The YouTube channel "Maximum Fury" conducted a technical test of the new Cyberpunk add-on called "Phantom Liberty" on an older AMD hardware system, testing it separately on Linux and Windows 11. The Linux system, specifically the Fedora distribution called Nobara, performed significantly better, delivering 31% more frames compared to Windows 11.
The hardware used for testing included an Asrock B550 motherboard with an AMD Ryzen 5 5600 CPU and an AMD Radeon RX 5700 XT GPU from the first RDNA generation, along with 16 GB of DDR4 RAM. The CPU, RAM, and GPU were overclocked, and the system utilized undervolting to save energy costs.
When testing the game at 1080p resolution with high textures, the Linux system achieved an average of 63.72 frames per second (fps), while Windows 11 managed only 48.55 fps. This suggests that the game should run noticeably smoother on the Linux system.
Wait, DLSS doesn't work on Linux at all? That's a pretty big thing to gloss over whenever someone is talking about linux gaming and how comparable it is to windows nowadays. I doubt I'd be able to get anything remotely close to a stable framerate on cyberpunk2077 without it, and same goes for other newer games like dying light 2 or starfield!
DLSS works. It took a while longer than Windows, but Nvidia themselves actually provide Wine-compatible DLL files. Also, there's a native way to implement DLSS for Linux which, I kid you not, zero games so far are using. The Windows version works fine though.
But DLSS Frame Generation and Ray Reconstruction do not work, and there are zero workarounds.
Oh, so he's just talking about DLSS3 features, gotchya. I thought DLSS 1 performance improvements are also frame generation but I see now thats different
DLSS is upscaler. Game is rendered at lower resolution and then image is upscaled in a bit smarter way than simple "stretching".
More precisely, DLSS is a set of models that use AI to interpolate an image. This interpolation can take many different forms:
Interpolation can be used to take a lower resolution image and upscale it, which is the main feature of DLSS.
You can also use DLSS to take a high resolution image and scale it down, with less artifacts, as a type of antialiasing. This is DLDSR.
You can also use it to take information from an image, combined with motion data, and interpolate how blocks of pixels might change into a new frame. This allows you to generate intermediary frames. This is Frame Generation.
You can also take a very noisy image, composed of discrete dots, and interpolate how neighboring pixels should look. This is Ray Reconstruction.
Aren't frame generation and ray reconstruction new? I'm sure they'll work one day, although I'm not a big Linux head I only use steamos on the deck I just see a lot of Linux posts on Lemmy so here I am lol.
They're new, yes. Though the folks at Proton already confirmed they'll provide no workaround to support it, Nvidia needs to build the Linux drivers with official support. We don't know if they'll do that and when.
Plain DLSS definitely works, I'm guessing they mean that specific reconstruction feature. I'm sure it'll be implemented eventually if it's possible at all though.
Side note, a kind of related feature that is missing for sure from the Linux drivers is DLDSR, and plain DSR for that matter. As a heavy user of both, it's a bit of a personal deal breaker.
DLSS works fine on Linux, but I don't know about frame generation and ray reconstruction specifically. It could be those two don't work yet.