Bona Fide:
Muscles:
I know only second-hand about water-cooling. When I started with building my "MOJO" in early 2004, I committed to air-cooling with heatpipes.
I was able to test something like six different heatsinks, some with heatpipes, and some fifteen or twenty fans of assorted sizes. I only stayed away from DELTA because they make powerful fan motors, and all the forum geeks who spoke of them joked about "the noise." But the bearings are good, and at RPMs commensurate with mid-range Panaflo fans, they aren't any noisier.
Another thing I found out was this. All heatsinks and heatpipes -- even water-cooling blocks under maximum flow conditions (assuming that water is not "chilled" but at room temperature) -- have a minimum thermal resistance. While I didn't tabulate my results with high-throughput fans, there is a CFM threshold beyond which thermal resistance remains the same.
Today, based on a contention about "what makes a cooler a 'good' or 'best' cooler," I tuned down my Delta Tri-blade 120x38mm fan mounted on my XP-120 cooler. Before today, I had it running at around 2,500 rpm, allowing it to spin up close to 3,500 rpm when a temperature had been reached on the way to full CPU load conditiions.
I tuned down the idle end of that speed range to about 1,500 rpm. Idle temperature didn't change, although you would think it would rise a bit. Perhaps that says something about heatpipes in general, or the ThermalRight coolers in specific -- I'm not sure. But it also turns out that pushing anything above 110 or 115 CFM doesn't reduce the load temperature any further, and I think even lower CFM than 110 would hold the load temperature at what it was under 130 or 140 CFM.
So a 200 CFM fan doesn't help beyond some level far below 200 CFM.
HOWEVER -- what high throughput fans do, despite a limit on their effect on the heatpipe CPU cooler, is to further cool the motherboard, chipset heatsink and memory modules. But even those temperatures did not change when I tuned the idle fan speed down to 1,500 rpm.
The more powerful fans have heavier motors, and one would want a lighter fan just as one would want a lighter heatpipe cooler. On the other hand, there is probably something to be said for overkill: if the motor was designed to push that much air, it probably has a greater longevity at half the speed than the 50K or 100K hours specified for the fan.