Nanomagnetic Computers Are the Ultimate in Efficiency

Nanomagnetic Computers Are the Ultimate in Efficiency: "

Computers that run on chips made from tiny magnets may be as energy-efficient as physics permits.

According to new calculations, if nanomagnetic computers used any less energy, they’d break the second law of thermodynamics. Such computers are still semi-theoretical, but they could someday be used in the deep oceans or even deep space, where energy is at a premium.

If nothing else, nanomagnetic laptops wouldn’t overheat.

“They’re actually maximally efficient, in the sense that they use up only the energy that is theoretically required to carry out a computation,” said electrical engineer Brian Lambson of the University of California at Berkeley. The results will be published in Physical Review Letters.

Conventional computers process information by shuttling electrons around circuits. But though electrons have miniscule mass, it takes a surprising amount of energy to move them. Even the most advanced computers use far more energy than they theoretically need.

That theoretical energy limit was set by IBM physicist Rolf Landauer, who argued in 1961 that altering a single bit of information will always produce a tiny amount of heat. No matter how the computer is built, Landauer claimed, no change can occur without an accompanying transfer of energy. Most computers devour up to a million times more energy than this “Landauer limit” every time they do a calculation.

Nanomagnetic chips are made from material similar to refrigerator magnets, etched with rows of rectangles. Each rectangle measures about 100 nanometers on a side and has magnetic poles. Information is stored in how they point: One configuration is 1, the other is 0. Because the magnets are so small, they can be packed close enough for their magnetic fields to interact. Information passes without any physical changes to the chip.

“Magnetic systems are unique in that they have no moving parts,” Lambson said. “Moving parts are really what dissipate a lot of energy in physical systems, whether it’s moving electrons or physical material.”

Nanomagnetic chip design is still in its infancy, far from optimally efficient. But to see how little energy the chips might consume, Lambson’s team estimated how magnetic fields would change during computation, then calculated the energy required to make those changes.

The results were close to Landauer’s limit. “We were surprised to see that they were almost exactly the same,” Lambson said.

Using magnets to build ultra-efficient computers is a powerful idea, said nanomagnetic logic pioneer Wolfgang Porod of the University of Notre Dame, who was not involved in the new work.

The Landauer limit, however, may not actually represent the limits of efficiency. “These arguments still are somewhat controversial,” Porod said. “The argument used to be more academic. But with devices getting smaller and smaller, these arguments are hitting closer to home.”

Image: A schematic of how nanomagnetic computers store and work with information using the direction of magnetic north. Courtesy of Brian Lambson.

Citation: “Exploring the Thermodynamic Limits of Computation in Integrated Systems: Magnetic Memory, Nanomagnetic Logic and the Landauer Limit.” Brian Lambson, David Carlton and Jeffrey Bokor. Physical Review Letters, in press.

See Also: