SP9 Neuromorphic Computing Platform
'Computers like brains': The Neuromorphic Computing Platform provides brain-inspired computer architectures and makes them available for applications and experiments in neuroscience and generic computing.
In SP9, we design, implement and operate the Neuromorphic Computing Platform (NCP). This Platform allows non-expert neuroscientists and engineers to perform experiments with two configurable Neuromorphic Computing Systems (the BrainScaleS and the SpiNNaker systems). These were released for use at the end of the Ramp-Up Phase of the Project and are fully integrated into the HBP Platform Infrastructure. Both systems implement simplified versions of brain models developed on the Brain Simulation Platform, and on generic circuit models. The systems comprise hardware devices that incorporate state-of-the-art electronic component and circuit technologies, as well as new knowledge arising from other areas of HBP research, such as experimental neuroscience, theory, and brain modelling.
The NCP provides Neuromorphic Computing Systems based on physical (analog or mixed-signal) emulations of brain models (the BrainScaleS system), running in accelerated mode, numerical models running in real time on digital multicore architectures (the SpiNNaker system) and the software tools necessary to design, configure and measure the performance of these systems.
Our scientists have successfully designed and produced prototypes of the next generation of computer chips that are expected to power the large-scale neuromorphic machines that are planned to be available by the end of the HBP in 2023. These have been designed in close collaboration with the HBP neuroscience groups and will have features beyond those currently accessible with conventional supercomputers, particularly plasticity and learning capabilities.
In SGA1, we intend to increase the scale, performance and online accessibility of the SpiNNaker system, with a real-time closed-loop virtual robotics environment from SP10. In addition, we intend to develop the feature set for the next-generation BrainScaleS system, and provide an architecture mode for the next-generation SpiNNaker system and test silicon, where appropriate.
SP Leader: Karlheinz MEIER
SP Co-Leader: Steve FURBER
Work Package Leaders:
- WP9.1 Platform Software and Operations: Andrew DAVISON
- WP9.2 Next Generation Physical Model Implementation: Johannes SCHEMMEL
- WP9.3 next Generation Many Core Implementation: Steve FURBER
- WP9.4 Computational Principles: Wolfgang MAASS
- WP9.5 Platform Training and Coordination: Karlheinz MEIER
Knight, J.C., & Furber, S.B. (2016). Synapse-Centric Mapping of Cortical Models to the SpiNNaker Neuromorphic Architecture. Frontiers in Neuroscience, 10: 420.
Painkras, E., Plana, L.A., Garside, J., Temple, S., Davidson, S., Pepper, J., Clark, D., Patterson, C., & Furber, S. (2012). SpiNNaker: A Multi-Core System-on-Chip for Massively-Parallel Neural Net Simulation. Custom Integrated Circuits Conference (CICC), 2012 IEEE: doi:10.1109/CICC.2012.6330636.
Furber, S. (2016). Bio-inspired massively-parallel computation. In: Advances in Parallel Computing vol. 27: Parallel Computing: On the Road to Exascale (Joubert, G.R., Leather, H., Parsons, M., Peters, F., & Sawyer, M. IOS Press, Amsterdam, pp. 3-10.
Friedmann, S., Schemmel, J., Grübl, A., Hartel, A. Hock, M. & Meier, K. (2016). Demonstrating Hybrid Learning in a Flexible Neuromorphic Hardware System. IEEE Transactions on Biomedical Circuits and Systems: 1-15.
Schemmel, J., Brüderle, D., Grübl, A., Hock, M., Meier, K., & Millner, S. (2010). A Wafer-Scale Neuromorphic Hardware System for Large-Scale Neural Modeling. Proceedings of the 2010 IEEE International Symposium on Circuits and Systems (ISCAS): 1947-1950.