I am not sure I understand the validity of this article. Did AMD and Nvidia base their TDP on furmark? Did the author of said article contact both Nvidia and AMD to find out what was used to determine TDP and use his results as a litmus as to their claims?

After reading this article I find myself asking more questions then what this article is suppose to answer. Another concern is that the use of furmark in this fashion can give the impression to others that those results are what you get in games and other programs. The truth is (if you used furmark) is that you don't draw nearly as much power and heat in games and programs as you do with furmark (at least based on my on examination of this).

Therefore, IMO this type of review should be about pushing power consumption to the max.