EASE 2026
Tue 9 - Fri 12 June 2026 Glasgow, United Kingdom

This program is tentative and subject to change.

Wed 10 Jun 2026 14:40 - 14:55 at JMS 743 - Performance and Optimisation 2 Chair(s): Taher A. Ghaleb

The Running Average Power Limit (RAPL) interface is widely used to estimate software energy consumption via CPU and DRAM counters, but tool design differences and high-frequency polling can introduce measurement overhead, namely, extra time and energy consumed by the tool itself.

This paper quantifies the impact of RAPL-based tools on high-frequency (1 kHz) energy monitoring and investigates mitigation strategies. We conduct two controlled experiments: the first evaluates seven tools, including a user-space application and a kernel module developed by the authors, against a no-tool baseline, using six NAS Benchmark functions to quantify overhead. The second experiment isolates and times key functions for polling Model-Specific Registers (MSRs) (rdmsr and sys/proc_read) to estimate their execution latencies and identify potential slowdowns.

The results show that existing user-space tools can introduce substantial time overhead at 1 kHz, whereas our tools significantly reduce system call overhead and inline math overhead. The time overhead of existing tools ranges from 0.25% to 46.75%. Our solutions maintain time overhead levels close to the baseline. We also find that system calls are slower than rdmsr, which in turn is slower than traditionally long-running instructions like cpuid. These findings indicate that RAPL-based energy measurement can be substantially improved by simplifying tool design and employing lower-level instructions to access RAPL values.

Our findings provide guidance for practitioners on how to develop high-frequency energy profiling tools, show possible situations that can skew energy values, and demonstrate that access to RAPL values can be faster using specific techniques.

This program is tentative and subject to change.

Wed 10 Jun

Displayed time zone: London change

13:30 - 15:00
13:30
10m
Talk
The Hidden Environmental Cost of Poor Coding Practices in TensorFlow and Keras Applications: A Study on Resource Leaks and Carbon Emissions
Short Papers and Emerging Results
Bashar Abdallah Polytechnique Montréal, Gustavo Santos Polytechnique Montréal, Rola Al Bataineh École de Technologie Supérieure ETS - Université du Québec, Alain Abran Ecole de Technologie Superieure, Mohammad Hamdaqa Polytechnique Montreal
13:40
15m
Paper
Verifier Warnings Do Not Improve Comprehensibility Prediction
Reproducibility and Negative Results
Nadeeshan De Silva William & Mary, Martin Kellogg New Jersey Institute of Technology, Oscar Chaparro William & Mary
Pre-print
13:55
15m
Talk
When Parsing Goes Wrong: An Empirical Study of Error Propagation and Data Augmentation in Log Anomaly Detection
Research Papers
Yicheng Sun City University of Hong Kong, Jacky Keung City University of Hong Kong, Xiaoxue Ma Hong Kong Metropolitan University, Yihan Liao City University of Hong Kong, Hi Kuen Yu City University of Hong Kong, Yishu Li Hong Kong Metropolitan University
14:10
15m
Talk
Decoding the Cost: A Phase-Level Analysis of LLM Inference in Software Development
Research Papers
Lola Solovyeva University of Twente, Fernando Castor University of Twente
14:25
15m
Talk
Evaluating the Environmental Impact of using SLMs and Prompt Engineering for Code Generation
Research Papers
Md Afif Al Mamun University of Calgary, Canada, Sayan Nath University of Calgary, Canada, Novarun Deb University of Calgary, Gias Uddin York University, Canada
Pre-print
14:40
15m
Talk
What Is the Cost of Energy Monitoring? An Empirical Study on the Overhead of RAPL-Based Tools
Research Papers
Jeremy Diamond Universität Zürich, Vincenzo Stoico Vrije Universiteit Amsterdam
Pre-print
Hide past events