We all know that copy and DRM protection eat CPU cycles, and that can influence game performance. The problem is that to be able to measure that you would need to compare a cracked game with an original one, that all by itself is a red flag. Recently, however, some game scrapped from Denuvo and that effect has been tested on seven games. 

First off, credits go to Overlord who posted a Youtube video on it. Secondly, it's good to see he is using our homegrown Afterburner RTSS overlay ;) Overlord tested 7 games with and without Denovo activated. The results are a bit all over the place. Now, he has been using a Core i7 2600K in combo with a GeForce GTX 1080 Ti, that was done on purpose to show a greater effect. Personally, I believe that you should test modern games with modern hardware, so the results are a bit let's call it 'saturated and bottlenecked', both pieces of hardware are not exactly a match either.

Among the titles tested are: Hitman - Mass Effect: Andromeda - Abzu - Sherlock Holmes the Devil's Daughter - Mad Max - Agents of Mayhem - Sniper Ghost Warrior 3. A few games showed performance drops with Denuvo enabled.  Mass Effect: Andromeda dropped 7 FPS towards 57 FPS. Mad Max saw a more meager that was 6 FPS coming from 60. Pretty much all other games including Hitman, Abzu, did not show a significant enough perf difference or scaled negative.

So on a Core i7 2600K with a GTX 1080 ti at worst the performance drop was 10%. While that is noticeable, I wouldn't call it substantial. With a modern age CPU that difference would have been smaller.

Seven games benchmarked before and after Denuvo got removed