Denuvo performance test by Richard Leadbetter. Bottleneck caused or eliminated? Am I right or wrong?

[Pelihemmo]

Here's the performance test.
https://www.eurogamer.net/digitalfoundry-2019-devil-may-cry-5-pc-denuvo-protection-tested

I'm having a debate about this performance test, why they (or Richard) did the test with extremely low graphics resolution, and did this cause a bottleneck or was it an attempt to eliminate a bottleneck.

The other guy says the test was ridiculous, they caused a CPU bottleneck because of low display resolution. In his opinion "CPU bottleneck" means the case where fast graphic card must wait the CPU. They caused this and it's a problem.

I'm saying they eliminated a possible bottleneck caused by the graphics card by keeping it almost unemployed. They were thinking the performance of the graphics card could slow the system. No bottleneck was caused, the test is basically valid. And that "CPU bottleneck" only means a case where CPU is the slow component in the system. In this test case they were thinking the graphic card could be the slow component of the system.

IMHO this shows clearly that the graphic card was the possible bottleneck, and that's why they reduced the load on it:
"the choice of graphics hardware and in-game settings is much more of a potential performance bottleneck. "
and
"the GPU removed from contention as much as possible, the CPU becomes the limiting factor in performance"

The game is said to be GPU load heavy, while Denuvo is said to be CPU heavy. So in my opinion that's why they wanted the load on the CPU.

 

[Kaamos_Llama]

Welcome to the forum!

The benchmarks youre referring too arent obvious me from the article you linked, am I missing something

You probably know but generally speaking, if youre benchmarking a CPU for gaming performance then its best to do so at a lower resolution, with the lowest graphical settings with the fastest GPU available. If the article you mention is trying to find out whether Denuvo is causing increased CPU load, then loading the CPU in that way would make sense in order to see exactly how much of an effect it was having.

If the game is so graphically demanding that it wont ever hit that CPU limit anyway its kind of not relevant to performance at high graphical settings, but interesting to see how much of an effect the DRM has anyway.

 

[Pelihemmo]

Welcome to the forum!

The benchmarks youre referring too arent obvious me from the article you linked, am I missing something

You probably know but generally speaking, if youre benchmarking a CPU for gaming performance then its best to do so at a lower resolution, with the lowest graphical settings with the fastest GPU available.

Ah! I copied the link from wrong article! Thank you! Also, copied the wrong article author name too. Fixed the link and the name.

Haven't read game benchmark articles, I've read hardware benchmarking. So, that's a new thing to me.

 

[Kaamos_Llama]

Ah yep Digital Foundry/Eurogamer does really good work with benchmarking games and showing the effects of different graphical settings. Some of the stuff they look at is kind of niche and general users may not notice some of the visual things they focus on. I know I dont really see a lot of the minor visual stuff and I'm not that sensitive to stuttering unless it gets really bad.