Hybrid Performance Prediction Models for Fully-Connected Neural Networks on MPSoC - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2022

Hybrid Performance Prediction Models for Fully-Connected Neural Networks on MPSoC

Résumé

Predicting the performance of Artificial Neural Networks (ANNs) on embedded multi-core platforms is tedious. Concurrent accesses to shared resources are hard to model due to congestion effects on the shared communication medium, which affect the performance of the application. In this paper we present a hybrid modeling environment to enable fast yet accurate timing prediction for fully-connected ANNs deployed on multi-core platforms. The modeling flow is based on the integration of an analytical computation time model with a communication time model which are both calibrated through measurement inside a system level simulation using SystemC. The proposed flow enables the prediction of the end-to-end latency for different mappings of several fully-connected ANNs with an average of more than 99 % accuracy.

Domaines

Electronique Autre
Fichier principal
Vignette du fichier
2022_GDR_SOC2_juin_poster_DARIOL.pdf (1.11 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03758026 , version 1 (22-08-2022)

Identifiants

  • HAL Id : hal-03758026 , version 1

Citer

Quentin Dariol, Sébastien Le Nours, Sébastien Pillement, Ralf Stemmer, Domenik Helms, et al.. Hybrid Performance Prediction Models for Fully-Connected Neural Networks on MPSoC. Colloque National du GDR SOC2, Jun 2022, Strasbourg, France. , 2022. ⟨hal-03758026⟩
53 Consultations
26 Téléchargements

Partager

Gmail Facebook X LinkedIn More