A cutting-edge optical computing technology called diffraction casting, developed by researchers in Japan, aims to transform the future of computing by offering increased speed and power efficiency while reducing heat generation.
A team of researchers from the University of Tokyo has unveiled a groundbreaking optical computing technology that promises to revolutionize digital devices by significantly enhancing speed and energy efficiency. The novel design, known as diffraction casting, builds upon earlier optical computing methods and aims to address current limitations in electronic computing.
As computing applications grow increasingly complex, traditional electronic technologies are straining under the demand for higher performance and efficiency. These electronic systems often generate excessive heat, and their fabrication techniques are nearing theoretical limits. To tackle these issues, researchers have been investigating alternative methods, including optical computing, which utilizes light waves to perform calculations.
Optical computing offers the advantage of leveraging light waves’ high speed and ability to interact in intricate ways with different optical materials without generating heat. This method allows for a potentially massive parallel processing capability, making optical computing highly efficient.
In the 1980s, Japanese researchers introduced an optical computing method called shadow casting. This approach, while innovative, was limited by its bulky geometric components and lack of flexibility. Now, the University of Tokyo’s Information Photonics Lab has proposed diffraction casting, which refines shadow casting by using the properties of light waves rather than geometric forms.
“In the 1980s, researchers in Japan explored an optical computing method called shadow casting, which could perform some simple logical operations,” Ryoichi Horisaki, an associate professor in the Information Photonics Lab at the University of Tokyo, said in a news release. “But their implementation was based on relatively bulky geometric optical forms, perhaps analogous to the vacuum tubes used in early digital computers. They worked in principle, but they lacked flexibility and ease of integration to make something useful.”
Diffraction casting aims to transform computing by introducing spatially efficient, functionally flexible optical elements that hold promise for creating universal computers. Horisaki and his team have conducted numerical simulations that demonstrate the feasibility of their design using small, 16-by-16-pixel black-and-white images as input data.
The system proposed by Horisaki’s team is entirely optical until the final output stage, where it converts to electronic and digital data. This process could be particularly beneficial for image processing and other data-heavy tasks, including those in machine learning.
“Diffraction casting is just one building block in a hypothetical computer based around this principle and it might be best to think of it as an additional component rather than a full replacement of existing systems, akin to the way graphical processing units are specialized components for graphics, gaming and machine learning workloads,” lead author Ryosuke Mashiko said in the news release.
While diffraction casting holds exciting potential, its physical implementation remains a work in progress. According to Mashiko, it could take around a decade before this technology becomes commercially available. The team is optimistic that diffraction casting could also extend into the realm of quantum computing, which represents another frontier in the computing world.
The findings have been published in the journal Advanced Photonics.