Okay gonna try to add some perspective here in the whole AIR v WATER debate.
On my 955/965 BE
platform, I ran
CM Hyper 212 plus (Single and Dual fan)
Corair h50 with single and P/P.
Custom loop with a HWLabs 360 Stealth dual DDC single pump dual pump operational... no GPU only MB and CPU on a EK Supreme HF block...
On my CIVFormula. All testing was F@H to give me reasonable feedback on temps. This was my only goal at the time.
On day one I ran stock cooling. F@H got my 955 up to 60c which is fine but I only ran stock for a base value. 60c over 24/7 is reasonable value for Stock. We know there isn't much cooling gains from a stock cooler.
So I replaced Stock with the h50 I purchased with the MB, CPU and 16gb of Corsair AMD sticks.
Single fan setup ran cooler but I saw that F@H maxed at 55c so I gutted the stock fan and used it between the Push fan I added and max temps with F@H showed at 48-50c with Push/Pull. Ran that for 6mos and decided that I wanted a custom loop.
It took longer than I expected to put the loop together so I got Hyper 212 and saw a 4c gain with dual fan setup and no gain at all with single fan setup. Single averaged roughly 48c which is what my h50 was conservatively getting in Push/Pull with a shroud. So air and water were giving me the same temps. I may never have figured this if I had simply gone with the 212 as h50 cost twice as much at the time. But I wanted to see for myself what water would give me when I started purchasing my sig rig.
So I ran 212 for a few months until I got my loop together. Having lapped my CPU, I got Temps down even further. Roughly 10c in gains at slightly above 40c average with the h50 and the 212.
Got rid of the h50 and kept the 212. But again my want of a quieter system was paramount as I game without a headset and I like to hear my games and not my system.
Played around and lapped the stock cooler and at Idle system ran 40c. When Folding it ran 52c. Again showing the futility of running the Stock cooler.
Around this time I got my hands on a lapped 965 for not much performance gain. Which increased Temps 4c average. Tried over clocking it and got it up to 4.0 but heat made a difference in stability. BSoD every time I tried to Fold at that clock speed.
Finally comailed the rest of the loop and averaged 40c @ 4.0ghz with a MB block and CPU run.
Dropped in a stock 1100T which bumped Temps up again at 5c over 40c with the 360 Radiator. But it showed me that water is the way to go if I intend to OC.
I have firsthand experience with both. And imho AIR is okay but water will give better results for OCing. Also as stated my main interest was heat and removing it from my system, but that soon changed to heat and quiet. F@H is simply a heat test for every system I build. I've uncovered bad chips with F@H including a freshly launched FX8125. That chip went beyond 65c right off a freshly built client system. No way I was going to saddle my client with a dog chip. I might've put in my sig rig but I didn't update the BIOS and had I done so I wouldn't have been able to go back to the 1100T. Kinda wish I did, if only to see how well water corrected the issue. But I suppose that it still wouldn'tve mattered with a bad chip anyway.
Both AIR and WATER have their merits but water is indeed better. Phase and Chillers are even better but the goal with those is to run under the Ambient wall. No AIR or WATER is going to do that without assistance of an external source. AC or Chiller. Now that I have two full time rigs I will likely add a chiller to them and run them off that. But for now I will stay with a 360(/+) but I am not looking to break the ambient barrier any time soon.