Energy-Efficient Code by Refactoring

Abstract

The rising number of mobile devices increase the interest in longer battery durations. To increase battery duration, researchers try to improve e.g. different hardware components, such as processors and GPS for lower energy consumption. Frequently, software optimization possibilities to save energy are forgotten. Hence, an approach is shown to reduce energy consumption of applications by reengineering. Therefor, energy-wasteful code in applications is searched by code analysis and then restructured to optimize their energy consumption. Energy savings are validated by different energy measurements techniques. MotivationThe scope of mobile devices is increasing, due to new user requirements, more powerful processors, and diversity of applications. Mobile devices are used for many tasks, and thus, users carry their smartphone with them the whole day. Hence, batteries of these devices need to endure this time. The Blackberry Z10 shows the contrary [1]: First tests showed that batteries of Z10s are empty after about five hours. These are only 21 % of a day which does not match users expectation's of at least one day uptime. Blackberry's solution is an external battery which loads the main battery to lengthen its operating time. This is not acceptable for most users to use two batteries. Therefore, manufacturers should be interested in energyefficient software to extend devices' battery duration.To improve batteries' duration by software optimization, software evolution techniques are used: Reverse Engineering produces an abstract view of components and relationships inside on applications and provides to analyze for energy wasting software behavior [2]. Reengineering is applied to improve existing applications regarding their energy consumption, in which the intended application behavior is not changed. Therefor, code analyzing and restructuring are used to obtain more energy-efficient code. Both steps describe refactoring which is concerned with detecting and restructuring inefficient code (code smells) without changing its behavior [3]. This paper is structured as follows: First, some energy code smells which have been validated to be energy-wasteful by measurement are presented in Section 2. Next, the process of identifying and restructuring of energy code smells is described in Section 3. Section 4 shows a validity of the energy consumption of applications before and after reengineering. Section 5 concludes this paper with an outlook. Energy RefactoringEnergy refactorings contain energy code smells which are energy-wasteful parts of code. These energy code smells and their restructurings are defined and described similar as the code smells of Fowler [3]. Two such energy code smells are Binding ressources too early and Third party advertising [4].

Evaluation of Strategies for the Development of Efficient Code for Raspberry Pi Devices

“…In addition, recent proposals have focused on code refactoring techniques [ 29 , 30 , 31 ] and restructuring codes without changing external behavior, but they only cover a limited number of techniques, and do not contain enough detail to help developers in writing more efficient code.…”

SENSEI: Software evolution service integration

“…To support the study of different research questions in this area (cf. [28], [29], [30]), a suitable tool infrastructure has to be built, combining tools for static and dynamic code analysis, metrics evaluation, refactoring, and visualization, as well as "helper" tools, e.g. for parsing, unparsing, or data transformation.…”

Tales from the Code #1: The Effective Impact of Code Refactorings on Software Energy Consumption

“…Other works also considered energy efficient refactoring for mobile applications (Gottschalk et al, 2013). In particular, the authors of (Rodriguez, 2017) presented some early experiments on different micro-benchmarks and discussed many coding aspects with a focus on implementation techniques, such as how to iterate on a matrix, avoid operations with immutable data types, evaluating strings, or the use the more specific numeric data types to save battery life.…”

“…Yet, code refactorings have also been considered as a mean to improve the performance and/or energy efficiency in a more or less automated way (Gottschalk et al, 2013;Anwar et al, 2019;Morales et al, 2018;Bree and Cinnéide, 2020). The large majority of the literature that has been published in this domain-especially for mobile application (Palomba et al, 2019;Gottschalk et al, 2013;Anwar et al, 2019;Linares-Vásquez et al, 2014)-based their study on a predefined set of refactoring rules, design patterns, or code smells. In most of these studies, the authors measure and analyze the effect of atomic code changes on the total energy efficiency of the software under study, before concluding on their effect.…”