I'm supposed to reorganize the heatsink review "Link" catalog here. I'm going to post a separate thread by way of "notes" -- so we can refrain from posting "replies to the "catalog" of review links, and post them on the latter thread for continuous discussion.
I had written a piece for that second thread on the factors of interest in selecting a heatsink. For performance alone, if you found comparison reviews that show "thermal resistance," you'd just pick the heatsink that had the lowest number. Thermal Resistance encapsulates all the information of idle and load temperatures, room-ambient test conditions, or just about everything but fan speed and CFM. These latter things will cause thermal resistance to vary over some range, and you'll find reviews (as at Overclockers.com) where the index is shown for "low" (reported) and "high" (reported) fan speeds and a certain size fan. Fans of a similar size are in a similar class when it would come to throughput, provided they can be made to run at the highest speed of any other fan in that class.
If the thermal resistance isn't reported in a review, then a comparison between heatsinks needs a common benchtest configuration and use of the same room ambient, the same thermal interface material, and the same stress-test software. Given those controls, the heatsink or cooler with the lowest load temperature would be showing the best performance.
But performance is not all there is, so to speak. Size, shape, weight, noise, price and other factors will be part of your decision. They are part of my own decision, even though I tend to pick the heatsink with the highest performance.
In a narrow range of thermal resistance measures, you will hear people say that inconsistencies in test conditions or just limits on accurate measurement leave the "clear winner" undetermined. That may be the case.
Here's one or two of the most recent Anandtech reviews. The focus of the reviews -- given by the titles -- may seem to bias or favor some choice, but the results are most likely as "kosher" as it can get -- given the test-conditions, the bench-configuration and controlled variance in the configurations, and so on.
A heatsink comparison review
Another heatsink comparison review
Before anyone says "He's pushing [this or that manufacturer]," and since I'm supposed to take over the "catalog of links," let me say that these reviews were about the easiest thing for me to find without leaving our (home) web-site here. Other information should be available from that link-catalog -- still posted under Lifeguard's name. that's why we want to compile that information.
Some reviews are filled with hype. Some reviews don't report useable facts. Sometimes, in a comparison, only favored advertisers are reviewed, even if the numbers used or obtained in the comparison are accurate, so the comparison leaves out useful information about other products.
Never base your decisions to purchase a cooler -- or anything else -- on a single review.