The chip’s design, optimized for commercial applications, holds promise for integration into graphics processing units (GPUs). (Source: Image by RR) 

Breakthrough Chip Combines Speed, Energy Efficiency and Enhanced Privacy

The silicon-photonic (SiPh) chip developed by Benjamin Franklin Medal Laureate Nader Engheta and Associate Professor Firooz Aflatouni combines Engheta’s research in nanoscale material manipulation with the SiPh platform, utilizing silicon for fast, light-based computations. Based on a report from phys.org, the chip explores the potential of light-matter interaction to overcome the limitations of traditional computer chips, which have remained largely unchanged since the 1960s.

Published in Nature Photonics, the collaborative effort between Engheta and Aflatouni aimed to create a platform capable of vector-matrix multiplication, a fundamental operation in neural networks powering modern AI tools. By varying the height of silicon regions on the chip, light propagation can be controlled to perform mathematical calculations at the speed of light without additional materials.

The chip’s design, already suitable for commercial applications due to constraints from the commercial foundry, offers potential for integration into graphics processing units (GPUs). This adaptation could significantly accelerate training and classification processes, addressing the rising demand for AI systems.

In addition to speed and energy efficiency, the Engheta-Aflatouni chip provides privacy benefits by eliminating the need to store sensitive information in a computer’s memory. With computations occurring simultaneously, there is no vulnerable working memory to access, making future computers powered by this technology virtually impervious to hacking attempts.

read more at phys.org