Android Display System Application Hardware Acceleration Technology

As an Embedded Multimedia Operating System, Android has well transplantation, powerful multimedia features, and the convenient software development process, so it is applied in the whole mobile platform area at present. In order to achieve the higher requirement introduced by the rapid development of the electronics, high performance embedded hardware graphic accelerators are widely and deeply accepted and adapted by Android gradually.However, because of the exclusive purpose of embedded system, hardware acceleration for embedded system has no unified standard, and for the lack of the specification about how to effectively use the hardware in practice, Android's design framework fails to efficiently explore the underlying hardware resources. All of those result in that Android could not gain the best performance.This paper focuses on how to apply hardware graphic accelerator for Android on MPRC Unicore platform. The key point is the design of the hardware acceleration layer through which Android can efficiently explore the benefit brought from the underlying hardware. The work of this paper includes these:1. The design and implementation of hardware acceleration layer based on UniGFX acceleration chip. We implement the interface that to operate UniGFX not only to shield the details of hardware acceleration, but also to reduce the complexity of programming on UniGFX.2. The optimization strategy for Android Display System based on hardware acceleration layer. The solution reduces the overhead of cross-layer function call, and the data swapping between main memory and video memory by utilizing the hardware channel of UniGFX.Meanwhile we design the evaluation to test the performance of Android Display System. The result shows that the graphic performance has been greatly improved. It indicates that when operating mouse in common cases, the occupancy rate of CPU reduces about 88% while keeping the friendly user experience, and the speed of frame rendering in Android Camera is increased by 1.5-6 times. All the result indicates the effectiveness of the optimization scheme, and shows the value and great potential of hardware accelerators for Android system.