Real-Time Cortical Simulation on Neuromorphic HardwareCitation formats

  • External authors:
  • Andrew Rowley
  • Christian Brenninkmeijer

Standard

Real-Time Cortical Simulation on Neuromorphic Hardware. / Rhodes, Oliver; Peres, Luca; Rowley, Andrew; Gait, Andrew; Plana, Luis A.; Brenninkmeijer, Christian; Furber, Steve.

In: Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, Vol. 378, No. 2164, 23.12.2019, p. 1-21.

Research output: Contribution to journalArticlepeer-review

Harvard

Rhodes, O, Peres, L, Rowley, A, Gait, A, Plana, LA, Brenninkmeijer, C & Furber, S 2019, 'Real-Time Cortical Simulation on Neuromorphic Hardware', Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, vol. 378, no. 2164, pp. 1-21. https://doi.org/10.1098/rsta.2019.0160

APA

Rhodes, O., Peres, L., Rowley, A., Gait, A., Plana, L. A., Brenninkmeijer, C., & Furber, S. (2019). Real-Time Cortical Simulation on Neuromorphic Hardware. Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 378(2164), 1-21. https://doi.org/10.1098/rsta.2019.0160

Vancouver

Rhodes O, Peres L, Rowley A, Gait A, Plana LA, Brenninkmeijer C et al. Real-Time Cortical Simulation on Neuromorphic Hardware. Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. 2019 Dec 23;378(2164):1-21. https://doi.org/10.1098/rsta.2019.0160

Author

Rhodes, Oliver ; Peres, Luca ; Rowley, Andrew ; Gait, Andrew ; Plana, Luis A. ; Brenninkmeijer, Christian ; Furber, Steve. / Real-Time Cortical Simulation on Neuromorphic Hardware. In: Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. 2019 ; Vol. 378, No. 2164. pp. 1-21.

Bibtex

@article{2027b4b4d7d44ea1bb2928e8c2a0c9c6,
title = "Real-Time Cortical Simulation on Neuromorphic Hardware",
abstract = "Real-time simulation of a large-scale biologically representative spiking neural network is presented, through the use of a heterogeneous parallelisation scheme and SpiNNaker neuromorphic hardware. A published cortical microcircuit model is used as a benchmark test case, representing 1mm2 of early sensory cortex, containing 77k neurons and 0:3 billion synapses. This is the first hard real-time simulation of this model, with 10 s of biological simulation time executed in 10 s wall-clock time. This surpassesbest published efforts on HPC neural simulators (3 slowdown) and GPUs running optimised SNN libraries (2 slowdown). Furthermore, the presented approach indicates that real-time processing can be maintained with increasing SNN size, breaking the communication barrier incurred by traditional computing machinery. Model results are compared to an established HPC simulator baseline to verify simulation correctness, comparing well across a range of statistical measures. Energy to solution, and energy per synaptic event are also reported, demonstrating that the relatively low-tech SpiNNaker processors achieve a 10 reduction in energy relative to modern HPC systems, and comparable energy consumption to modern GPUs. Finally, system robustness is demonstrated through multiple 12 h simulations of the cortical microcircuit, each simulating 12 h of biological time, and demonstrating the potential of neuromorphic hardware as a neuroscience research tool for studying complex spiking neural networks over extended time periods.",
author = "Oliver Rhodes and Luca Peres and Andrew Rowley and Andrew Gait and Plana, {Luis A.} and Christian Brenninkmeijer and Steve Furber",
year = "2019",
month = dec,
day = "23",
doi = "10.1098/rsta.2019.0160",
language = "English",
volume = "378",
pages = "1--21",
journal = "Royal Society of London. Proceedings A. Mathematical, Physical and Engineering Sciences",
issn = "1471-2946",
publisher = "Royal Society",
number = "2164",

}

RIS

TY - JOUR

T1 - Real-Time Cortical Simulation on Neuromorphic Hardware

AU - Rhodes, Oliver

AU - Peres, Luca

AU - Rowley, Andrew

AU - Gait, Andrew

AU - Plana, Luis A.

AU - Brenninkmeijer, Christian

AU - Furber, Steve

PY - 2019/12/23

Y1 - 2019/12/23

N2 - Real-time simulation of a large-scale biologically representative spiking neural network is presented, through the use of a heterogeneous parallelisation scheme and SpiNNaker neuromorphic hardware. A published cortical microcircuit model is used as a benchmark test case, representing 1mm2 of early sensory cortex, containing 77k neurons and 0:3 billion synapses. This is the first hard real-time simulation of this model, with 10 s of biological simulation time executed in 10 s wall-clock time. This surpassesbest published efforts on HPC neural simulators (3 slowdown) and GPUs running optimised SNN libraries (2 slowdown). Furthermore, the presented approach indicates that real-time processing can be maintained with increasing SNN size, breaking the communication barrier incurred by traditional computing machinery. Model results are compared to an established HPC simulator baseline to verify simulation correctness, comparing well across a range of statistical measures. Energy to solution, and energy per synaptic event are also reported, demonstrating that the relatively low-tech SpiNNaker processors achieve a 10 reduction in energy relative to modern HPC systems, and comparable energy consumption to modern GPUs. Finally, system robustness is demonstrated through multiple 12 h simulations of the cortical microcircuit, each simulating 12 h of biological time, and demonstrating the potential of neuromorphic hardware as a neuroscience research tool for studying complex spiking neural networks over extended time periods.

AB - Real-time simulation of a large-scale biologically representative spiking neural network is presented, through the use of a heterogeneous parallelisation scheme and SpiNNaker neuromorphic hardware. A published cortical microcircuit model is used as a benchmark test case, representing 1mm2 of early sensory cortex, containing 77k neurons and 0:3 billion synapses. This is the first hard real-time simulation of this model, with 10 s of biological simulation time executed in 10 s wall-clock time. This surpassesbest published efforts on HPC neural simulators (3 slowdown) and GPUs running optimised SNN libraries (2 slowdown). Furthermore, the presented approach indicates that real-time processing can be maintained with increasing SNN size, breaking the communication barrier incurred by traditional computing machinery. Model results are compared to an established HPC simulator baseline to verify simulation correctness, comparing well across a range of statistical measures. Energy to solution, and energy per synaptic event are also reported, demonstrating that the relatively low-tech SpiNNaker processors achieve a 10 reduction in energy relative to modern HPC systems, and comparable energy consumption to modern GPUs. Finally, system robustness is demonstrated through multiple 12 h simulations of the cortical microcircuit, each simulating 12 h of biological time, and demonstrating the potential of neuromorphic hardware as a neuroscience research tool for studying complex spiking neural networks over extended time periods.

U2 - 10.1098/rsta.2019.0160

DO - 10.1098/rsta.2019.0160

M3 - Article

VL - 378

SP - 1

EP - 21

JO - Royal Society of London. Proceedings A. Mathematical, Physical and Engineering Sciences

JF - Royal Society of London. Proceedings A. Mathematical, Physical and Engineering Sciences

SN - 1471-2946

IS - 2164

ER -