Those of a certain age, namely school children in the 1980s, will remember Acorn’s BBC Micro computer, designed for the BBC’s Computer Literacy Project. One of the principal designers behind the Micro was Steve Furber, who developed the computer’s microprocessor, ARM, which today is used in an array of computer-assisted technology, most notably mobile phones. Furber notes, “The BBC micro was very much a machine of the 80s, whereas the ARM processor has just gone on from strength to strength; it has passed its 30 billion shipment. I think ARM is probably responsible for more of today’s computing power than everybody else in the history of computing put together.”
After success at Acorn he was appointed the ICL Professor of Computer Engineering at Manchester University where he has pursued advancements in computation to this day. He is also a member of the IEEE, an organization that is concerned with advancing technology for the benefit of humanity.
Copying the human brain
Furber is currently leading a project called SpiNNaker (Spiking Neural Network Architecture). The aim of the programme is to build a new type of computer that can model the human brain.
Furber says, “SpiNNaker is a coarse contraption of spiky neural architecture. It’s a massively parallel computing machine that is a result of thinking about the issue of how to model brain-like functions. This is something I’ve been thinking about since 1998 and actually building since 2005 so it is a project that has been fairly long in gestation.”
In biology, electrical impulses in neurons known as spikes send messages along the neural networks of the brain. It is these spikes that the computer is trying to replicate. Furber says, “The key idea is to build a machine able to model the very high levels of connectivity found in the brain and run in biological real time so it operates at the same speed as the biology. The central architectural concept is that to model that connectivity the system has to be able to pass very large numbers of very short messages very quickly.”
SpiNNaker uses parallel computing, which unlike traditional computing does multiple processing simultaneously, i.e. the particular problem is divided into different parts, each executing its part of the algorithm at the same time.
The difference with SpiNNaker and other parallel computers, however, is in the way the messages are routed. Furber says, “In parallel super computers the messages are individually routed so each time you send a message you have to tell it where it has got to go whereas with SpiNNaker you effectively analyse the problem you are trying to model in terms of a set of processes and a graph that connects them together. The graph is effectively mapped into hardware in terms of routing so that when a processor that’s modelling a neuron wants to issue a spike all that processor has to do is drop a packet identifying the neuron that just spiked into the communications fabric and the fabric will then do all the routing.”
A neurosurgeon’s aid
So far the project is progressing well without any significant setbacks. They have created their own bespoke silicon design prototype with a single circuit board of 864 processors.
Furber says, “That board is meant to form the basic component that we then use to scale up to bigger machines. Our ultimate target within the current project is to build a machine with a million ARM cores all concurrently running the same model. To get this into proportion however, even with a million ARM cores we are only at about 1% of the complexity of the human brain.”
The purpose of SpiNNaker is to allow neurosurgeons and psychologists to test their hypotheses on neural patterns in an effort to unravel the mysteries of the brain. Furber says, “Clearly the goal is to understand the brain. The computer is not going to do it on its own. We rely on brain specialists to give us the inputs that we need to test the machine in all sorts of useful ways.”
Interestingly, though it has primarily been designed to help us understand the brain, this new type of computing, with its ability to process a large volume of information quickly and simultaneously, could be incorporated into conventional computers of the future. Once again, it is a perfect example of technology inspired by nature.