Durability of a CPU w/ and w/o OC

[jd254]

I've read so many post and articles about how the life a CPU is shorten due to overclocking. I've had about 7 or 8 processors through out my life from some Pentium 3 to my current quad, and all but the P3 was OC'ed to the best of my ability. Not one CPU has died on me yet. There was even one that I dropped and kicked 6 feet across the floor into a wall.

Am I just lucky, or can CPU's really die under normal or properly OC'ed operating conditions?

 

[veri745]

It's really the increased voltage, rather than the increased clock freq. that kills CPU's, especially if you go beyond the Mfc's recommended voltage ranges.

 

[2March]

Lots of people have lost hardware because of overclocking but it usually user error or overdoing it.

For me it depends on the reason for OC'ing. When you want extra performance extreme overclocking usually is not interresting because of the minute advantages performance wise. But if you really want that extra couple of MHz even when temps go overboard because of extreme voltages you have a good chance of losing it.

 

[jd254]

Originally posted by: 2March
Lots of people have lost hardware because of overclocking but it usually user error or overdoing it.

I agree completely, but has anyone ever had a CPU die on them after prolonged use with safe OC?

 

[jandlecack]

Originally posted by: jd254

Originally posted by: 2March
Lots of people have lost hardware because of overclocking but it usually user error or overdoing it.

safe OC

No such thing. There's your answer. Yes it can happen, yes it has happened, yes as soon as you go past manufacturer's specs you are risking your chip. No matter if that's clock freq, voltage or temps, since they are all correlated anyway.

 

[jd254]

Originally posted by: jandlecack

Originally posted by: jd254

Originally posted by: 2March
Lots of people have lost hardware because of overclocking but it usually user error or overdoing it.

safe OC

No such thing. There's your answer. Yes it can happen, yes it has happened, yes as soon as you go past manufacturer's specs you are risking your chip. No matter if that's clock freq, voltage or temps, since they are all correlated anyway.

lol ok... has it happened to you? right now, you're just spouting what you read, what I read, but I never seen it happen. I don't even hear stories of "my friend's friend's friend's chip fried out of no where, probably from his high OC" Like I said, all my CPU's are overclocked, the oldest and longest currently running chip is a P4 2.8c OC'ed to 3.2ghz. I bought that when it came out, and it's running strong now.

 

[jandlecack]

Of course I have had cases of dead chips, whether OC'ed or not. These things don't last forever and it's a fact that you decrease the lifespan of your chip by overvolting or overclocking it. Or your motherboard. Or your graphics card.

You seriously asking for proof of this?

 

[dguy6789]

The lifespan is decreased, but not by much. Voltage increase is far more harmful to overall chip life. Overclocking is very safe by the way, as long as you know what you're doing and you do it conservatively. A Q9550 overclocked from 2.83Ghz to 3.4Ghz just by changing the FSB from 1333 to 1600 for example will not die any sooner than a QX9770 that was never overclocked. It will die sooner than if it was at stock, but we're talking about a life span of more than 8 years either way so it's a moot point.

 

[jandlecack]

Originally posted by: dguy6789
The lifespan is decreased, but not by much. Voltage increase is far more harmful to overall chip life. Overclocking is very safe by the way, as long as you know what you're doing and you do it conservatively. A Q9550 overclocked from 2.83Ghz to 3.4Ghz just by changing the FSB from 1333 to 1600 for example will not die any sooner than a QX9770 that was never overclocked. It will die sooner than if it was at stock, but we're talking about a life span of more than 8 years either way so it's a moot point.

I never said the real world values are that significant that I wouldn't overclock my chip. But to say an OC is safe or to disregard the fact that any upward modification lowers the lifespan is simply silly.

 

[Dark4ng3l]

The lifespan of a CPU is many times what anyone could ever need anyways. Those things are so durable because they don't have moving parts so it's not like reducing it's expected life by 20 or 30% will have any real effect on anyone.

 

[dguy6789]

Originally posted by: jandlecack

Originally posted by: dguy6789
The lifespan is decreased, but not by much. Voltage increase is far more harmful to overall chip life. Overclocking is very safe by the way, as long as you know what you're doing and you do it conservatively. A Q9550 overclocked from 2.83Ghz to 3.4Ghz just by changing the FSB from 1333 to 1600 for example will not die any sooner than a QX9770 that was never overclocked. It will die sooner than if it was at stock, but we're talking about a life span of more than 8 years either way so it's a moot point.

I never said the real world values are that significant that I wouldn't overclock my chip. But to say an OC is safe or to disregard the fact that any upward modification lowers the lifespan is simply silly.

Safe is relative. Before the overclock, the chip will last 3x as long as I need it to, after the overclock, 2.8x as long as I need it to. Makes no difference to me. If you want to be picky enough, it's not "safe" to run your chip at all since that will make it die sooner.

 

[jandlecack]

Originally posted by: dguy6789

Originally posted by: jandlecack

Originally posted by: dguy6789
The lifespan is decreased, but not by much. Voltage increase is far more harmful to overall chip life. Overclocking is very safe by the way, as long as you know what you're doing and you do it conservatively. A Q9550 overclocked from 2.83Ghz to 3.4Ghz just by changing the FSB from 1333 to 1600 for example will not die any sooner than a QX9770 that was never overclocked. It will die sooner than if it was at stock, but we're talking about a life span of more than 8 years either way so it's a moot point.

I never said the real world values are that significant that I wouldn't overclock my chip. But to say an OC is safe or to disregard the fact that any upward modification lowers the lifespan is simply silly.

Safe is relative. Before the overclock, the chip will last 3x as long as I need it to, after the overclock, 2.8x as long as I need it to. Makes no difference to me. If you want to be picky enough, it's not "safe" to run your chip at all since that will make it die sooner.

Way to talk beside the point there. The subject here was the manufacturer's given specs in conjunction with expected lifespan, which is lowered by an OC no matter what. It could be 0.01%. You still won't be able to state that it's not lowered.

I don't care how long you want your chip to last, and I don't care how many inaccurate values you want to pull out of thin air and throw at me. OCing = decreasing lifespan, no twists and turns involved. Do it and you decrease your lifespan. Stop trying to make me look like I'm advising you against an OC. If you can't accept these bare basics of overclocking I think you should take some of your own advice.

 

[jd254]

Originally posted by: jandlecack

Originally posted by: dguy6789

Originally posted by: jandlecack

Originally posted by: dguy6789
The lifespan is decreased, but not by much. Voltage increase is far more harmful to overall chip life. Overclocking is very safe by the way, as long as you know what you're doing and you do it conservatively. A Q9550 overclocked from 2.83Ghz to 3.4Ghz just by changing the FSB from 1333 to 1600 for example will not die any sooner than a QX9770 that was never overclocked. It will die sooner than if it was at stock, but we're talking about a life span of more than 8 years either way so it's a moot point.

I never said the real world values are that significant that I wouldn't overclock my chip. But to say an OC is safe or to disregard the fact that any upward modification lowers the lifespan is simply silly.

Safe is relative. Before the overclock, the chip will last 3x as long as I need it to, after the overclock, 2.8x as long as I need it to. Makes no difference to me. If you want to be picky enough, it's not "safe" to run your chip at all since that will make it die sooner.

Way to talk beside the point there. The subject here was the manufacturer's given specs in conjunction with expected lifespan, which is lowered by an OC no matter what. It could be 0.01%. You still won't be able to state that it's not lowered.

I don't care how long you want your chip to last, and I don't care how many inaccurate values you want to pull out of thin air and throw at me. OCing = decreasing lifespan, no twists and turns involved. Do it and you decrease your lifespan. Stop trying to make me look like I'm advising you against an OC. If you can't accept these bare basics of overclocking I think you should take some of your own advice.

here's the problem I have with what you are saying. Fact: the life span of all of the same chips are not the same (example, all Q6600 will not have exactly equal life). Can we agree that there is some sort of deviation? If we do, and OC'ing does decrease at a certain percentage, then the decreased life span will only marginally decrease overall average lifespan of all of the same chip. now since we'll agree that it does lower the lifespan, you may think it's "correct" to say that it does indeed "lower the lifespan" of a chip. If you studied stats, you will know that this is a premature answer. If you go into hypothesis testing, the hypothesis of "will OC'ing lower the life of my processor," and you crunch the numbers, it'll say there isn't enough evidence to support the claim of lowered lifespan.

EDIT: also, you made it clear that you don't NOT support overclocking, so... ok.

 

[bryanW1995]

As mentioned, voltage increases do a lot more damage to a cpu than clock increases. I've overclocked everything I've ever owned, too, but I didn't start bumping voltages until june 07. I killed a cheap amd mobo that way, but the cpu still worked fine. Since then I've used better mobos and jacked up the voltages without issue. Hell, I ran an e6750 at 1.64v in bios for a few weeks until my room got too hot... I ended up running that one 24/7 at 1.53 in bios, 1.48 load. I sold it on fs/ft a year later and haven't heard any complaints from the new buyer. I told him the safe overclocking limits and my 24/7 settings, too, mainly because I thought they were slightly poor for an e6750 and wanted him to know.

I did a LOT of research before going so crazy on my e6750, then did more when I got a q6600, q9450, and x3350. The general consensus among knowledgeable oc'ers is that a typical cpu run at stock should last about 10 years. A mild oc with no voltage increase typical causes little to no shortening of life span. A moderate voltage increase, with or without an oc, will hurt a little more (lop off maybe a year or two) but still probably outlast just about every other component on the computer. A very aggressive oc with very good air or water will typically cut the cpu life in half. think q6600 at 3.6 here. A suicidal run with crazy good cooling will only last a few months (aigomorla has a good example or two of this). Poor cooling will cause a decrease in cpu lifespan, especially under high usage/oc scenarios, but not nearly as much as huge voltage increases. I believe that idontcare had a q6700 at 3.73 for quite a while on a vapochill, but he really needed that much performance and he only needed that setup to last until something better came out in a year or two. I don't think that he ever had cpu issues, but hopefully he'll chime in when he sees this.

 

[jandlecack]

I do not support overclocking but run my CPU daily at a 1.2GHZ overclock

If we do, and OC'ing does decrease at a certain percentage, then the decreased life span will only marginally decrease overall average lifespan of all of the same chip.

Denied. Without something to back this up, neither of us know how much a chip's lifespan degrades by how much of an overclock.

now since we'll agree that it does lower the lifespan, you may think it's "correct" to say that it does indeed "lower the lifespan" of a chip.

Yes, that is what you just said...twice.

If you studied stats, you will know that this is a premature answer.

But you just said it does for a fact lower the lifespan. What stats are you talking about? Mind sharing them?

If you go into hypothesis testing, the hypothesis of "will OC'ing lower the life of my processor," and you crunch the numbers, it'll say there isn't enough evidence to support the claim of lowered lifespan.

Denied. Crunch the "numbers" in front of me and I might give a damn what you spout.

If you call yourself an overclocker and can't admit that you're degrading your chip's lifespan, quite frankly, I don't want to continue this discussion and am a little bit disgusted by you. Continue living in a fantasy world for all I care.

 

[jd254]

I already told you what stats... hypothesis testing. and of course we don't have numbers to crunch, but u gave a clear "I don't care if it's 0.01%" and with that number, anyone who knows stats knows that isn't enough to claim that it lowers the lifespan due to it's variation of lifespan. so yes, losing 0.01% of it's lifespan can not be considered decreased life span.

you don't have to reply... but anyone who says "overclocking kills your chip" is merely having faith in another person's word.

as dguy's sig says: Faith = Nonthinking

 

[Gannon]

heat + volts = electric death, it's simple. Also beware of water! I fried a motherboard by overvolting the chipset with a bad heatsink. It's "do at own risk" and try to stay within tolerances by increasing voltage incrementally and then doing those burn in tests to see where the chip will begin to error out.

Ironically enough I've had water condensation spark directly on in the socket of a Core 2 duo and it lived after I dryed off the motherboard, the CPU socket and bottom of the chip had a bit of blackish gunk where the condensation formed and it sparked, I was surprised the board and motherboard were so hardy after my dinking around with extremely cooling experiments.

I benched it and the CPU works fine still, remarkably. Either way things get old, a lot of people simply do not use their hardware long enough (24/7) so they really don't stress their hardware much. So any overclocking and overvolting won't really stress their hardware too much.

I've got an older P4 box that is prone reboots as the video card chips have aged the card will just cut out after it gets too hot and I have to reboot the system.

So yes CHIPS do get old and they do eventually wear out, it totally depends on how much you use it, what the temps are like and whether the heatsink/fan still operates, or is sitll connected well, etc.

 

[waffleironhead]

" i'm in denial"

Yep pretty much. To deny the existence of voltage/heat/time in the degradation and lifespan of a processor is denial.

 

[Idontcare]

It wouldn't hurt some folks in this thread to review the contents of this page and this page regarding Anandtech's position on The Truth About Processor "Degradation". (their words, not mine)

 

[jandlecack]

Originally posted by: waffleironhead
" i'm in denial"

Yep pretty much. To deny the existence of voltage/heat/time in the degradation and lifespan of a processor is denial.

Forget it. Logic, let alone Ohm's law clearly don't exist for certain fellows in this thread. Just let them enjoy their toys.

 

[thilanliyan]

Originally posted by: Idontcare
It wouldn't hurt some folks in this thread to review the contents of this page and this page regarding Anandtech's position on The Truth About Processor "Degradation". (their words, not mine)

So does underclocking and undervolting (say at idle ONLY) help to effectively increase CPU life/decrease degradation? Or does the fact that you're running at regular speeds and voltages at load negate that?

Thanks for the links btw. I didn't know they actually degrade.

 

[Idontcare]

Originally posted by: thilan29
So does underclocking and undervolting (say at idle ONLY) help to effectively increase CPU life/decrease degradation?

Yes it does. Anything you do to decrease the temperature and voltage (which includes clockspeed) will decrease the pace of degradation relative to that of operating the chip at higher temperatures and/or higher voltages (and clockspeeds).

Its a continuous sliding scale. With your materials science background you are likely familiar with the concepts of S-N diagrams and how aluminum versus steel have very different behavior in that steel actually has a fatigue limit at which point it could in theory survive an infinite number of stress cycles without inducing mechanical failure whereas aluminum has no such minimum fatigue limit. Raise the ambient temperature and both the aluminum and steel parts fail at lower stresses.

Try as you might to lower the max stress during the cycle so as to extend the lifetime of the aluminum structure and your efforts will be defeated, the structure will inevitably fail.

This is the same with transistors. The cycling is the clockspeed, the stress is the voltage. Reducing the clockspeed will prolong the life. Reducing temperature will prolong the life. Reducing voltage will prolong the life. But try as you might there is no minimum threshold below which you can operate the transistor indefinitely without inducing degradation. It is like aluminum.

All you can do is try and be gentler and gentler with it (xtors or aluminum) and hope the resultant lifetime exceeds your needs.

 

[Lark888]

First, I am not intending to dispute the fact that higher duty cycles shorten the lifetime of a CPU. Just that the life cycle of CPUs is so incredible short that individual CPU lifetime has not been a factor in the many systems I have bulit. I have been overclocking since my Pentium MMX days and have not had a CPU die before it could be replaced by a $20 cpu that ran faster. In fact, I never did have one die in-service although I have a number of cpus in the drawer gathering dust (perhaps they have now died from neglect). It was hard to give up on my first Celeron 300a. However, I have had a cpu stop running stable at high overclocks (in the 50% range) after a year or two. The two examples I can recall were AMD A64 CPUs. The CPUs ran fine at milder overclock or stock speeds and continued to run without problems until I put them away. My 1996/1997 Pentium 233MMX CPU is still running just fine overclocked at stock voltages (I now use it to run a simple sound system for theatre productions).

I worry more about the capacitors in my old Motherboards than I do about the CPU.

 

[Idontcare]

My QX6700 just recently died on me after a couple years of some pretty solid use.

It would have been nice to have seen a couple more years of action out of it but I feel like I got what I paid for.

CPU death does occur. It's not an urban legend.