The answer to that is way beyond anybody here. It would require a model of the CPU and take operating temperatures, material properties, and voltage as inputs.

What we do have is the manufacturer's specifications for a similar processor as quoted a couple pages back.

I think anyone considering overclocking/volting any device should read about electromigration. I learned about it in class but the wikipedia article is really pretty good. See in particular this section: http://en.wikipedia.org/wiki/Electro...7s_equation.29

Note the exponential dependence on temperature and the (approximately) inverse square dependence on current density. You can make rough estimates as to how much raising CPU operational temperature will accelerate the process, and maybe make a rough guess as to the relative effects at different operating voltages based on the manufacturer stated lifetimes, but (1) this is only one mode of failure (2) doesn't capture the exact defects of your particular device edit: and (3) you can't really get at accurate on-chip temperature information, anyway.

So in other words, any estimate here would be only of academic interest.

Two different issues though, really. There's the immediate ability of the chip to run at a given frequency/voltage, and the lifespan of the chip. Running at elevated frequency/voltage/temperature reduces the lifespan of the chip (approximately) with the inverse square of current and exponentially with temperature. It's not linear at all. So while there is a linear scaling for temperature within a given range of frequencies, there's an exponential decline in lifetime.

Anyway, as i said, it's mostly a point of academic interest and not really of much value to this thread because those who buy a new $600 phone every six months (shocking statistic there from Javier) don't care, and those who do care are just not going to go crazy with the overclocking.