The chip is built using 200,000 neurons linked up by 50 million synaptic connections to mimic the way that data is transferred biologically.
While it is much less powerful than the brain at present there are no technical limitations to scaling the size of the design up.
It was developed by a team led by Karlheinz Meier, a physicist at the Heidelberg University, in Germany, as part of the Fast Analog Computing with Emergent Transient States project, or FACETS.
“Rather than simulating neurons we are building them,” he said.
The chip does not act as a brain per se, rather it simulated the parallel data handling the brain displays.
Much of the previous research into parallel data handling was done using software simulations but this requires a lot of computing power. Designing a physical chip is more efficient, and up to 100,000 times faster according to Meier.
"We can simulate a day in a second," he said.
The team now plan to scale the design up to a superchip with a total of a billion neurons and synapses numbering in the realm of 10 to the power of 13.
