Originally Posted by
Agent
This might have been the case years ago with some of the very old CPU's, but I'd love to see some evidence of it on any modern architecture.
At best you're talking about a few degrees - in a lot of instances the over-clock is limited by other factors before temperature (excusing crazy LN2 setups and the like), assuming you're using decent kit. Multiple times I've witnessed how little of an increase water cooling offers on a modern setup (I'm talking first hand experience, not reviews) over a decent air solution.
So thermal paste offering an increase in overclock, given how small of a difference it makes? I'd say it's going to be impossible to even measure that.