Embracing the unreliability of memory devices for neuromorphic computing - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Embracing the unreliability of memory devices for neuromorphic computing

Résumé

The emergence of resistive non-volatile memories opens the way to highly energy-efficient computation near-or in-memory. However, this type of computation is not compatible with conventional ECC, and has to deal with device unreliability. Inspired by the architecture of animal brains, we present a manufactured differential hybrid CMOS/RRAM memory architecture suitable for neural network implementation that functions without formal ECC. We also show that using low-energy but error-prone programming conditions only slightly reduces network accuracy.
Fichier principal
Vignette du fichier
Bocqet, IRPS2020, Embracing the Unreliability of Memory Devices for Neuromorphic Computing.pdf (1.59 Mo) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-04019222 , version 1 (08-03-2023)

Identifiants

Citer

Marc Bocquet, Tifenn Hirtzlin, Jacques-Olivier Klein, Etienne Nowak, Elisa Vianello, et al.. Embracing the unreliability of memory devices for neuromorphic computing. IRPS 2020 - IEEE International Reliability Physics Symposium, IEEE, Apr 2020, Dallas (Texas), United States. pp.1-5, ⟨10.1109/IRPS45951.2020.9128346⟩. ⟨hal-04019222⟩
23 Consultations
11 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More