eriko :
@ alidan:
Nope, no air cooler is going to come close to a dedicated water setup.
I'm rocking the Thermaltake P5, using a 480mm all-copper rad, and no I do not use the hard tubing, because as you said, it is inconvienent.
No matter what I do - gaming, folding overnight, my CPU has never seen over 42C, even with 14 cores flat out. Thats 168W continuous for half a day, or more sometimes. Cores typically sit in the high 20's.
Do that on air.
the d15 beats out most closed loop coolers, hell, depending on case the 212 evo can beat out a closed loop cooler, though I think the person who did that review was a moron.
now as for how cool the cpu is. does it matter? you have to dissipate the heat no matter what, so you are always dissipating X amount of heat, no cooler is going to change that, the most a closed loop can do is give you a few extra c to play with when overclocking.
http://www.tweaktown.com/reviews/6313/noctua-nh-d15-cpu-cooler-review/index6.html
http://www.tweaktown.com/reviews/7301/noctua-nh-d15s-style-cpu-cooler-review/index6.html
http://www.guru3d.com/articles_pages/noctua_nh_d15_review,11.html
There are more that pit noctua against just air coolers, there are also videos that have a plethora of coolers tested and shows, and noctuas never seems lt lose by much when it does. a good custom closed loop will likely be better, But I have watched the cleaning, when things go wrong, and now much trouble it is to never want to have a water cooler, or ever recommend one when they both cost more than a noctua and do not perform significantly better.
then with ducting the fans
http://www.heatsink-guide.com/fanduct.jpg
That is a poor example, but there are better ways to do it, something rarely benchmarked that significantly improves temperatures. Personally I have a 3d printer and will be making my own custom ducting when I get my new computer, both for getting the air our of my case, and getting it the hell out of my room, but you could make your own with a soda box and some tape if you are really tight on money.
I should also mention that if you look for reviews, see if they let the water in their closed loop cooler heat up first. there were a few where they just took idle and then benchmarked on stock where noctua, in almost any other benchmark is only a few c behind was nearly 20, kind of fun learning how methodology can skew results that should be simple and objective.