A million ARM chips to mimic one per cent of human brain

Work has started to link up one million ARM processors as part of a project to mimic the brain - but researchers say it will still represent only one per cent of the brain's capacity.

The University of Manchester has teamed up with ARM to create the "massive computer" called SpiNNaker (Spiking Neural Network architecture), a project lead by Steve Furber, the renowned ARM processor designer and now a researcher at the Manchester university.

The team has been working on developing the chips for two years, and has finally taken delivery from Taiwanese manufacturers last month, meaning work can start on creating the network, our sister title PC Pro reports.

The machine will be made up of chips featuring 18 ARM processors in a single 19mm package. "This package delivers the computing power of a PC in a tiny space and for around one watt of electrical power," the university said.

The SpiNNaker computer will mimic the signals sent in the brain by sending small packets of data. The machine will use one million ARM processors, but "this will enable them to recreate models of only up to one per cent of the human brain," the university said.

Understanding the brain

The project is looking to understand which functions are performed by different parts of the brain, to help researchers work out how best to treat injuries and diseases.

We don't know how the brain works as an information-processing system, and we do need to find out.

"Psychologists have already developed neural networks on which they can reproduce the clinical pathologies," Furber explained. "They then use these networks to test alternative therapies, to identify which is most effective in treating the patient's symptoms."

"At present they are limited in the fidelity they can achieve with these networks by the available computer power, but we hope that SpiNNaker will raise that bar a lot higher," he said. "We don't know how the brain works as an information-processing system, and we do need to find out. We hope that our machine will enable significant progress towards achieving this understanding."

The Manchester side of the project was awarded half of a total 5 million Engineering and Physical Sciences Research Council grant, with the rest going to related projects at Southampton, Cambridge and Sheffield.