Standalone VR Gives You 11 Milliseconds a Frame. The Editor Will Not Warn You.
If you have only ever shipped Unity to desktops and phones, a standalone VR headset will humble you. The editor says 200 fps. The headset says 61, then 58, then, twenty minutes into a session on a warm forehead, something worse. I spent a good part of my career in XR before moving into AI engineering: a clinical eye-tracking app on the Vive Focus 3 at Nystag, and XR work like the ERIS platform at…
Standalone VR headsets present a strict set of performance constraints that can be challenging for developers transitioning from traditional Unity projects targeting desktop and mobile platforms. The Unity editor may display 200 frames per second (fps), but the actual headset might only manage 61, 58, or even less as the session progresses. This discrepancy becomes particularly evident during extended use, as the device's internal temperature rises, causing the chip to throttle and further degrade performance.
In the author's experience across XR projects, maintaining a consistent frame rate is paramount. The editor's frame time metrics are merely decorative, as they do not reflect the headset's real-world performance. Developers must focus on actual device testing, using tools like the Unity Profiler and chip vendor profiling tools to monitor GPU load, clocks, and temperature.
On-device overlays can provide critical information about frame times, dropped frames, and temperature indicators, all while keeping allocations minimal to avoid causing stutter.
The author recommends planning the frame budget carefully. At 90 Hz, developers should aim to stay under 11.1 ms per frame, accounting for runtime overhead. At lower frame rates, such as 72 Hz, the budget tightens to about 13.9 ms. This budget should be divided between CPU and GPU work, with the slower component determining the final frame rate.
Strategies to optimize CPU performance include reducing draw calls through single-pass instancing, customizing shaders, and implementing efficient rendering techniques like static batching, GPU instancing, and texture atlases. Garbage collection pauses can also lead to dropped frames, making it essential to avoid allocations in Update methods and optimize hot paths.
On the GPU side, developers must be selective about what they render. High-resolution displays and double buffering mean that both eyes require rendering, making efficient use of pixels crucial. Full-screen post-processing effects like bloom, depth of field, and SSAO are typically too taxing for mobile GPUs, especially in stereo.
Baking lighting and utilizing lightmaps and probes can significantly reduce real-time lighting and shadow calculations. Moreover, minimizing overdraw by carefully managing transparent elements and glass materials can further alleviate GPU load.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
