Programming a million-core machine

Date/Time:
Wednesday 27 November 2013, 6.00pm - 8.00pm

Venue:
Manchester Metropolitan University Business School, All Saints Campus, Oxford Road, Manchester, M15 6BH | Venue and travel information

Speaker:
Steve Furber, CBE FRS FREng FBCS FIET

Summary:

This is a joint event with Manchester Branch and a special invitation is extended to all APSG members in the Northwest.

The SpiNNaker (Spiking Neural Network Architecture) machine is a massively-parallel computer whose architecture is inspired by the challenge of running brain models in biological real time. The hardware incorporates 864 ARM mobile phone processors on a circuit board; 1,200 of these boards will ultimately form a million-core machine capable of executing neural circuit models of up to 1% of the complexity of the human brain. Programming SpiNNaker demands a formulation of the problem into a very large number of small communicating processes (much like a neural network!), and while we have programming flows that work at the level of an individual board, many challenges remain in scaling to the million-core machine.

About the Speaker:

Steve Furber CBE FRS FREng FBCS FIET is ICL Professor of Computer Engineering in the School of Computer Science at the University of Manchester, UK. He obtained a BA in Maths in 1974 and a PhD in Aerodynamics in 1980, both from the University of Cambridge, and was Rolls Royce Research Fellow at Emmanuel College, Cambridge, from 1978-81. He then joined Acorn Computers where he was a principal designer of the BBC Micro, which introduce computing into most UK schools, and the ARM microprocessor, which today powers most of the world's consumer electronics, over 30 billion having been shipped by ARM Ltd's semiconductor partners. At Manchester his research interests include Systems-on-Chip, low-power and asynchronous logic design, and neural systems engineering.

Slides:

View the slides for this event (PDF)

View the slides and YouTube video from the event earlier this year

Read detailed notes on this talk