Originally Posted by Goldenpi
It depends on flow-rate. Under ideal conditions, the hotest and coldest points of a loop should have almost no temperature difference. But the slower the flow, the greater the difference becomes.
Exactamundo -- [Shatner Voice]
the question is, how much? [/Shatner Voice]
There is a detailed mathematical proof circulating (pun intended
), but I believe the gist of it is that water requires a lot of energy to change temperature--a watched pot never boils, climates near water have slower and smaller temperature swings, etc.--so the amount of temperature change required by the moving tube of water through your CPU to transport its heat away is surpisingly low.
Here's an easy test--run your CPU up to full load, let temps stabilize, and touch the tubes going in and out of your waterblock. When I did it, I couldn't tell a temperature difference, but maybe I'm just insensitive.
That's quite contrary to the image I have in my mind of hot water coming out of the waterblock, and cool water from the radiator. Actually, if you think about it, it points to where the current bottleneck in ambient cooling really is...
Hint--it's more obvious in air cooling.