By Topic

A PCI based high-fanout AER mapper with 2 GiB RAM look-up table, 0.8 µs latency and 66MHz output event-rate

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Daniel Bernhard Fasnacht ; Institute of Neuroinformatics, University of Zurich and ETH Zurich, Switzerland ; Giacomo Indiveri

Neuromorphic systems have been increasing in size and complexity in recent years, thanks also the adoption of the Address-Event Representation (AER) as a standard for transmitting signals among chips, and building multi-chip event-based systems. AER mapper devices that route Address-Events from multiple sources to different multiple destinations are crucial components of these systems, as they allow users to flexibly compose multi-chip setups, re-configure them, and program different architecture or network topologies. In this paper we present a PCI based high-fanout AER mapper with 2 GiB RAM look-up table, 0.8 μs latency and 66 MHz output event-rate. Integrated with a PC it forms a very flexible and affordable AER experimental platform which is suitable for prototyping and research projects. Indeed, multiple instances of this system are already being used to perform various types of AER experiments. In addition, the system's ability to implement probabilistic address-event mappings further extends the range of experiments that can be performed using this platform. We describe the hardware system implementation details, compare our approach to previously proposed ones, and present experimental results which demonstrate how the system provides optimal performance for experiments with high average fanout and how for low fanout mappings the limiting factor is given by the 0.8 μs latency induced by the PC on each random memory access.

Published in:

Information Sciences and Systems (CISS), 2011 45th Annual Conference on

Date of Conference:

23-25 March 2011