Personally I found it laughable that none of the testing was done with a noted room temperature and the degrees Celsius above room temp the CPU got to under XXX CPU cooler
He did specific the temp as Celsius above room temp
The big picture of 'air cooling sucks if you want to overvolt Haswell" comes across pretty well. Some of the finer points...like none of the coolers were operating with their fans at 12V when temps were recorded...but that's where the acoustic measurements were taken and now let's mix the two to create an efficiency chart is a little baffling.
It would appear that temps were also taken with CPU at full load and fan running at full 12v.
I am not trying to appear as though I am a big fan of tomshardware. However, it is fair to acknowledge that they have recovered a bit from their absolute worth and are trying to be scientific. It is hard to find good HSF review these day and this article does have a few good points to note.