CPU overvoltage question

[Synaps3]

I am overclocking a CPU. I am able to maintain a very consistent temperature using a radiator and compressor as well as a TEC. In order to overclock any more, I have to raise the voltage above what most people say is the "limit" (1.5V). It is an AMD FX-4100.

My question is: Does running a CPU above the specified voltage cause degradation? My guess is no. I would think that once you got to a certain voltage, it would short, and break, but running right below that voltage would not cause degradation, or would it?

[Neon_WA]

running any processor outside it's voltage limits will shorten the life of the processor
by how much.. is dependent on how well the processor is kept cooled

a 5 degree increase at the heat-spreader may mean a 30 degree increase at the core as the heat-spreader can only transfer so much heat

most modern CPUs have thermal trips built it.. but it the area of concern is running the processor just under this for long periods

[Guest]

Let's say the CPU was kept at the perfect temperature at all times. If the CPU was overvoltaged and remained at the same temperature, it would not shorten the life of the processor, right?

So insufficient cooling would shorten the life of the processor, not overvoltage?

[Neon_WA]

yes and no.. both are linked.. over-voltage creates more heat

if you can control the heat.. then it shouldn't affect lifespan
just means you will have to clean the heatsink/fan more often to maintain the processor in the preferred temperature range

due to the lag in the heat-spreader transferring the heat.. some people remove the heat-spreader so they can apply heat-sink directly to the core
but damage to the core becomes a lot more likely

[Guest]

[Neon_WA]

[Guest]

Thanks. This is the quickest responding forum I've been on.

[Neon_WA]

[Neon]

Both higher voltage and higher temperature can decrease the life of your processor.

One of the failure modes of integrated circuits involves Electromigration. The empirical Black's equation describes the mean time to failure for a conducting wire, and the failure time depends on temperature and current density. Current density depends on the driving voltage.

Of course, how much the lifetime is decreased is the question. Maybe it goes down from 20 years to 18 years; maybe it is much worse - it depends on how hard you push it, and the characteristics of each individual processor. I don't have experience with failure testing AMD FX processors, so I don't know the numbers.

I talked to a couple of low temp (LN2) overclockers, Kingpin and Sammi Makinen. They observe that when a chip is about to die, it is unable to overclock as high as it previously could, and then eventually it just won't POST at all.