Open dubigrasu opened 1 year ago
May I ask which game?
Oh, is Rust. https://store.steampowered.com/app/252490/Rust/
OK, thanks for the report.
A framerate drop around -40% is definitely not expected, as mentioned in the README the expected performance hit is around 5%. This is the first example I've seen such behavior, so I'm not sure what's going on here; for the "balance" option, I think it would take too much complexity and isn't very doable right now.
Let's see if others report similar cases.
I understand, thank you. I'll try to find maybe another game to test, excepting those risking bans though. I only tried the Ghostrunner demo, but unfortunately the demo doesn't have NR implemented.
Warhammer Darktide 40000. gamemoderun mangohud OBS_VKCAPTURE=1 PROTON_ENABLE_NVAPI=1 LFX=1 running at ~90-100 fps in lobby. gamemoderun mangohud PROTON_ENABLE_NVAPI=1 DXVK_NVAPI_ALLOW_OTHER_DRIVERS=1 LFX=1 %command% ~60-70 fps in lobby. DXVK_NVAPI_DRIVER_VERSION=49729 make main screen character selection is just black. Nvidia 535 driver + wayland.
It is my understanding that a decrease in FPS is to be expected with the feature enabled, but the difference I see is unexpected. To illustrate I measured the FPS in identical game scenarios using at first uncapped FPS, followed by capped at 80 >70 > 60 and finally again uncapped but with Nvidia Reflex enabled.
The first section uncapped is averaging around 100 FPS, but is clearly all over the place, while the following capped sections are much better, especially the one at 60 FPS. The Reflex-enabled one is puzzling because not only its roughly half of the initial FPS, but subjectively (and in the graph) seems rather rougher than what I expected.
My questions would be: Is this significant difference to be expected, or maybe is something wrong in my setup/testing? Is there any option to compromise between reducing latency and assuring a minimal FPS, as in: "do what you can, but don't go below 60 FPS"?