News report

NVIDIA's Wayland vs X.Org Test Finds No Clear Latency Winner

NVIDIA's XDC 2026 testing found close Wayland and X.Org input latency on an RTX 5070 setup, while VRR produced the largest responsiveness improvement.

On this page
  1. NVIDIA measured Wayland and X.Org end-to-end latency
  2. XWayland no longer showed the old extra-frame penalty
  3. VRR mattered more than the session choice in this test
  4. The result arrives as Linux desktops continue moving toward Wayland

NVIDIA measured Wayland and X.Org end-to-end latency

NVIDIA engineer Kamil Łysik presented an automated Wayland-versus-X.Org latency comparison at XDC 2026 on September 30. The experiment measured end-to-end response from a generated key press to the resulting display change, using a microcontroller and light sensor rather than relying on frame-rate or software timing alone.

According to reporting on the presentation, the tested system used a GeForce RTX 5070 with NVIDIA's R610 driver branch, KDE KWin 6.7.4 and GNOME Mutter 49.7 for Wayland testing, and X.Org Server 1.21.1.24 for the X11 path. The results were close enough that NVIDIA did not establish a consistent overall latency winner between Wayland and X.Org.

Reported NVIDIA XDC 2026 latency-test environment
ComponentReported test configurationEvidence boundary
GPUGeForce RTX 5070One NVIDIA GPU; not a cross-vendor result
NVIDIA driverR610 branchResults depend on this driver generation
Wayland compositorsKWin 6.7.4 and Mutter 49.7Does not cover every compositor or version
X11 stackX.Org Server 1.21.1.24Comparison target used in the presentation
MeasurementKey press to displayed responseEnd-to-end latency rather than FPS

XWayland no longer showed the old extra-frame penalty

One of the more useful findings for Linux gaming is that XWayland did not show the extra frame of delay associated with older configurations in NVIDIA's current test. That matters because many games and applications still use X11 interfaces even when the desktop session itself runs on Wayland.

This result should not be generalized into a claim that every XWayland game has identical latency to every native Wayland application. Game engines, compositors, presentation modes, frame caps and driver behavior can all change the complete input-to-display path. The measurement establishes what NVIDIA observed on the tested configuration.

VRR mattered more than the session choice in this test

Variable Refresh Rate produced the clearest responsiveness improvement in the reported measurements. With VRR enabled, NVIDIA found roughly comparable latency between the tested Wayland and X.Org sessions, making display timing a larger practical factor than choosing one of those session types on this setup.

For a Linux gamer deciding whether Wayland is inherently too slow, the result is therefore useful but narrow: current NVIDIA hardware and software can deliver Wayland latency close to X.Org in this measured environment. It does not prove that Wayland, X.Org or VRR will behave identically across AMD and Intel GPUs, different monitors, other compositors, or individual games.

The result arrives as Linux desktops continue moving toward Wayland

The XDC presentation is relevant because the Linux desktop ecosystem continues moving away from traditional X11 sessions while compatibility remains a practical concern for gaming. XDC itself focuses on the Linux graphics stack, including Wayland, X11, Mesa and DRM development.

The useful takeaway is not that the Wayland-versus-X11 debate is permanently settled. It is that the historical assumption that a Wayland session necessarily costs an extra frame of latency is not supported by NVIDIA's current measured setup, and XWayland's compatibility path has materially improved compared with older implementations.

Sources

Primary and technical sources

These sources support the reporting and analysis above. Current stories are updated when later evidence materially changes the facts.

  1. 01 X.Org Developer's Conference / freedesktop.org

    Wayland vs X.org latency comparison — XDC 2026
  2. 02 Phoronix

    NVIDIA's Results Of Conducting Wayland vs. X.Org Latency Testing

Related

Technical guide

16 GB vs 32 GB vs 64 GB RAM for Gaming PCs

Choose 16 GB, 32 GB, or 64 GB of system RAM for a gaming PC by measuring the games and simultaneous workloads you actually run instead of relying on a universal capacity rule.