Lightweight resource estimation model to extend battery life in video playback

Abstract

Modern mobile computing devices are integrating more and more processing units into a single platform. For this reason separate resource managers are needed to control computing resources effectively and in an energy efficient way. Multimedia signal processing applications, such as a video playback, have huge variations in their resource needs depending on the input sequence characteristics. To improve this situation a resource usage estimation model has been developed and evaluated. By using the model it is possible to estimate the resource needs with less than 10% error marginal on average. A proper exploitation of prior information improves both the user experience and the battery life. On the testing platform already a single step in the dynamic voltage and frequency scaling scheme corresponds to 30 minutes in high definition video playback time. The approach does not require changes to the existing encoders or decoders.