Industry News

Scientists have developed "superchips" that can process nearly 2 billion images per second

Update time : 2022-06-13 09:23:21

Recently, researchers at the University of Pennsylvania published a paper saying that they have developed a powerful new optical chip, which can process nearly 2billion images per second. The device consists of a neural network that can process information into light without slowing down the components of traditional computer chips.

The chip is based on neural network, which is a system modeled on the way the brain processes information. These networks are composed of interconnected nodes like neurons, which "learn" in a way similar to the organic brain by training the data set, such as recognizing objects in images or words in speech. As time goes on, their learning ability becomes better.
During the test, the team made a chip with a size of 9.3 mm2 (0.01 in2) and used it to classify a series of handwritten characters similar to letters. After training the relevant data sets, the chip can classify the images containing two character sets, with an accuracy of 93.8%, and the classification accuracy of four types of images is 89.8%.
The most exciting thing is that the chip can classify each character in 0.57 nanoseconds, enabling it to process 1.75 billion images per second. The team said that this speed comes from the chip's ability to process information into light, which gives it a certain advantage over existing computer chips.
Another advantage is that the chip does not need to store the information being processed, so it does not have to send data to the memory card. The research team said that it is safer not to store data because it can prevent any possible leakage.The team plans to expand the chip scale and adjust the technology to handle other types of data.

Previous : Infineon is the first to use Infineon's new 750 V automotive grade discrete IGBT EDT2 devices

Next : Intel freezes hiring in PC chip unit for at least 2 weeks

AMD XC7K160T-2FFG676I - FPGAs: A Comprehensive Guide

May .16.2025

The AMD XC7K160T - 2FFG676I, a member of the Kintex - 7 FPGA family, offers a compelling combination of features. It boasts a rich array of logic resources, allowing for the implementation of intricate digital designs, making it suitable for complex tasks

Cypress/Infineon CY7C028V-20AXI Memory: A Comprehensive Guide

May .14.2025

The Cypress/Infineon CY7C028V - 20AXI is a 3.3V, 64K×16 dual - port static random - access memory (SRAM). It features a 0.35 - micron CMOS process, enabling a balance between high - speed access (20 ns) and low power consumption (120 mA typical operating

LFXP2-30E-5FN484C LATTICE -FPGA: A Comprehensive Guide

May .12.2025

The LATTICE LFXP2-30E-5FN484C is a Field-Programmable Gate Array (FPGA) based on the flexiFLASH architecture of the LatticeXP2 series. It adopts a 90nm process and has a core power supply voltage of 1.2V. With 29,000 logic units, it is rich in logic resou

Intel / Altera 5M160ZT100C5N - CPLDs: A Comprehensive Guide

May .09.2025

The Intel/Altera 5M160ZT100C5N is a 100 - pin TQFP packaged CPLD. It has 160 logic elements, rich I/O resources with 79 user I/O pins supporting various standards. It features internal oscillator, user flash memory block, and is suitable for low - power,

About Us

About Us What we do Quality People

Products

CHIP/IC Active devices Passive device Components OTHER

China electronic components supplier----Wah Luen Shing Electronic.----www.wlschip.com

Leave a message