Coca4ai: checking energy behaviors on AI data centers

TL;DR

This paper presents a lightweight system for monitoring energy consumption in AI data centers, enabling better energy management and user engagement through software wattmeters validated against external measurements.

Contribution

It introduces a novel, easy-to-deploy energy monitoring approach at various scales within AI data centers, validated with accurate external wattmeters.

Findings

Effective energy monitoring at data center, user, and job levels

Potential for improving energy efficiency in AI data centers

Engages users in energy conservation efforts

Abstract

Monitoring energy behaviors in AI data centers is crucial, both to reduce their energy consumption and to raise awareness among their users which are key actors in the AI field. This paper shows a proof of concept of easy and lightweight monitoring of energy behaviors at the scale of a whole data center, a user or a job submission. Our system uses software wattmeters and we validate our setup with per node accurate external wattmeters. Results show that there is an interesting potential from the efficiency point of view, providing arguments to create user engagement thanks to energy monitoring.