The impact of traffic localisation on the performance of NoCs for very large manycore systems

Al Khanjari, S. and Vanderbauwhede, W. (2015) The impact of traffic localisation on the performance of NoCs for very large manycore systems. Procedia Computer Science, 56, pp. 403-408. (doi: 10.1016/j.procs.2015.07.227)

Abstract

The scaling of semiconductor technologies is leading to processors with increasing numbers of cores. The adoption of Networks-on-Chip (NoC) in manycore systems requires a shift in focus from computation to communication, as communication is fast becoming the dominant factor in processor performance. In large manycore systems, performance is predicated on the locality of communication. In this work, we investigate the performance of three NoC topologies for systems with thousands of processor cores under two types of localised traffic. We present latency and throughput results comparing fat quadtree, concentrated mesh and mesh topologies under different degrees of localisation. Our results, based on the ITRS physical data for 2023, show that the type and degree of localisation of traffic significantly affects the NoC performance, and that scale-invariant topologies perform worse than flat topologies.

• Item Type: Articles
• Additional Information: The 10th International Conference on Future Networks and Communications (FNC 2015) / The 12th International Conference on Mobile Systems and Pervasive Computing (MobiSPC 2015) Affiliated Workshops
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Vanderbauwhede, Professor Wim
• Authors: Al Khanjari, S., and Vanderbauwhede, W.
• College/School: College of Science and Engineering > School of Computing Science
• Journal Name: Procedia Computer Science
• Publisher: Elsevier B.V.
• ISSN: 1877-0509
• ISSN (Online): 1877-0509
• Copyright Holders: Copyright © 2015 The Authors
• First Published: First published in Procedia Computer Science 56:403-408
• Publisher Policy: Reproduced under a Creative Commons License