I was wondering what is the lowest possible sustained temperature achievable with cooling? Not looking for something like Liquid Nitrogen where it lasts for 40seconds or less, but something you can actually use day after day.
Probably around -25oC to -40oC, which is cold, but I would think it would be possible to go lower, say down to -80oC to -100oC? Would that be possible to sustain?
Minus 100? Not unless you go for liquid nitrogen setups, which are the domain of laboratories and engineers (just forget about that one, OK guys?). Peltiers will take things down to minus 20 degrees C at most, but issues of condensation make it more hassle than it is worth, IMHO, as for this you exceed overclocking values obtained by water cooling alone by only a small margin. You'd have to look to liquid nitrogen setups for the full whack, which create an effect approaching super-conduction (although not exactly), and then you run into other limitations of the chip (impurities in design/material etc.). Good watercooling will get you a sustained and stable +30 C, which is a good operating temperature for a CPU, and allows for overclocking to its practical maximum.
Colder is not always better. At least it shouldn't be. I remember having a discussion with my semiconductor materials lecturer about cooling processors. I can't remember the exact details of what he said (haven't started revising yet), but it was along the lines of, there is a balance to be found between intrinsic and extrinsic carriers in the semiconductor and another property which is also temperature dependent. anyway, there reaches a point where further temperature reductions have detrimental effects on performance. He also said this temperature is not as low as you would expect. This is because a amd and intel will often tailor the dopant concentrations to give the best performance around there expected operating temperatures (I'm sure they don't expect everyone to use LN2). I'm gonna have to drag my notes out and maybe post again if I've missed something out.