BrainScaleS Large Scale Spike Communication using Extoll

Tobias Thommes; Niels Buwen; Andreas Grübl; Eric Müller; Ulrich Brüning; Johannes Schemmel

BrainScaleS Large Scale Spike Communication using Extoll

Hardware Architecture 2021-12-15 v2 Neural and Evolutionary Computing Networking and Internet Architecture

Authors: Tobias Thommes , Niels Buwen , Andreas Grübl , Eric Müller , Ulrich Brüning , Johannes Schemmel

Abstract

The BrainScaleS Neuromorphic Computing System is currently connected to a compute cluster via Gigabit-Ethernet network technology. This is convenient for the currently used experiment mode, where neuronal networks cover at most one wafer module. When modelling networks of larger size, as for example a full sized cortical microcircuit model, one has to think about connecting neurons across wafer modules to larger networks. This can be done, using the Extoll networking technology, which provides high bandwidth and low latencies, as well as a low overhead packet protocol format.

Keywords

neuromorphic computing

Cite

@article{arxiv.2111.15296,
  title  = {BrainScaleS Large Scale Spike Communication using Extoll},
  author = {Tobias Thommes and Niels Buwen and Andreas Grübl and Eric Müller and Ulrich Brüning and Johannes Schemmel},
  journal= {arXiv preprint arXiv:2111.15296},
  year   = {2021}
}

Comments

3 pages, 2 figures, submitted to the Neuro Inspired Computational Elements 2020 (NICE'2020) conference, accepted and presented as a poster in March 2021; 1st replacement: add acknowledgement of DFG (German Research Foundation)

BrainScaleS Large Scale Spike Communication using Extoll

Abstract

Keywords

Cite

Comments

Related papers