Boosting GPU Frequency for Better App Performance on Meta Quest 2

Using the full potential of Meta Quest device resources helps ensure the highest quality content and the best overall experience for users. As a result, optimizing app performance to take into account the amount of GPU available is something familiar to all Meta Quest developers.

We’re continuing to work on ways to improve developer tools and experiences, including finding ways to help you improve pixel quality and app performance. And today, we’re excited to announce that you now have more GPU compute power to work with on Meta Quest 2 devices—7% more, to be exact.

When it comes to app performance, increased GPU power should improve your ability to leverage higher pixel density without substantially reducing the resolution in order to hit the target frame rate. This means your apps’ visuals will look even better—without sacrificing resolution quality.

All developers can start taking advantage of the new 525 MHz GPU frequency (up from 490 MHz) today. No integration or coding is needed to start reaping the benefits of this improvement—the dynamic clocking system will automatically increase the frequency as it detects that your app would benefit from it; however, on the current v47 OS build, either a doff/don or sleep cycle (clicking the power button twice) is needed before the added GPU frequency is granted to the device. Starting with v49, the additional compute power will be available immediately and will not require additional steps after firing up your headset.

The increase in GPU power will also automatically integrate with dynamic foveation if you have enabled it in your app. Upon an increase in compute needs, the GPU will maintain higher visual quality as long as possible by increasing from 490 MHz to 525 MHz first, instead of increasing foveation.

We’re excited about the impact this improvement will bring across our app ecosystem, and early experiments strongly suggest that a 7% increase in GPU clocks will provide a better experience for both you and your app’s community.

Understanding how GPU memory is allocated and how much memory is being used can provide you with a foundational reference point to start optimizing app performance. In this blog post, we walk you through how GPU memory can be queried, go over tips for reducing memory usage, and highlight useful tools from Unity and Unreal that can help you analyze what you currently have in memory at any given time.

Our documentation also provides details on how you can use a variety of tools to identify bottlenecks and diagnose problems in order to optimize app performance:

Have feedback about this GPU improvement? Let us know in the Developer Forum⁠.