An efficient on-chip configuration infrastructure for a flexible multi-ASIP turbo decoder architecture - Laboratoire des Sciences et Techniques de l'Information, de la Communication et de la Connaissance, site ENIB Brest Accéder directement au contenu
Communication Dans Un Congrès Année : 2013

An efficient on-chip configuration infrastructure for a flexible multi-ASIP turbo decoder architecture

Résumé

Dynamic reconfiguration of multiprocessor platforms is an important challenge for System-on-Chip designers. Addressing this issue is mandatory in order to manage the increasing number of applications and execution conditions that multiprocessor platforms have to face. In this paper, a novel configuration infrastructure for the UDec multi-ASIP turbo decoder architecture is presented. Our approach leads to split the interconnection architecture in two subsets, one dedicated for data and another dedicated for configuration. Indeed both types of communication do not have the same requirements. Our novel configuration infrastructure, which proposes an area efficient and low latency solution, has been validated through a two-step approach. First a SystemC/VHDL mixed simulation model has been developed to perform an early performance evaluation, second a hardware FPGA prototype has been built. Results show that up to 64 processing elements can be dynamically configured in 5.352 µs.
Fichier principal
Vignette du fichier
ReCoSoC_final.pdf (570.8 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-00873978 , version 1 (09-03-2021)

Identifiants

Citer

Vianney Lapotre, Hübner Michael, Guy Gogniat, Purushotham Murugappa Velayuthan, Amer Baghdadi, et al.. An efficient on-chip configuration infrastructure for a flexible multi-ASIP turbo decoder architecture. ReCoSoC 2013 : 8th IEEE International Workshop on Reconfigurable and Communication-Centric Systems-on-Chip, Jul 2013, Darmstadt, Germany. ⟨10.1109/ReCoSoC.2013.6581518⟩. ⟨hal-00873978⟩
119 Consultations
71 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More