Posts by Collection

Estimation of energy consumption in machine learning

E2-Train: Training State-of-the-art CNNs with Over 80% Energy Savings

Energy-Efficient Machine Learning on the Edges

Carbontracker: Tracking and Predicting the Carbon Footprint of Training Deep Learning Models

IrEne: Interpretable Energy Prediction for Transformers

Green AI: do deep learning frameworks have different costs?

Uncovering Energy-Efficient Practices in Deep Learning Training: Preliminary Steps Towards Green AI

A systematic review of Green AI

An Energy-Aware Approach to Design Self-Adaptive AI-based Applications on the Edge

Green AI Quotient: Assessing Greenness of AI-based software and the way forward

Tensor-Aware Energy Accounting

A Synthesis of Green Architectural Tactics for ML-Enabled Systems

Specification-Driven Code Generation for Inter-parameter Dependencies in Web APIs

Authors: Saman Barakat, Enrique Barba Roque, Ana Belén Sánchez, Sergio Segura

Published in Service-Oriented Computing – ICSOC 2022 Workshops, 2022

Automatic code generation for APIs considering inter-parameter dependencies.

Recommended citation: Barakat, S., Barba-Roque, E., Sánchez, A.B., Segura, S. (2022). "Secification-Driven Code Generation for Inter-parameter Dependencies in Web APIs." Service-Oriented Computing – ICSOC 2022 Workshops.
Download Paper

Unveiling the Energy Vampires: A Methodology for Debugging Software Energy Consumption

Authors: Enrique Barba Roque, Luis Cruz, Thomas Durieux

Published in International Conference on Software Engineering, Ottawa, Canada, 2025

Energy debugging methodology for identifying and isolating energy consumption hotspots in software systems, with Alpine consuming up to 20.2% more power than Ubuntu in certain operations.

Download Paper

Energy Aware Development of Neuromorphic Implantables: From Metrics to Action

Authors: Enrique Barba Roque, Luis Cruz

Published in ICT for Sustainability Conference, Dublin, Ireland, 2025

Literature review of energy metrics for neuromorphic computing and Spiking Neural Networks (SNNs), identifying open challenges and research avenues with respect to fidelity and actionability of metrics

Download Paper

Green AI by Default: Energy Reduction Techniques for LLMs in SE

Authors: Enrique Barba Roque

Published in International Conference on Software Engineering - Doctoral Symposium, Rio de Janeiro, Brazil, 2026

Extended abstract for Green AI research, identifying open challenges and mapping out a research plan to improve metrics and optimization methods for LLMs.

Download Paper

FLOPs vs Real Work: The Importance of Replication in AI Efficiency Assessment

Authors: Enrique Barba Roque, Luis Cruz

Published in International Conference on Evaluation and Assessment in Software Engineering (EASE) 2026, Glasgow, Scotland, 2026

Replication study for a execution time estimation for CNNs. We do not manage to completely replicate the original study, due to the impact of newer GPU hardware and limited replication package

Download Paper

Finding Energy Inefficiencies in shared libraries with Docker

Published: June 12, 2024

I share my work on my Master Thesis on finding energy inefficiencies in different Linux distributions and libc implementations with the KDE Eco community, focused on improving energy efficiency of software. Minutes, Slides

Reliable cross-platform energy measurements with EnergiBridge

Authors: Carolin Brandt, Enrique Barba Roque

Published: June 10, 2026

Find our slides for this tutorial here