PC Cooling Science Project

[lowman]

My only suggestion would be to incorporate overclocking into your experiment. Benchmark your performance stats with no cooling - "stock" and 1 or 2 "overclocked" settings. Then, run the "stock" and "overclocked" settings with 1 type of "standard" cooling (if there is a standard or basic type of cooling), and then maybe another, more robust type of cooling. The other thing to consider is the load you have on the CPU in each test, and making sure it is exactly the same in each one. You should also verify the temp of each cooling method and the room temp as well for each test, and ensure they are both uniform - otherwise those variables will leave you open for innacuracies. That's my 2 cents, and I'm very interested in seeing how this goes...

 

[NoisySilence]

Any updates ?

 

[Carrot]

my 2 penneth worth.

I have to agree with with a lot of what has been said, however, this does not mean the experiment will be a failure, even with just the original outlook.

For example, if the stock cooling temp is just that bit higher than it should be, then you should see the difference in the benchmarks.

However, this difference, in my opinion, is not going to be huge, and will not necessarily create a successfull experiment.

I would highly recomend pushing the CPU to its limits. I would suggest that "throttling" be removed for the sake of the experiment, if you feel this is safe. As this would leave it entirley to the processor.

I would also suggest, in addition to the overclocking of the CPU, underclocking it a little, maybe see how cold you can get the CPU by adjusting the clock speed rather than the cooling (and then maybe with the cooling). Maybe that is a few too many options.

But I would definately try to see if the better cooling makes any diff at stock speeds (shouldn't do, but you never know).

However, I strongly recomend you attempt to overclock the CPU to its maximum under every cooling scenario. IMO, only this way will we see how much temperature bears an influence on processing power.

Don't know much about superconducters though, be careful, and good luck.