| 10 Jan 2026 |
 Atemu | I've seen different swapchain settings produce unexpectedly high reported GPU usage | 07:16:19 |
| Atemu | Or does it use dynamic resolution perhaps? | 07:17:21 |
 woobilicious | Na, I assume it's async compute stuff | 07:17:58 |
| Atemu | hm | 07:18:27 |
| Atemu | Anyhow; if CPU usage is anywhere near 100%, that'd imply a massive CPU bottleneck | 07:19:05 |
| Atemu | What's the hardware? | 07:19:13 |
| woobilicious | 5900X, the game was causing a full gpu crash followed by a CPU error, it now seems stable after disabling PCIe power managemnt. but after twaeking settings for a month I can't figure out how to get it working better again | 07:22:37 |
| woobilicious | and a Radeon 5700 XT | 07:22:53 |
| Atemu | Something doesn't add up here. I'm cofused, therefore an assumption I made is wrong. That usually points to some indirect cause like a thermal or power issue.
You could isolate the bottleneck first by lowering the resolution to something tiny
| 07:30:30 |
| woobilicious | Finally I'm getting 50fps, I changed WINE_CPU_TOPOLOGY from cores 12-17 to 0-5, no clue why that matters | 07:31:56 |
| woobilicious | But it still doesn't help that the game is using denuvo lol | 07:40:08 |
 K900 | Denuvo hurting performance is honestly mostly a meme | 07:50:23 |
| K900 | Denuvo is, fundamentally, just a big fat targeted obfuscatir | 07:50:41 |
| K900 | * | 07:50:45 |
| K900 | As long as the game doesn't use it on hot code, it should not actually matter | 07:51:17 |
| woobilicious | Well yeah but good luck stingy execs allocating resources to do that last bit | 07:52:49 |
| woobilicious | It also doesn't help that a rich mans cache hit is a poor mans cache miss, I bet this game ran better on release simply due to the lack of X3D and 1MB L2 cache that devs have access to, once you release a few DLCs games run so much worse | 08:05:23 |
 SigmaSquadron | didn't several media outlets test denuvo and non-denuvo copies of games and found that most of time, denuvo-enabled games had lower performance? | 08:12:18 |
| K900 | I remember seeing some opposite results | 08:12:47 |
| SigmaSquadron | I remember ars technica doing a few articles about denuvo | 08:12:55 |
| SigmaSquadron | Okay it looks like most news outlets are reporting on one youtube video that conducted some testing on denuvo | 08:14:51 |
| SigmaSquadron | and there's one japanese video game producer that blamed denuvo for framerate issues on his game | 08:15:16 |
| SigmaSquadron | And there was one game that was misusing denuvo's API and their mistake caused denuvo to perform terribly | 08:16:08 |
| SigmaSquadron | seems like the evidence is inconclusive at best | 08:16:26 |
 magic_rb | From what i know, if you hold denuvo wrong and ask it to perform integrity checks 40 times a frame itll perform terribly. If you use it the way you're supposed to, it doesnt have any real performance impact. | 08:17:51 |
| magic_rb | (This is coming from someone who'd prefer if DRM was illegal) | 08:18:07 |
| SigmaSquadron | * | 08:18:23 |
| SigmaSquadron | Yeah, this seems to be accurate. There aren't enough tests to measure a performance downgrade accurately. | 08:19:13 |
| SigmaSquadron | * | 08:19:21 |
| SigmaSquadron | Denuvo is still very bad, of course. In 2021 they forgot to renew their domain name which caused every denuvo game to fail to launch. | 08:20:53 |
