Linux GPU Gaming Benchmarks on Bazzite OS

The video from Gamers Nexus benchmarks GPU gaming performance on Linux using the Bazzite OS (0:49), responding to community requests. The main goal is to give data to players considering a move from Windows to Linux for gaming (0:20).

Linux isn’t for everyone (3:25): Gaming support has improved a lot, but compatibility with professional software can still be a problem. Even so, with Windows becoming increasingly invasive and telemetry-heavy (4:31), the creator encourages users to at least experiment with Linux (4:29).

Methodology and disclaimers (5:27): The Linux testing workflow is less automated and polished than the channel’s Windows-based benchmarks (5:59). As a result, direct one-to-one comparisons between Linux and Windows charts are discouraged because of differences in OS environments, game builds, APIs, and capture tools (6:13).

Choosing Bazzite (7:29): Bazzite is selected over SteamOS (not officially released for desktop) and ChimeraOS because it is popular, gaming-focused, recommended by Wendell from Level1Techs, and the devs are accessible for support (7:56). As an immutable distribution, Bazzite helps maintain a stable, predictable test image (8:38), prioritizing stability, compatibility, and recoverability over chasing the absolute newest software (10:24).

Testing challenges (11:37): Linux components update frequently and on different schedules, which makes locking down a consistent test environment difficult (12:09). To mitigate this, the team freezes Bazzite, GPU drivers, and Proton versions for each test cycle and carefully documents every version used (13:00).

Gaming performance benchmarks (19:27): The video shows detailed results across multiple titles and resolutions (1080p, 1440p, 4K), including Dragon’s Dogma 2 (19:27), Black Myth Wukong (21:49), Resident Evil 4 (26:36), Starfield (28:38), Dying Light 2 (30:37), Cyberpunk 2077 (32:50), Baldur’s Gate 3 (34:45), and separate ray tracing runs (37:24). In Black Myth Wukong, Nvidia GPUs suffer from poor 1% and 0.1% lows and uneven frame pacing, while AMD GPUs deliver much smoother performance (24:29). In Resident Evil 4, AMD appears unusually strong versus its Windows results (27:07). In Starfield, Nvidia performance is described as “straight up screwed,” likely due to a driver or game interaction issue, with AMD cards clearly ahead (28:48, 29:25). In Cyberpunk 2077, Nvidia’s flagship card still struggles with low frametime metrics, though brute-force performance keeps it playable (33:14).

Problems encountered (39:28): The team runs into bugs, anti-cheat conflicts, and extremely long shader compilation times (4:55). They also note that tools for measuring animation error—important for their frametime analysis—are not yet where they need them to be on Linux (22:20).

Conclusion (45:45): Despite the hurdles, Linux is now performing better for gaming than ever before (5:05). The team views the effort required to benchmark Linux as worthwhile, especially in light of growing concerns over Windows’ telemetry and perceived “spyware” behavior (16:21).