Physicists at New York University have developed a means to create a new form of memory, making a significant step forwards in high-performance computing. The work builds on research into neuromorphic computing, which seeks to mimic the function of the human brain.
With demands for ever more powerful computing devices, researchers are pushing at the limits of physics to explore alternatives to conventional computing, such as with photonics, quantum simulators, and spintronics.
“Quantum materials hold great promise for improving the capacities of today’s computers,” said Professor Andrew Kent, a senior investigator. “The work draws upon their properties in establishing a new structure for computation.”
Kent worked alongside collaborators from the University of California-San Diego and the University of Paris-Saclay on the project. Professor Ivan Schuller, a San Diego physicist, explained: “Since conventional computing has reached its limits, new computational methods and devices are being developed. These have the potential of revolutionising computing and in ways that may one day rival the human brain.”
The physicists built on work in neuromorphic computing, a field of computing which seeks to mimic the human brain’s complex functionality. Due to the human-like features of these computers, they may offer more efficient and innovative ways to process data using approaches impossible with conventional computers.
The team built a new neuromorphic computing device: a nanoconstriction spintronic resonator. Put very simply, this manipulates known physical properties in usual ways.
Resonators can generate and store waves of well-defined frequencies; they can be imagined as the box of a string instrument. In this context, the resonator is capable of storing and processing information in a similar way to synapses and neurons in the human brain. The Scientific Reports paper describing the research details a resonator which combines the unique properties of quantum materials with that of spintronic magnetic devices.
While conventional electronic devices use the charge of an electron, spintronic devices also use an electron’s spin to process information. This could enormously boost energy efficiency while increasing storage and processing capacity.
A common device, a spin torque oscillator, operates at one set frequency. By combining it with a quantum material, however, the scientists created a device which allows for frequency tuning. This opens up the possibility of much broader and richer applications, including in neuromorphic computing.
“This is a fundamental advance that has applications in computing, particularly in neuromorphic computing, where such resonators can serve as connections among computing components,” Kent explained.
