We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Hacker News·21d ago·5 min read
Measuring Input Latency on Linux: X11 vs. Wayland, VRR, and DXVK
TL;DR
Input latency on Linux gaming setups can vary significantly between X11 and Wayland. A comprehensive testing framework was developed to measure these latencies accurately.
✦ Why It Matters
Engineers can use the provided testing framework to validate their own latency reduction strategies in Linux gaming environments.
Key Takeaways
- Build a device to measure input latency accurately.
- Use native Wayland for lower latency compared to XWayland.
- Enable Variable Refresh Rate to reduce latency fluctuations.
- Optimize dxvk settings for better frame pacing in games.
Full Summary
After switching to Linux for gaming, the author sought a reliable method to measure input latency, leading to the development of a custom device using a light sensor to track the time between mouse clicks and screen changes. Testing focused on three areas: X11 vs.
Wayland, Variable Refresh Rate (VRR), and the dxvk-low-latency frame pacer. Results indicated that while X11 consistently outperformed Wayland in latency measurements, the differences were only between 0.14 ms to 0.22 ms.
VRR improved latency by 0.26 ms to 0.45 ms and flattened latency distribution. Notably, using XWayland significantly increased latency, highlighting the importance of native Wayland for optimal performance.
Overall, the findings suggest that while X11 is slightly faster, the improvements from VRR and low-latency settings can enhance gaming experiences on Linux.