A recent study by Marco Nett measures input latency on Linux, comparing X11 and Wayland, with Variable Refresh Rate (VRR) and DXVK low-latency fork [1]. The study uses a custom-built device to measure the time between a mouse click and the resulting screen update. The results show that X11 has lower latency than Wayland, but the difference is small, ranging from 0.14 to 0.22 ms [2]. VRR has the biggest effect, reducing latency by 0.26 to 0.45 ms [3].
The study also finds that the DXVK low-latency fork reduces latency by 0.10 to 0.29 ms in capped scenarios and by 0.84 ms in uncapped scenarios [4]. Additionally, the study concludes that XWayland should be avoided, as it adds 3.13 ms of latency [5]. Other studies have also found similar results, with David Ramiro's study finding that native Wayland is on par with native X11, while XWayland roughly doubles the latency [6]. Farnoy's study also concludes that XWayland should be avoided [7].
Overall, the study provides valuable insights into optimizing Linux for gaming and reducing input latency.
