For the life of me I can't find the video.

There was one on YouTube where the guy was using Mesen.  He showed how loading one of his sprites or palettes I forget. Anyway, he was able to show the number of cycles the NES used to process the PRG code before and after he changed one of the subroutines. There was an option he set on Mesen that displayed the cycles actively on the screen.  I've gone through just about every option on MESEN and can't figure out. 

Anyone know how to setup MESEN for this "cycle test" that allows me to see if the code is more efficient or not? It was a really interesting setup and I'd find it valuable, especially since we are dealing with hardware that can't afford to waste cpu cycles and I'd like to test some of my own theories as well. 

Thanks!

OleSchool