EduardoS (no@spam.com) on 6/23/11 wrote:
---------------------------
>Correct if I'm wrong but the small area to dissipate heat from the CPU itself is a space constrain...

true, but that is why we use heat-spreaders and (active) heat-sinks. (As opposed to passive convection directly on the die.)
The advantage of water is of cause that it's heat-capacity/volume is soo much higher, allowing for a slower and more quiet flow. While the enthusiast will appreciate the noise reduction in his living room, a computing-center is always noisy.
BUT in a computing-center the air on the components is not the air from outside there is always an air-conditioning therefore my "it is always air in the end anyhow" argument fails. In a computing-center the water will go right into the unavoidable "air"-conditioning.