The way overlcking reduce the life span of CPU

[hohyss]

I know that high voltage reduces the life span of CPU.

But what if you overclock and can still safely stay on stock voltage?

Will that still reduce the life span of CPU? even if you are on stock voltage?

I once read an article saying that
CPU is sold on market at its most stable speed ( some are slightly underclocked to be OC friendly )

Will any oc activity reduce the life span? or is it negligible up to certain point?

 

[VirtualLarry]

I would say that lifespan is proportional to voltages and temps.

 

[hohyss]

hmm.. So its ok if I maintain the temperature ( with aftermarket cooler and good cae )
and voltage at stock.?

Then the life span wont be affected?

 

[AleleVanuatu]

No one knows about lifespan. Think LOGICALLY. Core 2 Duo has been on the market, what 2.5 years? LoL. Anyone telling you they know about lifespan for sure, is full of shit.

Obviously don't juice your chip with too much voltage. Personally I like clocking up on stock voltage, with the stock cooler, and with good airflow in the case, things are usually about 50C AWAY FROM TMAX.

 

[hohyss]

Originally posted by: AleleVanuatu
No one knows about lifespan. Think LOGICALLY. Core 2 Duo has been on the market, what 2.5 years? LoL. Anyone telling you they know about lifespan for sure, is full of shit.

Obviously don't juice your chip with too much voltage. Personally I like clocking up on stock voltage, with the stock cooler, and with good airflow in the case, things are usually about 50C AWAY FROM TMAX.

I never asked about the specific lifespan of processor. Think LOGICALLY. ;p Read the post.
I am asking what aspect of overclocking would REDUCE the LIFE SPAN of physical CPU itself. You seemed to have confused it with life cycle of CPU.

 

[Idontcare]

Originally posted by: AleleVanuatu
No one knows about lifespan. Think LOGICALLY. Core 2 Duo has been on the market, what 2.5 years? LoL. Anyone telling you they know about lifespan for sure, is full of shit.

There are hundreds, if not thousands, of engineers who know specifically about this topic. It is a rather mature field, as is insurance actuary tables for mortality rates and lifespans from which life-insurance policies are underwritten.

Originally posted by: VirtualLarry
I would say that lifespan is proportional to voltages and temps.

It's actually more proportional to the exponential of electric field strength (which itself is proportional to the voltage).

Double the operating voltage, while keeping temperature the same, and the chip's lifespan can be expected to be reduced by a factor of 10.

This phenomenon is actually what enables accelerated lifespan testing that the chip vendors undertake during their qualification routines. (weibull plots of operating voltage versus lifetime where the voltage is 50x higher than the spec'ed lifetime, etc)

This article isn't the best to cover the topic but it was the best public example I could find with a quick google search.

This article is a little older, NASA paper, but explains things in more of a laypersons terminology I think.

And if you want to delve into the mathematics then this article is not too bad. (not published, just googled and inspected by me to be reasonably accurate as is)

 

[AleleVanuatu]

Theory Theory Theory. Wonderful. Now tell me, WHO has a Core 2 For 5+ years? In Production Use? LoL, Just LoL.

Course, theres tons of data on juicing up IC's that goes back years. Buddy, I'm an old fart, I know this. But Specifics? LoL, Just LoL.

 

[Idontcare]

Originally posted by: AleleVanuatu
Theory Theory Theory. Wonderful. Now tell me, WHO has a Core 2 For 5+ years? In Production Use? LoL, Just LoL.

Course, theres tons of data on juicing up IC's that goes back years. Buddy, I'm an old fart, I know this. But Specifics? LoL, Just LoL.

I have no idea what your being old has to do with us ignoring your ignorance of the existance of an entire field of engineering dealing with accelerated lifetime modeling.

You can't possibly be for real, nor from the industry, nor aware that there are plenty of us here who are.

LOL all you want, I spent >10yrs immersed in lifetime quals for SUN microprocessors at Texas Instruments, until your posts can reflect a semblance of engineering prowess of intelligence you simply won't be and can't be taken seriously.

 

[AleleVanuatu]

If you think that accelerated lifetime modeling is anywhere close to the real world, you're in for a world of hurt. This is something the old guys know: Things are a lot more complex in real life.

 

[Idontcare]

Originally posted by: AleleVanuatu
If you think that accelerated lifetime modeling is anywhere close to the real world, you're in for a world of hurt. This is something the old guys know: Things are a lot more complex in real life.

Care to enlighten the world on how you qualified the device's lifetime as adhering to the spec lifetime without relying on any modeling of accelerated lifetimes to realworld applications? Care to take a guess HOW operating voltage and temperature specifications are determined?

When was the last time (if ever) you were involved in the lifetime qualification of a microprocessor or IC of any kind?

I personally know how this part of the industry gets done, and your viewpoint on it is in direct conflict with reality. I can't state it any more plainly than that.

 

[solog]

Originally posted by: Idontcare
I have no idea what your being old has to do with us ignoring your ignorance of the existance of an entire field of engineering dealing with accelerated lifetime modeling.

Don't you get it, LOL? Were talking specifics. LOL. Fart, lol. Listen to us old guys. When you accelerate into old age, you are going to hurt a lot, in this complex real world. LOL.

 

[Idontcare]

Originally posted by: solog

Originally posted by: Idontcare
I have no idea what your being old has to do with us ignoring your ignorance of the existance of an entire field of engineering dealing with accelerated lifetime modeling.

Don't you get it, LOL? Were talking specifics. LOL. Fart, lol. Listen to us old guys. When you accelerate into old age, you are going to hurt a lot, in this complex real world. LOL.

Thanks for that! I really needed a good laugh, for real. That was priceless.

 

[Zenoth]

As long as you don't go with something like 1.5v on air and stock cooling you won't have to worry about the physical lifespan for a good four to five years if not more, outside of suicide runs (extreme over-clocking, basically asking for problems) I've never heard about a CPU failing after "a few months" or "too early" because of reasonable over-clocks and/or over-voltages and so on, and somewhat high temperatures too, a CPU is much more durable than people might think, as long as you don't screw things up seriously due to complete ignorance or actually doing such a thing on purpose for experimental or benchmark-worthy tests.

I would go as far as saying that when it comes to over-clocking and especially over-voltages one should worry more about the Power Supply / Motherboard / RAM rather the CPU, if "reasonable" conditions are respected, of course. Again I say it, if you play around with voltages and just ignore about OC'ing in general or even about electricity and the risks behind that and electrical components and so on then yes you'll have problems and the CPU might just cease to function after a very short period of time compared to how long it "should have" lasted in "not so risky" conditions.

 

[hohyss]

Originally posted by: AleleVanuatu
Theory Theory Theory. Wonderful. Now tell me, WHO has a Core 2 For 5+ years? In Production Use? LoL, Just LoL.

Course, theres tons of data on juicing up IC's that goes back years. Buddy, I'm an old fart, I know this. But Specifics? LoL, Just LoL.

Again no one asked for specifics. Unless you are having fun talking to yourself.
Good luck with that.

Originally posted by: Idontcare
It's actually more proportional to the exponential of electric field strength (which itself is proportional to the voltage).

Double the operating voltage, while keeping temperature the same, and the chip's lifespan can be expected to be reduced by a factor of 10.

This phenomenon is actually what enables accelerated lifespan testing that the chip vendors undertake during their qualification routines. (weibull plots of operating voltage versus lifetime where the voltage is 50x higher than the spec'ed lifetime, etc)

This article isn't the best to cover the topic but it was the best public example I could find with a quick google search.

This article is a little older, NASA paper, but explains things in more of a laypersons terminology I think.

And if you want to delve into the mathematics then this article is not too bad. (not published, just googled and inspected by me to be reasonably accurate as is)

Thanks for profound material.

Originally posted by: Zenoth
As long as you don't go with something like 1.5v on air and stock cooling you won't have to worry about the physical lifespan for a good four to five years if not more, outside of suicide runs (extreme over-clocking, basically asking for problems) I've never heard about a CPU failing after "a few months" or "too early" because of reasonable over-clocks and/or over-voltages and so on, and somewhat high temperatures too, a CPU is much more durable than people might think, as long as you don't screw things up seriously due to complete ignorance or actually doing such a thing on purpose for experimental or benchmark-worthy tests.

I would go as far as saying that when it comes to over-clocking and especially over-voltages one should worry more about the Power Supply / Motherboard / RAM rather the CPU, if "reasonable" conditions are respected, of course. Again I say it, if you play around with voltages and just ignore about OC'ing in general or even about electricity and the risks behind that and electrical components and so on then yes you'll have problems and the CPU might just cease to function after a very short period of time compared to how long it "should have" lasted in "not so risky" conditions.

The answer I wanted to hear. Thanks so much

 

[MarcVenice]

Hold up though, Zenoth might say 1.5v, but that depends on the chip. I think the 65nm c2d's could take 1.5v tops, if that much allready. I doubt you want to be giving a q9450 or similar CPU 1.5v.

As an example, many oc-ers here have been overclocking for ages, I bet most of them had opterons running oc-ed for at least 2 years. For myself, my CPU is running oc-ed for like 1.5 years now, at it's stock voltage though, because many CPU's can reach nice oc's on stock voltage, so if you're worried, keep it at stock vcore. Also, it made my $100 CPU worth $200. Right now it's worth jack though, like $25 tops. So if it dies, I still come out ahead