Low-Power Chips to Model a Billion Neurons

It’s a little sobering, actually. The average human brain packs a hundred billion or so neurons—connected by a quadrillion (10¹⁵) constantly changing synapses—into a space the size of a cantaloupe. It consumes a paltry 20 watts, much less than a typical incandescent lightbulb. But simulating this mess of wetware with traditional digital circuits would require a supercomputer that’s a good 1000 times as powerful as the best ones we have available today. And we’d need the output of an entire nuclear power plant to run it.

Fortunately, we don’t have to rely on traditional, power-hungry computers to get us there. Scattered around the world are at least half a dozen projects dedicated to building brain models using specialized analog circuits. Unlike the digital circuits in traditional computers, which could take weeks or even months to model a single second of brain operation, these analog circuits can model brain activity as fast as or even faster than it really occurs, and they consume a fraction of the power. But analog chips do have one serious drawback—they aren’t very programmable. The equations used to model the brain in an analog circuit are physically hardwired in a way that affects every detail of the design, right down to the placement of every analog adder and multiplier. This makes it hard to overhaul the model, something we’d have to do again and again because we still don’t know what level of biological detail we’ll need in order to mimic the way brains behave.

To help things along, my colleagues and I are building something a bit different: the first low-power, large-scale digital model of the brain. Dubbed SpiNNaker, for Spiking Neural Network Architecture, our machine looks a lot like a conventional parallel computer, but it boasts some significant changes to the way chips communicate. We expect it will let us model brain activity with speeds matching those of biological systems but with all the flexibility of a supercomputer.

Another team, led by Dharmendra Modha at IBM Almaden Research Center, in San Jose, Calif., works on supercomputer models of the cortex, the outer, information-processing layer of the brain, using simpler neuron models. In 2009, team members at IBM and Lawrence Livermore National Laboratory showed they could simulate the activity of 900 million neurons connected by 9 trillion synapses, more than are in a cat’s cortex. But as has been the case for all such models, its simulations were quite slow. The computer needed many minutes to model a second’s worth of brain activity.

One way to speed things up is by using custom-made analog circuits that directly mimic the operation of the brain. Traditional analog circuits—like the chips being developed by the BrainScaleS project at the Kirchhoff Institute for Physics, in Heidelberg, Germany—can run 10 000 times as fast as the corresponding parts of the brain. They’re also fabulously energy efficient. A digital logic circuit may need thousands of transistors to perform a multiplication, but analog circuits need only a few. When you break it down to the level of modeling the transmission of a single neural signal, these circuits consume about 0.001 percent as much energy as a supercomputer would need to perform the same task. Considering you’d need to perform that operation 10 quadrillion times a second, that translates into some significant energy savings. While a whole brain model built using today’s digital technology could easily consume more than US $10 billion a year in electricity, the power bill for a similar-scale analog system would likely come to less than $1 million.

Read more