Consumption and Runtime Automatic Software Performance Optimization on Android Smartphones

Energy consumption has become a major concern in com-puting nowadays. Hardware designs are more and more energy efficient,especially in the field of battery-based devices, as smartphones, the sub-ject of study in this work. However, software ultimately drives hardwarebehavior, and it needs to be carefully designed to get the best perfor-mance. This task requires deep knowledge on the hardware architecture,and performing it by experts is unpractical for all existing hardware in-frastructures. In this work, we propose an automated methodology tomodify the source code so that the performance of the software is opti-mized, in terms of both its energy consumption and runtime. For that,a novel combinatorial multi-objective optimization problem is defined,and a micro cellular genetic algorithm is designed to address it. Resultsshow important savings of up to 78% runtime and 64% consumptionperformance on a commercial Android device.