Home » today » Business » Lightning: The Future of Computing – MIT’s Photonic Computing Power System Accelerating Machine Learning

Lightning: The Future of Computing – MIT’s Photonic Computing Power System Accelerating Machine Learning

Physics is against Moore’s law, but there is a way out. “Lightning” usheres in the era of future computing

MIT introduces “Lightning,” a photonic computing power system for accelerating machine learning.

Computing technologies are moving to a new stage. Moore’s claim that the number of transistors on a chip will double every year is no longer valid due to physical limitations. This pushes engineers to look for new ways to improve computing capabilities.

Photonic computing: a new stage in IT

One answer to the growing needs of machine learning is photonic computing. Instead of transistors and wires, photons are used to perform computational operations. With lasers, these small beams of energy travel at the speed of light, complementing existing computer systems.

Scientists from the Massachusetts Institute of Technology (MIT) are one of the leading technical schools and research centers in the United States and the world. It was founded in 1861 and is located in Cambridge, Massachusetts.

MIT is known for its focus on research and engineering in various fields, including computer science, artificial intelligence, energy, biotechnology, physics, chemistry and others. The university offers bachelor’s, master’s and doctoral programs.

MIT is renowned for its academic rigor, innovative approach to teaching, and strong emphasis on practical application of knowledge. At MIT, students and researchers have access to state-of-the-art labs, equipment, and resources to work on cutting-edge projects.

” data-html=”true” data-original-title=”MIT”>MIT demonstrated the potential of photonics in the field of machine learning with their new “Lightning” system . This prototype combines the advantages of photonic and electronic components, delivering amazing speed and becoming the first system to answer machine learning queries in real time.

Overcoming the Limitations of Photonics

The main problem was that photonic devices are passive and have no memory or instructions to manipulate data. However, “Lightning” eliminates this drawback. MIT’s Zhizhen Zhong emphasized that managing the flow of data between photonic and electronic components has been the Achilles heel of previous designs.

The Lightning system connects photonic and electronic circuits, providing a single platform for data processing. It provides fast real computing process without stopping.

Eco-friendly solution in IT

Modern machine learning services such as ChatGPT are computationally intensive. “Lightning” offers a faster and more energy efficient method. The Ghobadi team compared their device to standard GPUs and other accelerators, finding that “Lightning” was significantly more efficient in terms of power consumption.

This approach not only speeds up response times to machine learning queries, but also offers a way to reduce the environmental footprint of these centers.

2023-09-16 09:00:41
#Physics #Moores #law #Lightning #usheres #era #future #computing

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.