The Energy Impact of Batch Testing in Continuous Integration
An Empirical Study of Static and Dynamic Batching Strategies
M. Oszko Mate (TU Delft - Electrical Engineering, Mathematics and Computer Science)
C.E. Brandt – Graduation committee member (TU Delft - Software Engineering)
X. Liu – Mentor (TU Delft - Software Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Continuous integration pipelines execute auto-mated tests on every commit, consuming substan-tial energy. Batch testing, which groups multiplecommits into a single test run, has been shown toreduce test executions in simulation studies, but noprior work has measured whether these reductionstranslate into actual energy savings.This study measures CPU package energy con-sumption of CI builds under a baseline run-all ap-proach and two batching strategies (BatchStop4and linear-4 lwd) across eight open-source Javaprojects. BatchStop4 achieves energy savings be-tween 57% and 88% (mean: 80.3%), while linear-4 lwd achieves savings between 59% and 94%(mean: 85.2%). Energy savings correlate almostperfectly with time savings (r >0.99), and base-line failure rate strongly predicts achievable sav-ings, while CPU utilisation shows no significant re-lationship. These findings provide empirical evi-dence that batch testing effectively reduces the en-vironmental footprint of continuous integration.