Design of a concentrated torus topology with channel buffers and efficient crossbars in NoCs

Dominic Ditomaso, Randy Morris, Evan Jolley, Ashwini Sarathy, Ahmed Louri, Avinash Kodi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Excess power dissipation along with increased leakage currents in router buffers and crossbars are becoming a major constraint that is affecting the performance of Network-on-Chips (NoCs) architectures. In this paper, we design channel buffers and router crossbars in a concentrated torus topology (CTorus) which is a dual network without the additional area overhead. When compared to other dual networks, CTorus improves saturation throughput by 11-20% for synthetic traffic and improves speedup by 1.78-2.15X for real benchmark traces such as PARSEC and SPEC CPU2006. When the energy-efficient buffer and crossbar organization was inserted into our CTorus topology, we reduced energy dissipation by 32% and area by 53% on average over mesh2X, CMesh2X and FBfly2X.

Original languageEnglish (US)
Title of host publicationProceedings - IEEE 27th International Parallel and Distributed Processing Symposium Workshops and PhD Forum, IPDPSW 2013
PublisherIEEE Computer Society
Pages876-883
Number of pages8
ISBN (Print)9780769549798
DOIs
StatePublished - 2013
Externally publishedYes
Event2013 IEEE 37th Annual Computer Software and Applications Conference, COMPSAC 2013 - Boston, MA, Japan
Duration: Jul 22 2013Jul 26 2013

Publication series

NameProceedings - IEEE 27th International Parallel and Distributed Processing Symposium Workshops and PhD Forum, IPDPSW 2013

Conference

Conference2013 IEEE 37th Annual Computer Software and Applications Conference, COMPSAC 2013
Country/TerritoryJapan
CityBoston, MA
Period7/22/137/26/13

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Software
  • Theoretical Computer Science

Fingerprint

Dive into the research topics of 'Design of a concentrated torus topology with channel buffers and efficient crossbars in NoCs'. Together they form a unique fingerprint.

Cite this