Why colder = better o/c

[shuRe]

I was just wondering the physics or whatever of why when a cpu is cooler / below 0 you can o/c it higher.

Any info would be appreciated.

Also, this might sound a bit dumb, but o/cing isnt physically changing anything in the cpu (i pressume), so what happens when you change the multi and htt/fsb in computing terms?

Thanks

[mtb856]

As I understand it, silicon conducts better when it is colder (up to a certain point). This graph shows what I'm talking about:

http://www.ioffe.rssi.ru/SVA/NSM/Semicond/Si/Figs/151.gif

While the thermal properties of the silicon used in processors may be slightly different, the same idea applies. The graph shows that silicon has it's highest conductivity at -200°C.... which coincidently enough is very close to the temperature of LN2 (-196°C), which is why it is so heavily used in extreme overclocking. That's also why liquids colder than LN2 (such as liquid helium.... although it's a super-fluid, so it wouldn't be wise to use it anyway) are not used, because it would actually start to lower your overclock.



As for your other question, what you are changing is the amount of work that is done every second by the cpu.

Say that this is the clock cycle of your stock cpu:

_/-\__/-\__/-\__/-\__/-\

______<data>_________

and it takes 4 seconds for it to complete that operation.

Now when you overclock, you're increasing the speed of the clock pulses in the cpu, so the same operation that took 4 seconds at stock speed might only take 2 seconds at 4 gigs.

I hope this helps a little bit.... I'm not an expert in silicon yet, I just started studying it in my EE classes.

[shuRe]

thx for the explanation, massive help.