An Energy Aware Linter: Helping programmers make more energy efficient code on the fly
Translated title
An Energy Aware Linter
Authors
Term
4. term
Education
Publication year
2022
Submitted on
2022-06-09
Pages
91
Abstract
Energy consumption of software has become a prominent focus area for re- searchers and developers. Much work has gone into researching patterns and constructions that consume more en- ergy than others, across programming languages and platforms. Code smells have been examined to a large extent, however not much research has gone into examining if the results carry over to other languages or platforms. In ad- dition to this, not many tools for de- tecting these issues exist. This report examines a collection of code smells that have been proven to have an im- pact in Java on desktop systems and C++ on embedded systems to see if they have the same impact in C# on a desktop system, in both a micro- and macrobenchmarks. Initial tests show that removing a number of code smells from a macro system can save up to 23% in energy consumption. To ease the process of such refactoring, an ef- fort is made to create a linter-like tool that can detect such patterns using the analysis engine CodeQL.
Keywords
Energy ; Linter ; Energy Awereness ; Power ; Strøm
Documents