Energy Consumption of Neural Networks on NVIDIA Edge Boards: an Empirical Model

TL;DR

This paper empirically measures and models the energy consumption of neural network inference on NVIDIA Jetson edge boards, providing a practical tool for optimizing energy efficiency in edge AI applications.

Contribution

It introduces a simple, realistic energy consumption model for neural network inference on NVIDIA Jetson TX2 and Xavier boards based on extensive experimental data.

Findings

Developed a practical energy consumption model for inference tasks.

Collected extensive energy data for convolutional and fully connected layers.

Model can guide architecture search, pruning, offloading, and efficiency evaluation.

Abstract

Recently, there has been a trend of shifting the execution of deep learning inference tasks toward the edge of the network, closer to the user, to reduce latency and preserve data privacy. At the same time, growing interest is being devoted to the energetic sustainability of machine learning. At the intersection of these trends, we hence find the energetic characterization of machine learning at the edge, which is attracting increasing attention. Unfortunately, calculating the energy consumption of a given neural network during inference is complicated by the heterogeneity of the possible underlying hardware implementation. In this work, we hence aim at profiling the energetic consumption of inference tasks for some modern edge nodes and deriving simple but realistic models. To this end, we performed a large number of experiments to collect the energy consumption of convolutional and fully connected layers on two well-known edge boards by NVIDIA, namely Jetson TX2 and Xavier. From the measurements, we have then distilled a simple, practical model that can provide an estimate of the energy consumption of a certain inference task on the considered boards. We believe that this model can be used in many contexts as, for instance, to guide the search for efficient architectures in Neural Architecture Search, as a heuristic in Neural Network pruning, or to find energy-efficient offloading strategies in a Split computing context, or simply to evaluate the energetic performance of Deep Neural Network architectures.