Overclocking: voltages and cpu longevity

[Mark Heath]

After a somewhat lengthy and technical talk with b2 (banthracis) I've come here to get all my half-answered question answered properly.

I am already familiar with overclocking in general and have successfully overclocked the cpu in my computer (late at night when everyone else was asleep because they think it'll absolutely kill it if I o'c)

I do however still need to grasp some of the concepts of longevity, heat and voltages.

I originally thought (before banthracis helped me out) that o'cing on stock voltage would increase heat a bit, enough to notice and possibly be looking at, but nothing massive. After b2 told me more about it, in summary he said that the heat increase when o'cing on stock voltage is so negligible that it's not going to be ... well .. big enough to even bother thinking about.

So what sort of change in the longevity of the cpu would I be looking at if I were to find a completely stable overclock on stock voltage (benchmarks run for hours without problems, not noticeable stability issues) compared to both a higher stable overclock on a higher voltage and also compared to the the stock speed and voltage? (by cpu life I don't mean futureproofing, I mean how long will it chug on until it dies)

thanks, especially if you understood that well, since I got lost

btw, I'm talking about both stock cooling and decent budgetish aftermarket cooling like the Cooler Master Hyper 212+

 

[iqvl]

CPUs are healthiest at highest frequency that can be achieved on stock voltage, since they will be well within the specified TDP, temperature and have minimal power loss under this condition.

 

[Mark Heath]

I know there's lots of threads around about this sort of thing and yes I have gone through and read 3 or 4, but my brother's convinced that he knows a lot about computers when he actually doesn't so something about current tech would be needed... cos he says that it's gonna completely screw it and we all know he's wrong.

 

[jsc]

Stay under Intel's (or AMD's) maximum recommended voltage and temperatures and, short of a random failure, your overclocked PC will more likely be obsolete before it fails.

Two summers ago, I discarded a working P233MMX system I had OC'd to 333 MHz. Last summer I dumped a working, OC'd PII Celeron system.

 

[Mark Heath]

PCs round here stay here for.. ages and don't get replaces much. Still got ye old Pentium computers up and in use every few days.

So I'm not asking about anything to do with will it last until it's obsolete, but rather what sort of life difference would there be overall?

 

[wolfram23]

It all depends on how high you go. Since you don't state your CPU, I'll give an example with mine: i5 750. The (new) Intel specs for maximum voltages are 1.40V Vcore and 1.40V Vtt/IMC. Stock volts are something like 1.15V Vcore and 1.08V Vtt or thereabouts. Now, firstly, OCing at stock volts is completely safe. Core speed doesn't really harm the PC it's the voltage that will melt it. If you want to make it go a little faster, let's say you go from 1.15V to 1.25V you should still be totally safe for a long time. However, as you approach the voltage limit you are getting riskier in terms of longevity. Intel states that if you go over their specs, longevity will probably be negatively affected, however, returning to stock settings should something go wrong should allow it to continue to function for a some time.
As for heat, speed doesn't really make heat. The heat comes from the power usage, so more voltage = more power/heat. Again, depends on you CPU but they all have a max running temp from the manufacturer. Intel says for my CPU it's 72C.
I'm running my CPU OC from 2.66ghz to 3.68ghz, voltage from 1.15 to 1.306 Vcore and 1.08 to 1.206 Vtt, and my max heat reaches 62C under stress testing (which it never reaches under normal usage). I fully expect my CPU to last several years like this.

 

[Mark Heath]

hmm... thanks.

Not really asking about either of my cpus (C2D 6750, i5 750 btw), just sort of a general question.

On an o'c like that would you expect it to last say, 10 years though? (seeing as avg stock cpu life is above 10)

 

[wolfram23]

Hmm 10 years? Couldn't say. I'm sure it'll be at least 4 or 5 tho.

 

[Mark Heath]

k.

what about the life difference when o'cing on stock voltage?

 

[wolfram23]

k.

what about the life difference when o'cing on stock voltage?

Probably none. More speed isn't an issue, at worst it causes errors which you fix by either lowering the speed or increasing the voltage. It's the voltage and heat that kill CPUs.

 

[Mark Heath]

Well insignificant amounts of extra heat and current anyway.

Thanks so far guys. As always you guys are great

Hopefully this and all my other evidence'll be enough to convince my thickheaded bro one day that o'cing on stock doesn't make the computer die a quick and painful death.