Floadia Develops Memory Technology That Retains Ultra-high-precision Analog Data for Extended Periods

TOKYO, Dec. 10, 2021 /PRNewswire/ --

- This Technology, Set for Presentation at SEMICON JAPAN at Tokyo Big Sight, Is to Be Applied to Computing-in-Memory Chips That Can Perform AI Inference Operations at Overwhelmingly Low Power -

Floadia Corporation, headquartered in Kodaira-shi, Tokyo, has developed a prototype 7-bit-per-cell flash memory chip that can retain analog data for 10 years at 150 degrees Celsius by devising a memory cell structure and control method. With the existing memory cell structure, the problem of characteristic change and variation due to charge leakage was significant, and the data retention was only about 100 seconds.

Logo: https://kyodonewsprwire.jp/prwfile/release/M104461/202112064501/_prw_PI1fl_5N25P74m.jpg

Floadia will apply the memory technology to a chip that realizes AI (artificial intelligence) inference operations with overwhelmingly low power consumption. This chip is based on an architecture called Computing in Memory (CiM), which stores neural network weights in non-volatile memory and executes a large number of multiply-accumulate calculations in parallel by passing current through the memory array. CiM is attracting worldwide attention as an AI accelerator for edge computing environments because it can read a large amount of data from memory and consumes much less power than conventional AI accelerators that perform multiply-accumulate calculations on CPUs and GPUs.

This memory technology is based on SONOS-type flash memory chips developed by Floadia for integration into microcontrollers and other devices. Floadia made numerous innovations such as optimizing the structure of charge-trapping layers, i.e. ONO film, to extend the data retention time when storing 7 bits of data. The combination of two cells can store up to 8 bits of neural network weights, and despite its small chip area, it can achieve a multiply-accumulate calculation performance of 300 TOPS/W, far exceeding that of existing AI accelerators.

The content of this technology will be exhibited at a booth (booth number 1746) of SEMICON JAPAN to be held at Tokyo Big Sight from December 15 to 17, 2021, and Floadia's CTO, Yasuhiro Taniguchi will give a lecture at "TechSTAGE" from 11:15 a.m. (JST) on December 15.

About Floadia Corporation
The company was established in 2011 by a group of experienced engineers who had been developing embedded non-volatile memory chips for more than 20 years at Hitachi, Ltd. and Renesas Technology (now Renesas Electronics Corporation). Floadia is engaged in the business of licensing the processes and circuit designs required to manufacture an embedded non-volatile memory (which retains its memory content even when the power is turned off) for use in microcomputers, power semiconductors, sensors, and other devices to semiconductor manufacturers as intellectual property (IP). Compared with competing memory technologies, Floadia's non-volatile memory technology requires extremely low power consumption of only one-millionth that of its competitors when writing and erasing data, while having excellent heat resistance and the ability to reduce the additional cost of integrating it into a chip to about one-third. These features have already led to its use in automotive microcontrollers by a Japanese semiconductor manufacturer, and it has also been adopted by a Taiwanese foundry for use as embedded memory chips in smartphone components manufactured by the foundry.

SOURCE Floadia Corporation