SpiNNaker Support for the Neural Simulator NEST
SpiNNaker (Spiking Neural Network Architecture) is a novel computer architecture inspired by the working of the human brain and has been designed by the Advanced Processor Technologies (APT) Research Group at the School of Computer Science, University of Manchester. A SpiNNaker machine is a massively parallel computing platform, targeted towards three main areas of research: neuroscience, robotics and computer science.
The neural simulator NEST is designed for the simulation of large heterogeneous spiking neural networks of point neurons. NEST provides tools for modern computational neuroscience and scales on a wide range of machines from laptops to the world’s largest supercomputers.
The aim of the project is to add the SpiNNaker hardware to the platforms supported by NEST. Exact reproducibility and correctness of simulation results are important. Since the NEST simulator serves as a reference platform for hardware systems, the integration of SpiNNaker will not only extend the number of targeted platforms, but will also allow the direct validation of simulation data produced by SpiNNaker applications. In addition SpiNNaker users will benefit from the NEST functionality, especially when creating point neuron models in NESTML.
Figure: SpiNNaker board (from APT Group, University of Manchester)
Our Contribution
- Extending NEST and NESTML functionality for the generation of SpiNNaker-compliant neuron models in C code.
- Development of a framework for SpiNNaker binary code creation and deployment.
- Development of reference examples for the validation of simulation results.
Simlab Contact
Project Partners
Advanced Processor Technologies Research Group, University of Manchester