Intel warranty when overclocked

[AD-15]

hello,

I recently purchased a core i7 920 D0 from Scan. I currently have it overclocked, and have kept to reasonably low voltages so far (1.33V vcore, 1.35V VTT).

From what I understand, Intel's warranty is not invalidated automatically by overclocking. However, at what point exactly does overclocking void the warranty?

What if I keep it at these levels, but it develops a failure? How will Intel/I know if the failure is due to a fault, or simply because of the OC?

Thanks,

Odai.

[AD-15]

Anyone?

[wesleyaldred]

Warranty is voided if the cpu fails as a direct result of overclocking, IE the voltage is increased to high and the CPU blows.

Intel will know if the failure is a direct result of incorrect overclocking.

[AD-15]

So if I have the voltage quite high (1.4V+), but the temps are under the limits, I have tested it's stability, and I can run it without any problems/degeneration for a long period of time, then Intel will honour the warranty if it fails suddenly?

[Ph4ZeD]

No they won't. You are deluded if you think you can modify a product then expect a company to honour a warranty if it fails.

[Durzel]

Temperature isn't the only thing that kills CPUs, too much voltage will cause electromigration over time as well. The question is not so much if it will fail, but when (it could be days, weeks, months or years).

Suffice to say that you shouldn't expect warranty service if you have deliberately run the product outside of its specifications and it can be proven that the CPU failed to excess voltage, heat or whatever.

[AD-15]

If it runs fine though, without any degeneration over a long period of time, then wouldn't this mean a sudden failure cannot be due to long term heat or electromigration?

[Agent]

If it runs fine though, without any degeneration over a long period of time, then wouldn't this mean a sudden failure cannot be due to long term heat or electromigration?

No, why would it?

You cant stop electromigration - it's impossible.

[AD-15]

I just thought it was something that happens over a long period of time, and that the effects you'd see from it (degeneration?) would happen over a long period of time too. Hence my question, if you don't see any degeneration for a long time, even with 24/7 use, then it means no significant electromigration is happening, and that a sudden failure out of the blue can't be due to the OC?

[wesleyaldred]

Unfortunately there are no blanket answers that you seek, Intel will deal with each case individually and make a decision based of facts with that case.

Wesley

[Agent]

....if you don't see any degeneration for a long time, even with 24/7 use, then it means no significant electromigration is happening....

And how exactly are you going to measure this "degeneration"?

[Bugbait]

I just thought it was something that happens over a long period of time, and that the effects you'd see from it (degeneration?) would happen over a long period of time too. Hence my question, if you don't see any degeneration for a long time, even with 24/7 use, then it means no significant electromigration is happening, and that a sudden failure out of the blue can't be due to the OC?

Running a CPU beyond it's rated/recommended specification usually shortens its life expectancy. How much is hard to determine as there are so many variables. If you expect a CPU to last 10 years at the rated specification, it's not unreasonable for it to fail in 2-3 years depending how far you've pushed it.

Your logic is a bit flawed. You should view it more like this (generalizing a lot): The CPU will function "perfectly" at 75%. It started at 100% but due to overclocking it has been degrading over time, but while it's above 75% you won't notice it. After a few years it drops below 75% and all of the sudden, being a binary world and all you start seeing failures in the form of blue screens. So in some ways you're right, the degradation occurs over time. However, the flaws may not be immediately apparent.

My Q6600 runs perfectly at up to 3.8GHz in Vista Ultimate x64. When I installed Windows 7 Ultimate x64 I encountered all kinds of problems under load. Surely this couldn't be the overclcok as it was fine for over a year in Vista? Wrong. Dropped the settings back to 3.6GHz and everything has been fine ever since. Give it another year and it might fail again. This is most likely due to my overclock and I wouldn't even think about asking Intel for a warranty should it fail.

BTW, I think the word you're after is degradation not degeneration. An electronic component degrades, it doesn't really degenerate.

[AD-15]

hello,

Thanks for all the input!

I was going to watch for degeneration by simply looking out for signs of instability, or if I needed to up the vcore a bit etc.

Guess I'll have to be careful then.