According to research published in Applied Physics Reviews, the performance of photon-based TPU is 100 times higher than that of the electrical processor.
This new approach enhances not only the speed but also the efficiency of machine learning neural networks significantly. Current processors used for machine learning have certain limitations in performing complicated operations due to the power required to process the data. The more demanding the task, the more complex the data, therefore the greater the power requires. In addition, these networks are also limited by the slow transmission of electronic data between the processor and the memory.
Elite U.S. researchers discovered that using photons within the neural network (tensor) processing units (TPUs) could overcome these limitations and create more powerful and power-efficient AI. In addition, their photon-based TPU was able to perform between 2-3 orders of magnitude higher than an electric TPU.
By replacing electricity with light-based TPU, machines now can learn unsupervised at a speed of lightning. (Photo: Analytics Insight)
“We found that integrated photonic platforms that integrate efficient optical memory can obtain the same operations as a tensor processing unit, but they consume a fraction of the power and have higher throughput,” said one of the paper’s authors. “When opportunely trained, [the platforms] can be used for performing interference at the speed of light,” he added.
Potential commercial applications for the innovative processor include 5G and 6G networks, as well as data centers tasked with performing vast amounts of data processing. Photonic specialized processors can save a great deal amount of energy, improve response time, and reduce data center traffic to a significant amount.
Source: Independent UK
About us: TMA Solutions was established in 1997 to provide quality software outsourcing services to leading companies worldwide. We are one of the largest software outsourcing companies in Vietnam with 2,500 engineers.
Visit us at https://www.tmasolutions.com/
No comments:
Post a Comment