What kills CPUs in the longrun? Voltage or Temps?

[RA3KW0N]

Some of us can have low temps by using different cooling methods, but have to pump more juice to obtain a high overclock because of CPU variations (e.g. high VIDs)..

others have good chips, and similar temps, because not as much voltage is being put into the CPU to obtain the same overclock..

my question is... in the long run, what is it that kills CPUs?

Excessive Voltage? Temps? Both?

I am aware that more voltage tends to mean higher temps... but let's say we all had access to liquid nitrogen cooling, would high voltages kill CPUs?

[enaberif]

both.

but voltage is the worst and will degrade a cpu fairly quickly.. and by quickly I mean it'll have 10 years instead of 15.

[Tazer-[X]]

Wow, in ten years, I dont think ill be able to run a next gen, or 3 gens from now OS on my q6600. Considering, that the next gen CPU from Intel, is supposed to have up to 16 cores per die on the server chips.

[Jon_di2]

Originally Posted by RA3KW0N

Some of us can have low temps by using different cooling methods, but have to pump more juice to obtain a high overclock because of CPU variations (e.g. high VIDs)..

others have good chips, and similar temps, because not as much voltage is being put into the CPU to obtain the same overclock..

my question is... in the long run, what is it that kills CPUs?

Excessive Voltage? Temps? Both?

I am aware that more voltage tends to mean higher temps... but let's say we all had access to liquid nitrogen cooling, would high voltages kill CPUs?

Temps aren't as bad as most people think... but what do you consider excessive? Anything up to 70degrees shouldn't take too much life off your cpu. Over that, yes I would say effects would start to be worse as you rise, but your cpu will shut off automatically before the temps get so high that immediate damage is immanent.

Voltage on the other hand is up to you how much goes in. And anything beyond the voltage recommended by intel for instance is going to reduce the life of the cpu. But really when you think about it, like enaberif said you will reduce it from like 15-10 years, and will you be using this comp in 10 years anyways? Probably not.

Ps. The people who use liquid nitrogen to compete in overclocking contests do not OC on liquid nitrogen on a constant basis. These voltages would reduce the life of the processor dramatically if done often.

[Boldeagle]

That's a fair question and you are correct in stating that higher volts will lead to higher temps. Generally people will increase voltage and use the temps as a guide for upper thresholds, i.e. they may use vCore 1.8 to get to temps off 70C load and use this as an upper threshold with the temp being the dictator. Under LN2 you can run the CPU at 1.8-1.9 or higher and if you could maintain those low temps for a long time (have a constant supply of LN2) you may keep these high voltages but you wouldn't be able to run the CPU at say 80C for prolonged periods. I suposse you would really need to conduct 4 experiments running 4 equally matched systems one with high volatges and low temps, one with low volts and high temps, one with high volts and high temps and finally one at stock (consider this low volts and temps) and then try and quantify "product" life cycles. To really provide basis to such an experiment you would have to do 3 replications of all sample types and generate a mean (average) of each sample type, therefore 12 experimental units, got a lazy 10 grand?

[MpG]

It's not really an either-or kind of situation. More of an inter-relationship between the voltage, current, and temperature levels occurring within the CPU. The voltage is what pushes the current through the chip, but the current is effectively what creates the heat and wear, but the temperature of the CPU affects how much current actually flows for a given voltage.

Which is why using temperature as a gauge for how much strain you're putting on a CPU isn't entirely without merit. It's more complicated than that, and the exact math behind it all is probably best left to the engineers, but generally speaking, a colder chip can handle more input voltage without necessarily incurring more wear in the process.

[DK2]

Increasing voltage is what will kill your cpu more than temps.

Cpu’s are spec’d to operate within a certain temperature range and frequency.
Generally keeping temps within and smaller range allows a better OC.

[sswilson]

One of the things that hasn't been mentioned (and I can't for the life of me remember the actual name for it.... :( ) is that over time higher voltages will cause the electrons from one internal trace to jump to another internal trace eventually creating an electrical bridge between two points which aren't intended to be connected.

If you think of a spark plug with the connector removed, as you bring the contact closer to the spark plug you get to a point where you'll get a spark. Increasing the voltage will lengthen the distance that a spark will occur.

Over time at high voltages the electrical barrier between internal traces breaks down allowing current to more easily jump from one trace to another. I'm not positive, but I suspect this is part of what's going on with the folks who've had their max stable OC go down, and why lowering voltages (at a lower clockspeed) generally returns the CPU to a stable state.

[3oh6]

Originally Posted by Jon_di2

Ps. The people who use liquid nitrogen to compete in overclocking contests do not OC on liquid nitrogen on a constant basis. These voltages would reduce the life of the processor dramatically if done often.

haha, define "often" and "constant basis"

[SugarJ]

Originally Posted by sswilson

One of the things that hasn't been mentioned (and I can't for the life of me remember the actual name for it....

I'm thinking it's electron saturation, but I'm not sure.