Today's a scheduled holiday after the festivities yesterday (Go Hawks!), so I decided to run a few benchmarks and compare for myself. I daily drive a Ryzen 7 5700X3D paired with a Geforce RTX 4070 Super and 32 GB of DDR4, so not upper-upper echelon but pretty robust in my experience. I chose to run a set of blender benches for a standardized score, Black Myth Wukong benchmark for DX12 performance, and finally some lighter fare in a default Mangohud overlay in Fallout: 4. As for control variables, both games were set to run in Proton compatibility setting 10.4 in Ultra Settings without motion blur (makes me sick, literally). And my conclusion is... there's extremely little difference to the naked eye.
There is some difference, such as in BM:W there is a sharper frame drop at around a quarter through with the open driver set. Yes, there is the DX12 issue remaining and I personally will use whatever driver set first alleviates that issue. Hope the data helps in any way possible! It's been a while since I last benchmarked, feels like my ol' Windows days when I was paranoid about FPS. Ahh, how times change. I'm actually very impressed by the open driver's performance, and am considering changing to the open set. Huh, times really do change! Driver parity!
Open = Top Image : Closed = Bottom Image. Drivers utilized were selected from the current KDE Discover hardware driver section. I ensured as well I was using the most updated nvidia-open-current, and the respective proprietary version.
