There’s a rather interesting video doing the rounds showing the performance impact of Denuvo when tested across several titles. It’s notoriously difficult to gauge the performance impact of Denuvo as cracks often bypass Denuvo rather than remove it, while developers also tend to push patches which remove Denuvo without warning. In an ideal world, we could benchmark a game the day before Denuvo is removed, but this is next to impossible.
Still, YouTuber Overlord has given it their best shot, benchmarking seven games before and after Denuvo DRM was removed. The full list of games benchmarked includes Hitman, Abzu, Sherlock Holmes the Devil’s Daughter, Mad Max, Agents of Mayhem, and Sniper Ghost Warrior 3. The downside is the comparisons still often aren’t perfect matches, although Overlord gets damn close.
He also only utilises a single system for the test, which is a GeForce GTX 1080 Ti paired with an Intel Core i7-2600K at stock clock speeds. This is intended to make the CPU the performance bottleneck as Denuvo would tax the CPU if it affected anything. Unfortunately, this is still a quad-core, multithreaded system, so this doesn't delve into the impact on low-end CPUs.
Here’s how the seven games shaped up both with Denuvo and without Denuvo.
|Sherlock Holmes: The Devil's Daughter||Yes||86|
|Sherlock Holmes: The Devil's Daughter||No||92|
|Mass Effect Andromeda||Yes||89|
|Mass Effect Andromeda||No||92|
|Agents of Mayhem Indoors||Yes||126|
|Agents of Mayhem Indoors||No||133|
|Agents of Mayhem Outdoors||Yes||51|
|Agents of Mayhem Outdoors||No||51|
|Sniper Ghost Warrior 3||Yes||71|
|Sniper Ghost Warrior 3||No||75|
While it's difficult to draw concrete conclusions, based upon the raw frame rate data we can see there is little noticeable performance difference between games running Denuvo and those without. Some games do demonstrate a greater frame rate drop than others, but the typical average FPS loss with Denuvo is around 2.8%. For most PC gamers this drop in performance should be imperceptible.
At the end of the day, a game that performs badly or is badly optimised is going to run badly, regardless of Denuvo. There appears to be little correlation between game performance and Denuvo implementation. While stripping out Denuvo may have bumped up the average FPS by 4, Sniper Ghost Warrior 3 simply doesn't run well.
What these results don't let us though is the impact of Denuvo on lower-end hardware. The Intel Core i7-2600K may be over seven years old now, but it's still a decent quad-core CPU with multi-threading. We know for sure that Denuvo does use CPU resources. It has to, it's an application running in the background. For weaker CPUs, it's going to be a proportionately larger hit. What this indicates though is that the performance impact of Denuvo should become less of an issue as time goes by and gamers have access to more powerful processors.
However, there is also the issue of frame times. Frame rates only tell us half the story, and in Overlord's we can see a better look at frame times and, in particular, the highest frame time for each benchmark. A high frame time is where we'd notice stutters and performance hiccups, even if the frame rate is being maintained.
As you'll see in Overlord's video below though, the frame time analysis is still inconclusive. In examples like Sherlock Holmes we can see the max frame time is actually lower with Denuvo enabled. It's the same story with Mad Max, but in another test, Agents of Mayhem has a lower max frame time without Denuvo.
During the video, there’s also a reference to an article here on GD benchmarking Doom both with and without Denuvo. As Overlord says, it’s an imperfect art when the game was benchmarked eight months apart, but you can see the full article for yourself here.
Okay, so it feels as if we're still no closer to answering the question once and for all of whether Denuvo affects frame rates in PC games. The statistics tell us there is the smallest of dips, although it's practically imperceptible.
What are your thoughts? Are you convinced Denuvo is responsible for poor performance during your gaming sessions? Or is it an easy scapegoat when poor performance may be the result of the game itself? Let us know what you think below!