• DocumentCode
    3175656
  • Title

    Address-Event Communication Using Token-Ring Mutual Exclusion

  • Author

    Imam, Nabil ; Manohar, Rajit

  • Author_Institution
    Comput. Syst. Lab., Cornell Univ., Ithaca, NY, USA
  • fYear
    2011
  • fDate
    27-29 April 2011
  • Firstpage
    99
  • Lastpage
    108
  • Abstract
    We present a novel Address-Event Representation (AER) transmitter circuit to communicate pulses of neural activity (spikes) within a neuromorphic system. AER circuits allow an ensemble of neurons to achieve large scale time-multiplexed connectivity through a shared communication channel. Our design makes use of token-ring mutual exclusion where two circulating tokens in a 2D array of neurons provide exclusive access to the shared channel. Compared to traditional arbitration-tree-based designs, our design has a higher throughput and lower latency during high spiking activity. This allows the circuit to serve larger neuronal densities and higher spiking activity while maintaining the temporal precision required by the system. Our design also eliminates the address line loads that restricted scalability in previous designs. In addition, our simpler circuit topology leads to area and power savings.
  • Keywords
    asynchronous circuits; bioelectric phenomena; network topology; neurophysiology; token networks; transmitters; address-event communication; address-event representation transmitter circuit; arbitration-tree-based designs; circuit topology; large scale time-multiplexed connectivity; neural activity; neuromorphic system; neuron 2D array; power savings; shared communication channel; spiking activity; token-ring mutual exclusion; Arrays; Encoding; Neurons; Radiation detectors; Servers; Throughput; Transmitters; AER; Address-Event Representation; Interconnection Networks; Neuromorphic Chips;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Asynchronous Circuits and Systems (ASYNC), 2011 17th IEEE International Symposium on
  • Conference_Location
    Ithaca, NY
  • ISSN
    1522-8681
  • Print_ISBN
    978-1-61284-973-7
  • Type

    conf

  • DOI
    10.1109/ASYNC.2011.20
  • Filename
    5770573