live for 5 years or so with good temps (dual core)

 

[Leper Messiah]

Originally posted by: Unkno
one question, would increasing the voltage by like 10% still be able to live for 5 years or so with good temps (dual core)

yes. The engineers have to build some kind of tolerance into the cores, due to the fact that not all voltage regulators on mobo's are created equal. 10% over shouldn't adversely affect the life of the CPU by any ammount, after that, it gets kinda iffy. so with a dualie, your default vcore is 1.35, so you can go up to like 1.475 probably without killing the core, IMO.


Newer A64s don't repsond very well to uber ammounts of volts anyways, the gains after about .2 over are very small.

 

[tvdang7]

um..........since we ar on this topic....... my max on my 3200 venice 1.575volts running 2650mhz cooled by a alcu zalman7000b is 48 during super PI and 52 on prime. is this bad even tho its rated for 65 max?

edit : iono if the software is reading right cuz i fell barely any heat on the heatink when i touch it....i feel nada

 

[SickBeast]

Just an FYI, I killed an Athlon 1400mhz by volt modding it. It was 60C at the time of death, and it only took about a minute to die. I smelled burning afterward, poor thing. I was using the maximum voltage of my MSI motherboard, combined with a variable resistor to futher juice things up. Unfortunately I couldn't resist lowering the resistance more and more.

 

[Leper Messiah]

Originally posted by: SickBeast
Just an FYI, I killed an Athlon 1400mhz by volt modding it. It was 60C at the time of death, and it only took about a minute to die. I smelled burning afterward, poor thing. I was using the maximum voltage of my MSI motherboard, combined with a variable resistor to futher juice things up. Unfortunately I couldn't resist lowering the resistance more and more.

jeez. You must have hit like 2.5vcore+ if you killed a T-bird. Those things were beasts.

 

[aka1nas]

In general, if you look at the highest rated model that is the same core of your CPU, then you can at least use that models voltage without frying the chip. Provided you can cool it of course.

 

[The Sauce]

I have never really had to max out volts to get max o/c. I have found that with every processor there is a sweet spot with regard to Vcore and MHz. One I go past that sweet spot I can dramatically increase voltage and get virtually no significant speed gains.

For instance right now with my Opty 165 I can do 2.55GHz (285x9) @ 1.4V but even taking the voltage up to 1.55 I cant get stable at 2.6GHz.

This has been the case for most procs I have owned so there has never really been a good reason to crank the volts....just find your sweet spot and then run it at as low voltage there as you can.

 

[SickBeast]

Originally posted by: Leper Messiah

Originally posted by: SickBeast
Just an FYI, I killed an Athlon 1400mhz by volt modding it. It was 60C at the time of death, and it only took about a minute to die. I smelled burning afterward, poor thing. I was using the maximum voltage of my MSI motherboard, combined with a variable resistor to futher juice things up. Unfortunately I couldn't resist lowering the resistance more and more.

jeez. You must have hit like 2.5vcore+ if you killed a T-bird. Those things were beasts.

Something like that.

I hope you have nothing against beasts.