News Cyberpunk Phantom Liberty Runs 31% Faster on AMD GPUs in Linux Compared to Windows 11

[Admin]

YouTuber Maximum Fury benchmarked Cyberpunk 2077's latest 2.0 update in a modified version of Fedora Linux known as the Nobara Project and unveiled that the OS is a whopping 31% faster running Cyberpunk 2077 compared to Windows 11, at least on an AMD RX 5700 XT.

Cyberpunk Phantom Liberty Runs 31% Faster on AMD GPUs in Linux Compared to Windows 11 : Read more

 

[thently]

Why would anyone play this game on an AMD video card.. Nothing against AMD but the pull for most people on this game is the new tech. I myself could not get into it but I tried it to see the new tech myself.

 

[Viking2121]

Why would anyone play this game on an AMD video card.. Nothing against AMD but the pull for most people on this game is the new tech. I myself could not get into it but I tried it to see the new tech myself.

You know "New Tech" isn't the only thing this game offers right? Seem to forget about that I guess.

 

[ezst036]

The future of gaming on Linux is very bright! This particular item has a likelihood of being an outlier, however, Linux is so easy to use and it offers the ability for game manufacturers themselves to create performance tweaks for their own games in a way that's not possible on closed-source alternatives.

 

[ezst036]

Why would anyone play this game on an AMD video card

Three simple words: open source drivers.

Everything in Linux "just works" with open source drivers in a way that the average Windows user can understand like plugging in a USB mouse or keyboard. Just plug the video card in and it works, plug in and go. Compare that to having to deal with a vendor's binary blobs/installshield packages. That's an old-school way of accessing video drivers.

(To be fair, Intel also has open source drivers which will make the plug and go experience the same)

 

[Makaveli]

Those are interesting results would love to know what the cause is.

 

[tstng]

I guarantee something is off here. I'm almost certain that the Linux bench was done carefully with a million tweaks and optimizations while the Windows test was just done haphazardly. Cp2077 is notorious for changing settings on you when you don't expect it (like turning dlss/FSR on or off without telling you).

 

[watzupken]

Cyberpunk Phantom Liberty is very CPU intensive, or at least that's what the developer said. So if this holds true, then this just shows how resource hungry Windows 11 really is.

 

[watzupken]

Why would anyone play this game on an AMD video card.. Nothing against AMD but the pull for most people on this game is the new tech. I myself could not get into it but I tried it to see the new tech myself.

This "game" is popular mostly because it is like a graphic showcase. I don't even think many people are actually still playing thus game. So no surprises, this game is still largely used as a benchmark.

 

[TheOtherOne]

Who would've thought Micro$oft and nVidia have products that not only work better together when paired but ALSO it affects negatively on a third party product?

View: https://media.giphy.com/media/3DudglxvlDmJfCVQjL/giphy.gif

 

[kerberos_20]

Who would've thought Micro$oft and nVidia have products that not only work better together when paired but ALSO it affects negatively on a third party product?

you know that on linux, starfield runs on nvidia with twice more fps, right?
that is with extra dx12 to vulkan translation

 

[mitch074]

I guarantee something is off here. I'm almost certain that the Linux bench was done carefully with a million tweaks and optimizations while the Windows test was just done haphazardly. Cp2077 is notorious for changing settings on you when you don't expect it (like turning dlss/FSR on or off without telling you).

One massive tweak : open source drivers. What does it mean? Developer that own a game can look into the driver for a bottleneck. Bottleneck found? Driver updated right away, fix is tested, validated and part of the next RC. At worst, it comes as stable with the next Mesa stable release once every 3 months.
A specific bottleneck can hinder performance by a LOT, and be very hard to fix when the driver's dev can't communicate with the game's dev directly. Now, either the bottleneck was solved in the open source driver faster than it was on the official AMD driver, or it was never there in the first place.

 

[coolitic]

It's probably because DXVK, the DirectX-to-Vulkan translation layer, is optimizing some poorly-made graphics code on CP's end.

 

[Shadow_Death]

I wonder if it has anything to do with exploit protection or Control Flow Guard in Win 11. Supposedly people managed to fix hitching in DX12 games by bypassing that for the games executable. I haven't tried it since the update so maybe?

 

[Joomsy]

It's probably because DXVK, the DirectX-to-Vulkan translation layer, is optimizing some poorly-made graphics code on CP's end.

I have to agree with you. This can usually be tested with notoriously bad PC ports, too. My favorite is GTA IV, as Rockstar really flubbed the DX implementation, and it's what first showed me just how egregious the performance difference is between Windows and Proton in certain scenarios. Under Windows, I'd be lucky to pull 30 to 45 frames with mid to low settings running at 720p. With Proton and DXVK, I can play with ultra settings at 1080p, and it'll never dip below 55 frames.

For extra context, this was all on a mid-2015 MacBook Pro with an AMD GPU. Under Windows (10 installed through Bootcamp), I used "custom" AMD drivers, as the ones provided by Bootcamp were severely dated. This was necessary to even get the performance I mentioned above. I also used Afterburner with Mac's Fan Control to squeeze out a tiny boost. This machine easily triggers thermal throttling, so any overclocking usually leads to negligible differences. I never applied any under Linux, though, I did install a fan control daemon to help with the throttling. It should also be noted that I used the Xanmod kernel, because it has fsync and esync support. I may not have gotten the same performance with the stock kernel, and this is something important to know if somebody's new to Linux. The kernel you use will largely determine the performance you see, and they vary from distro to distro. This article doesn't go into detail about what tweaks and mods the tester may have put in place, but I'm not familiar with the distro they tested on, either, and it sounds like it has gaming in mind. I just don't want people to read this article, think they can slap Ubuntu on something, and expect to see similar results. This leads to them just going back to Windows.

 

[sundragon]

Why would anyone play this game on an AMD video card.. Nothing against AMD but the pull for most people on this game is the new tech. I myself could not get into it but I tried it to see the new tech myself.

As a Steam Deck deck owner (AMD GPU/Linux) any additional performance is a good thing. The game runs decently on the SD and people were worried about the expansion's performance. This may improve sales of the DLC.