Fault-Tolerance Mechanisms for Permanent Failures in a Coherent Shared-Memory Many-Core Architecture - Archive ouverte HAL
Poster De Conférence Année : 2014

Fault-Tolerance Mechanisms for Permanent Failures in a Coherent Shared-Memory Many-Core Architecture

Résumé

Exponential augmentation on the transistor density in nowadays integrated circuits increases probability of faults. This work proposes an "On the field" fault-tolerance strategy for many-core processors allowing self-healing against permanent faults and therefore, increasing manufacturing yield and lifetime. Fault recovery is achieved by reusing of processor cores as test generators, test analyzers and reconfiguration drivers through the execution of embedded software procedures at every processor power-on or reset.
Fichier principal
Vignette du fichier
gdr2014_cfuguet_agreiner.pdf (156.63 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01011990 , version 1 (25-06-2014)

Identifiants

  • HAL Id : hal-01011990 , version 1

Citer

César Fuguet, Alain Greiner. Fault-Tolerance Mechanisms for Permanent Failures in a Coherent Shared-Memory Many-Core Architecture. Colloque GDR SoC-SiP, Jun 2014, Paris, France. 2014. ⟨hal-01011990⟩
265 Consultations
279 Téléchargements

Partager

More