Trading Runtime for Energy Efficiency
summary
Empirical study runs 9 CLBG benchmarks across 20 programming languages, compiling 108 solutions and measuring with Intel energy counters and power‑capping. After calibrating caps at 2, 10, 15 and 25 W the team selects a 15 W cap that maximizes savings. Every program shows lower energy and longer runs: average energy drops ~14% overall (up to ~27% for some languages) while execution time rises ~91% on average; DaCapo shows ~20.8% energy cut with ~45.6% slower runs.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:04Energy efficiency background
- 3:112017 language energy ranking
- 5:00Power capping analogy
- 6:03Power capping research
- 7:55Benchmark methodology
- 10:20Language benchmark game
- 12:07Energy measurement tools
- 13:13Power capping results
- 15:20Language ranking with power capping
- 16:25Python energy impact
- 17:29Results validation
- 20:26Faster not always greener
- 21:11Conclusion
- 22:58Q&A: Design patterns
- 25:57Q&A: Language features
- 28:45Q&A: Optimal power cap
related talks
50mEnergy Efficient Design Principles of Software Engineering
39mCreate an Eco-Friendly World with Green Software Engineering
34mCarbon Footprint–Aware Software Development
39mApplying refactoring techniques ...
42mEnergy-Efficient Software Architecture for Developers
27mKeynote: Performance tuning
01h 20mGreen Software Architecture - Dos Don'ts and Some Surprises
14mIs your Code Net zero?