Neurobiology-based processors provide speed boost
Tags: transportation, mobility, cyber, smart infrastructure
What? On June 6th, San Diego-based KnuEdge Inc., a private company founded by former NASA Chief Dan Goldin, launched an innovative computer processor technology designed using neurobiological principles. According to the company, its new Lambda FabricTM processor surpasses the world's top supercomputer processors in terms of scalability, latency (the time it takes for messages to travel between processors) and energy use. For example, the company claims its processor provided record fast latency while using as little as one-sixth the energy of current processors. The technology was developed to support the company's voice recognition and authentication technologies.
So what? The application of neurobiological principles to computer processor design has the potential to dramatically improve neural computing capabilities in machine learning, signal processing and the internet of things. The performance improvements achieved by the Lambda FabricTM processor are particularly critical to supporting large data centres, where energy use (and thus expensive cooling) and latency issues related to the vast number of processors have become major constraints. The need for this technology is further demonstrated by Google’s announcement to focus on the development of faster processors in order to achieve machine-learning advances beyond the reach of current processors. Integrating this technology into everyday life could aid in a diverse range of computationally hard problems, including the calculation of traffic flows.