Microchip launches analog embedded SuperFlash® technology to facilitate AI application system architecture implementation

As artificial intelligence (AI) processing moves from the cloud to the edge of the network, battery-powered deeply embedded devices are facing challenges when performing AI tasks such as computer vision and speech recognition. Microchip Technology Inc., through its subsidiary TPV Semiconductor (SST), launched memBrain™ neuromorphic memory solution, an analog memory technology that can significantly reduce power consumption, to effectively address this challenge. Microchip's analog memory solution is based on the industry-recognized SuperFlash® technology and is optimized for the vector matrix multiplication (VMM) execution of neural networks. It improves the system architecture implementation of VMM through analog storage computing methods and improves edge AI reasoning capabilities.

Since current neural network models may require 50M or more synapses (weights) to process, it becomes difficult to provide sufficient bandwidth for off-chip DRAM, becoming a bottleneck for neural network calculations and leading to an increase in overall computing power consumption. In contrast, the memBrain solution stores synaptic weights in on-chip floating gates, significantly improving system latency. Compared with traditional digital DSP and SRAM/DRAM-based approaches, the new products consume 10 to 20 times less power and significantly reduce the overall bill of materials (BOM).

"As technology vendors in the automotive, industrial and consumer markets continue to implement VMM for neural networks, our architecture can help improve the power, cost and latency of these forward-looking solutions," said Mark Reiten, vice president of technology licensing at SST. performance. Microchip will continue to provide highly reliable and versatile SuperFlash memory solutions for AI applications."

Companies looking to improve machine learning capabilities in edge devices have begun adopting memBrain solutions. With the ability to significantly reduce power consumption, memBrain analog memory computing solutions are ideal for all AI applications.

"Microchip's memBrain solution provides ultra-low power storage computing for our upcoming analog neural network processors," said Kurt Busch, CEO of Syntiant. "We are enabling ultra-low power storage computing for voice, imaging and other sensor modalities on edge devices. "Working with Microchip brings a number of key benefits to Syntiant as it delivers a variety of ubiquitous machine learning capabilities for always-on applications."

SST will demonstrate this analog memory solution at the 2019 US Flash Memory Summit, and will also demonstrate Microchip's flash memory performance expansion architecture based on the memBrain product block array at the Artificial Intelligence/Machine Learning Professional Forum. The 2019 US Flash Memory Summit will be held from August 6-8, 2019 at the Santa Clara Convention Center in Santa Clara, California.