Static Energy Consumption Analysis in Software: the Worst-Case Scenario

5/10/2017

By Marco Couto, HASLab, INESC TEC & Minho University.

Abstract. Energy consumption is becoming an evident concern to software developers. While programming languages provide several compiler optimizations, memory profiler tools, benchmark and time execution monitoring frameworks, there are no equivalent tools/frameworks to profile/optimize energy consumption.This is even more notorious due to the propagation of mobile devices. Such propagation of devices is also influencing software development: a software system is now developed as a set of similar products sharing common features. This talk will focus on explaining how energy consumption analysis can be performed in software in a very challenging and promising manner: statically, without execution. Moreover, a case study will be used to better explain how to achieve this, while showing the speaker's ongoing work in the area.

Keywords. Marco Couto is a 3rd year MAP-i PhD student affiliated with HASLab/INESC TEC & the University of Minho. He is currently working on energy consumption analysis in software, and working under the supervision of Prof. João Saraiva. Marco completed his MSc degree in Informatics Engineering in 2014, with a thesis entitled “Monitoring Energy Consumption in Android Applications”. Currently, his areas of research are green computing, source-code analysis and manipulation, and configurable systems. After that, he started his PhD thesis entitled “Energy-aware Software Product Lines”, with the goal of statically analyzing the source code of similar software systems, and compare the energy efficiency with each other. Since the beginning of his PhD, Marco has also been involved in 2 FCT bilateral projects, 2 FLAD/NSF projects and 1 FCT project.