Computing at the speed of light? How to achieve faster and more energy-efficient memories and processors
Thanks to the rise of Big Data and AI applications, there’s an ever increasing demand for information processing and memory storage technologies with higher speed, lower energy consumption and smaller size. To address these issues, scientists have been focusing on creating the next generation of computing systems that store and process data in the same place, with a focus on the use of light. Enter the EU-funded Fun-COMP project that is developing nanoelectronic and nanophotonic devices and systems that blend the central information processing tasks of computing and memory. With partial funding from the project, a team of researchers has recently developed an electro-optical, integrated nanoscale device that is programmable with either photons or electrons. “This provides an elegant solution to achieving faster and more energy efficient computer memories and processors,” as stated in a news release by project coordinator the University of Exeter. “Computing at the speed of light is an enticing prospect, and with this development it’s now within our grasp. While the use of light to perform various computing processes has previously been demonstrated, a compact device to interface with the electronic architecture of traditional computers has so far been lacking.”
Incompatibility
The same news release explains that “the fundamentally different interaction volumes for electrons and photons – the wavelength of light being much larger than that of electrons” is the main reason for the incompatibility between electrical and light-based computing. The team provided a solution to this problem and “combined concepts from integrated photonics, plasmonics and electronic memory technologies to deliver a compact device that can operate simultaneously as an optical or electrical memory, and as a processor.” The news release adds: “Information can be stored and processed using either light or electrical signals, or indeed by any combination of the two.” The researchers published their study in the journal ‘Science Advances’. “This is an unprecedented demonstration of an integrated, reversible, and nonvolatile phase-change memory cell that fully bridges the gap between electro-optic mixed-mode operations.” They conclude: “We anticipate that a plethora of novel devices and platforms should arise in the coming years, which will capitalize on the bridge between the electrical and photonic domains that are demonstrated here.” The ongoing Fun-COMP (Functionally scaled computing technology: From novel devices to non-von Neumann architectures and algorithms for a connected intelligent world) project seeks “to develop a new wave of industry-relevant technologies that will extend the limits facing mainstream processing and storage approaches,” as noted on the project website. In May 2019, another study that is partially funded by Fun-COMP focused on designing hardware mimicking neurons and artificial synapses – connections between the neurons that can store and process information in ways similar to that of the human brain. “Such hardware, when connected in networks or neuromorphic systems, processes information in a way more analogous to brains,” the researchers say in an article in the journal ‘Nature’. “Here we present an all-optical version of such a neurosynaptic system, capable of supervised and unsupervised learning.” For more information, please see: Fun-COMP project website
Countries
United Kingdom