Efficient parallel implementation of multilayer backpropagation networks on SpiNNaker

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

  • External authors:
  • Xin Jin
  • Mikel Luján
  • Alexander D. Rast
  • Stephen R. Welbourne


This paper presents an efficient implementation and performance analysis of mapping multilayer perceptron networks with the backpropagation learning rule on SpiNNaker - a massively parallel architecture dedicated for neural network simulation. A new algorithm called pipelined checker-boarding partitioning scheme is proposed for efficient mapping. The new mapping algorithm relies on a checker-board partitioning scheme, but the key advantage comes from introducing a pipelined mode. The six-stage pipelined mode captures the parallelism within each partition of the weight matrix, allowing the overlapping of communication and computation. Not only does the proposed mapping localize communication, but it can also hide a part of or even all the communication for high efficiency. © 2010 author/owner(s).

Bibliographical metadata

Original languageEnglish
Title of host publicationCF 2010 - Proceedings of the 2010 Computing Frontiers Conference|CF - Proc. Comput. Front. Conf.
Place of PublicationNew York, USA
PublisherAssociation for Computing Machinery
Number of pages1
ISBN (Print)9781450300445
Publication statusPublished - 2010
Event7th ACM International Conference on Computing Frontiers, CF'10 - Bertinoro
Event duration: 1 Jul 2010 → …


Conference7th ACM International Conference on Computing Frontiers, CF'10
Period1/07/10 → …