Skip to main navigation Skip to search Skip to main content

Evaluating the energy impact of device parameters for DNN inference on edge

  • Stony Brook University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

With advancements in edge inference accelerating the shift from cloud to edge computing, there is a need for research on sustainable edge deployments of DNN workloads. However, the energy consumption of DNN workload execution is affected by numerous parameters and knobs. This paper studies the impact of hardware knobs (CPU and GPU frequency) across six different DNN inference workloads on two different Jetson edge devices. We find that default parameter settings need not be energy optimal; for example, tuning the CPU and GPU frequency can save as much as 19% energy over DVFS.

Original languageEnglish
Title of host publicationProceedings of the 14th International Green and Sustainable Computing Conference, IGSC 2023
PublisherAssociation for Computing Machinery
Pages52-55
Number of pages4
ISBN (Electronic)9798400716690
DOIs
StatePublished - Oct 28 2023
Event14th International Green and Sustainable Computing Conference, IGSC 2023 - Toronto, Canada
Duration: Oct 28 2023Oct 29 2023

Publication series

NameACM International Conference Proceeding Series

Conference

Conference14th International Green and Sustainable Computing Conference, IGSC 2023
Country/TerritoryCanada
CityToronto
Period10/28/2310/29/23

Fingerprint

Dive into the research topics of 'Evaluating the energy impact of device parameters for DNN inference on edge'. Together they form a unique fingerprint.

Cite this