Analog flash memory solution enhances AI inference at the edge

August 07, 2019 //By Ally Winning
Microchip Technology’s Silicon Storage Technology (SST) subsidiary is reducing the power required for edge processing with its new memBrain neuromorphic memory.
Microchip Technology’s Silicon Storage Technology (SST) subsidiary is reducing the power required for edge processing with its new memBrain neuromorphic memory, which the company says provides substantial reduction in compute power for AI processing functions like computer vision and voice recognition.

Based on Microchip’s SuperFlash technology and optimized for vector matrix multiplication (VMM) calculations for neural networks, the analog flash memory solution improves implementation of vector matrix multiplication (VMM) through an analog in-memory compute approach.

“As technology providers for the automotive, industrial and consumer markets continue to implement VMM for neural networks, our architecture helps these forward-facing solutions realize power, cost and latency benefits,” says Mark Reiten, vice president of the license division at SST. “Microchip will continue to deliver highly reliable and versatile SuperFlash memory solutions for AI applications.”

memBrain stores synaptic weights in the on-chip floating gate, offering an improvement in system latency. When compared to digital DSP and SRAM/DRAM based approaches, it provides 10 to 20 times lower power usage and reduces BOM.

SST also offers design services for memBrain and SuperFlash technologies, along with a software toolkit for neural network model analysis.

More information

www.sst.com

Related news

Hyperstone presents at the Flash Memory Summit 2019

Global semiconductor revenue hits $120.8 billion in Q2

ARM supports  eMRAM on Samsung's FDSOI process

Memory solution for STM32MP1 MPUs from Rutonik


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.