The Study of Android Parallel Programming Based on the Dual-Core Cortex-A9

Abstract

This study is based on the Samsung Exynos 4210 dual-core Cortex-A9 and Android 4.2.1. The performances of the APPs are improved by tuning CPUs' resources allocation and adding parallelism using the OpenMP compiler directives. The Cgroup and Cpuset are used in this paper to manage the CPUs' resources allocation. Besides, the Android's Native Development Kit is modified to support the Android Apps with OpenMP library in this paper. The study takes the Canny edge detection of the OpenCV as an example. The result shows that the processing time of the one picture can be improved from 939ms to 671ms with 28.5% enhancement.