PowerDL: A unified in-memory GPU energy profiling and visualization toolkit for deep learning
Year: 2026
Type: software-paper
Abstract: This study presents PowerDL, an in-memory software tool designed for deep learning workloads that profiles GPU energy consumption, power usage, and hardware utilization rates. PowerDL provides a unified interface for PyTorch and TensorFlow frameworks, enabling high-time-resolution analysis of GPU behavior during training and inference phases. The software supports interactive exploration of energy and performance metrics without requiring mandatory file writing. Time-series-based power traces, cumulative energy consumption, and phase-aware visualizations enable users to conduct energy-aware deep learning experiments. Additionally, optional output export mechanisms are provided to support reproducible research.
Source: SoftwareX
Author: Yasin Sancar
Institution: Atatürk University
Language: English
Subfield: Hardware and Architecture
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,200
OpenAlex ID: W7141759536