Geminiwave

Its kinda a silly idea, but I was thinking about this the other day.

When you shrink the die size of a processor, it requires less power, creates less heat, and can generally run faster. All by simply making the space smaller. Making shorter pathways for the electons etc etc etc.

I had this kinda crazy idea and I think I'm going to experiment with it later on when I have more time. I got it while at the Boeing museum of flight and watching the SR-71 Blackbird tour.

When the blackbird gets going at mach 3.3, the cockpit grows over a foot in length from the extreme heat. The engines increase in size by more than 10% in some cases. Well the opposite happens when you put materials under extreme cold conditions.

Now back to the CPUs...I am wondering if a processor would actually run FASTER if you supercooled it. I am not talking about overclocking potential. I am wondering if a stock 3ghz Pentium 4 or an Athlon XP 2800+ or whatever proc you want would run faster at subzero temps than at normal temps. I had hoped that THG would have measured such a thing when they did their race for the fastest proccessor. Sadly no luck there.

Anyway what do you guys think? If the processor is super cooled, then the die would shrink from the atoms compressing and then electrons would have a shorter distance to travel inside the processor...meaning faster processes. Probably not by much, and you'd still have memory and harddisks etc etc running at normal speeds. But I think i'd like to try this. Sandra would be just about the only program that'd pick up the difference.

lemme know if anyone has tried this before and theres an article on it or something. Thanks

__________________

dj_stick

mmmm the theory seems right, don't know if its been tested though

__________________

Judas

yes it does have an impact. SuperConductors are usually created when the material is brought down to an extremely low temperature. The lower the temp, the more effeicent and usually quicker the flow. Now if you slapped a midrange rig today up with say -100*C, i doubt you'd see a 5% increase in preformance from stock. However, like anything, it provides a much higher overclocking ceiling. Or it's one hell of a way to make your system cool and quiet

__________________

Geminiwave

Yeah I knew that the increase in performance would be low since we're simply talking about the processor here. You would have to supercool every part of your computer and I think that some parts such as optical drives and\or harddrives can be damaged if the temps are too low.

Anyway I dont know if I can get some liquidnitrogen to test this out sadly...I really want to see if theres a difference even at somewhat more modest (like not -100C but still below zero) lowtemps. I think Sandra would be just about the only thing capable of detecting that difference.

__________________