Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group
  • Asia Communications and Photonics Conference (ACP) 2018
  • OSA Technical Digest (Optica Publishing Group, 2018),
  • paper Su2A.38

Increasing Communication Path Diversity in Mesh-Based ONoC by Using Multi-Interface Processing Cores

Not Accessible

Your library or personal account may give you access

Abstract

Communication path diversity has been greatly limited in most of existing Optical Network-on-Chip (ONoC) architectures, because one processing core is connected to only one optical router. This paper considers multi-interface processing core to increase the path diversity in ONoC. According to mesh architecture, we propose communication protocols that suits the multi-interface ONoC, including routing strategy, and the switching mechanism. Besides, a network interface and an 8 × 8 optical router architecture are designed to support the multiple interfaces and the protocols. We analyze the performance per area of the multi-interface ONoC architecture. The simulation results show that multi-interface ONoC reduces 66.7% in end-to-end delay per area and increases by 1.7× in maximum throughput per area, compared with the single-interface ONoC architecture under hot-spot traffic pattern.

© 2018 The Author(s)

PDF Article
More Like This
A Nesting Ring Optical Network on Chip (ONoC) Architecture for Multi-chip Systems

Wenzhe Li, Shanguo Huang, Yu Zhou, Shan Yin, Jie Zhang, and Wanyi Gu
ASu1H.1 Asia Communications and Photonics Conference (ACP) 2015

A High-Performance Optical Network-on-chip Architecture Based on Sub-network Division

Ke Chen, Huaxi Gu, Zheng Chen, and Na zhang
AW3H.5 Asia Communications and Photonics Conference (ACP) 2013

Impact of multi-port card diversity constraints in mesh optical networks1

Eric Bouillet, Jean-François Labourdette, and Subir Biswas
NThB5 National Fiber Optic Engineers Conference (NFOEC) 2006

Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.