Usability and Performance Analysis of Embedded Development Environment for On-device Learning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Scaffi, Enzo, Bonneau, Antoine, Mouël, Frédéric Le, Mieyeville, Fabien
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909166948319232
author Scaffi, Enzo
Bonneau, Antoine
Mouël, Frédéric Le
Mieyeville, Fabien
author_facet Scaffi, Enzo
Bonneau, Antoine
Mouël, Frédéric Le
Mieyeville, Fabien
contents This research empirically examines embedded development tools viable for on-device TinyML implementation. The research evaluates various development tools with various abstraction levels on resource-constrained IoT devices, from basic hardware manipulation to deployment of minimalistic ML training. The analysis encompasses memory usage, energy consumption, and performance metrics during model training and inference and usability of the different solutions. Arduino Framework offers ease of implementation but with increased energy consumption compared to the native option, while RIOT OS exhibits efficient energy consumption despite higher memory utilization with equivalent ease of use. The absence of certain critical functionalities like DVFS directly integrated into the OS highlights limitations for fine hardware control.
format Preprint
id arxiv_https___arxiv_org_abs_2404_07948
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Usability and Performance Analysis of Embedded Development Environment for On-device Learning
Scaffi, Enzo
Bonneau, Antoine
Mouël, Frédéric Le
Mieyeville, Fabien
Software Engineering
Artificial Intelligence
Machine Learning
This research empirically examines embedded development tools viable for on-device TinyML implementation. The research evaluates various development tools with various abstraction levels on resource-constrained IoT devices, from basic hardware manipulation to deployment of minimalistic ML training. The analysis encompasses memory usage, energy consumption, and performance metrics during model training and inference and usability of the different solutions. Arduino Framework offers ease of implementation but with increased energy consumption compared to the native option, while RIOT OS exhibits efficient energy consumption despite higher memory utilization with equivalent ease of use. The absence of certain critical functionalities like DVFS directly integrated into the OS highlights limitations for fine hardware control.
title Usability and Performance Analysis of Embedded Development Environment for On-device Learning
topic Software Engineering
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2404.07948