The University of Pennsylvania creates a “silicon photonic” computer chip(called SiPh) capable of performing mathematical calculations using light, a faster and more efficient alternative to electricity. All this, because of the current and growing need in the field of artificial intelligence.
Today, more efficient and less power-hungry supercomputers are required. And despite current developments, computer technology continues to operate under principles dating back to the 1960s.
SiPH: Improved data processing and security
[post_relacionado]SiPh was designed to perform vector matrix multiplications, a fundamental computation in the development of neural networks. Its silicon structure, a widely available material, ensures rapid expansion of the technology.
The researchers, led by Nader Enghata, strategically reduced the chip height in specific areas, allowing precise control over light propagation. This resulted in a chip that directs light in a straight line, optimizing its performance.
With the approach to a commercial foundry for the production of SiPh chips, they are ready for implementation. It is estimated that these chips could replace the graphics processing units (GPUs) used in AI, providing a faster and more efficient alternative. Associate Professor Firooz Aflatouni suggests that the SiPh platform can complement existing infrastructure in AI companies, marking a milestone towards mass adoption of this innovative technology.
Cost-effectiveness in commercial computer chip production, privacy protection and power consumption
Its ability to perform faster and more efficient calculations is its main asset. In addition, these chips offer solutions to data privacy issues. By performing multiple calculations in parallel, they eliminate the need to store information in working memory, which strengthens the security of valuable information.
Professor Aflatouni stresses that the absence of a hackable memory reinforces the protection of information. The results of this pioneering research were published in the journal Nature Photonics, marking a significant advance in light-based computing.
Don’t miss any of our posts and follow us on social media!
Inspenet.com YouTube LinkedIn Facebook Instagram
Source: interestingengineering.com