An Empirical Study on Method-Level Performance Evolution in Open-Source Java Projects
Kaveh Shahedi, Nana Gyambrah, Heng Li, Maxime Lamothe, Foutse Khomh · 2025
Read An Empirical Study on Method-Level Performance Evolution in Open-Source Java Projects by Kaveh Shahedi, Nana Gyambrah, Heng Li, Maxime Lamothe, Foutse Khomh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-07084.txt.
Description
Kaveh Shahedi and co-authors present a 2025 empirical study of performance change in Java projects. Developers guess which code edits slow a program down, but method-level proof was scarce. The team mines fifteen mature open-source projects on GitHub. They isolate 739 commits holding 1,499 method-level changes. A microbenchmark harness measures each method's speed precisely. Bytecode instrumentation captures the execution metrics behind the numbers.
Statistical tests then grade which change types truly move performance, and by how much. Four angles structure the analysis, from time trends to change categories. Software teams get evidence where folklore ruled code review. Performance regressions slip in through innocent-looking edits; knowing the risky edit types focuses review effort. The fine grain is the contribution: project-level speed data hide the very methods that changed.
Catalog record
| CREATOR | Kaveh Shahedi, Nana Gyambrah, Heng Li, Maxime Lamothe, Foutse Khomh |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 141 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-07084