Understand how scalers dynamically adjust asset quality levels to optimize performance based on system conditions.
For practical examples of scaler use, refer to the Adaptive Performance samples.
Scalers are components that represent individual features, such as graphics or physics, and control the quality of these features in response to real-time system conditions.
The Indexer captures these system conditions and tracks thermal and performance state to produce a quantified quality index.
Each scaler has default settings that specify the minimum scale, maximum scale, and maximum scale level of the associated feature. As the scale level increases, the associated feature’s quality (as measured by level of detail) decreases. The scale step is calculated as (highest scale - lowest scale) / maximum scale level. The default scale level always starts at zero.
Scalers respond to priorities that the Indexer supplies, using the following targets:
Scalers only work when the Indexer is active. When you enable scalers, they’re automatically added to your application.
To enable scalers:
Adaptive Performance provides a set of built-in scalers with default settings.
You can also create custom scalers to control additional features in your application.
Each scaler stores its tuneable settings separately for each operation mode. For example, a scaler can use one set of bounds in normal mode and a different set in battery mode. Adaptive Performance applies the settings for the active operation mode, so switching modes changes how each scaler behaves. You configure these per-mode settings in the provider settings.
Once you have several scalers active, managing their settings individually can be complex. To simplify this, you can group scaler settings into scaler profiles.
A profile lets you create and change between different performance configurations (like high quality or power saving) with a single command.
To create and load scaler profiles at runtime, refer to Use scaler profiles.