Originally Posted by Phaedrus2129
Cooler transistors don't take less power... True, higher heat means more resistance, but it's a small factor, not 20%+. Maybe 2%. And that's more than compensated by the overclocking that watercooling entails.
Wrong. I've actually tested this one myself! I hooked up a digital volt & current meter to my computer, and then hooked that up to the EPS12V cable. I then took readings every 100ms for an hour (while running LinX) and averaged the amount.
i5 750 @ Stock:
Air Cooling (85c average core temp) : 71.7w
Water Cooling (40c average core temp): 65.1w
That's a 10% drop!
R = kT (roughly)
85c --> 40c is a 14% drop in Kelvins.
With the i5 750 @ 3.8 GHz the power difference was 12W. (105 vs. 93)
I have tested with other CPUs and at other clock speeds and the results are similar. With high power GPUs consuming 200W+ a piece you are going to be saving 25W+ per GPU.
10W per CPU = 20W
25w per GPU = 75w
Total: 95W+, depending on overclocks.
GPUs I assume would be similar, and due to their increased power draw,