Winbond Electronics’ semiconductor memory technology is supporting a breakthrough in edge computing from California-based Flex Logix for artificial intelligence (AI) applications such as object recognition.

Winbond is a Taiwanese supplier of semiconductor memory. Its low-power LPDDR4X DRAM technology has been paired with Flex Logix’s Infer X X1 edge inference accelerator chip, which is based on an architecture that features arrays of reconfigurable Tensor processors.

This provides higher throughput and lower latency at lower cost than existing AI edge computing when processing complex neural networking algorithms such as Yolo v3 or full-accuracy Winograd.

“We chose the Flex Logix Infer X X1 edge accelerator because it delivered the highest throughput per dollar, which is critical to drive volume mainstream applications,” said Robert Chang, technology executive at Winbond. “The price-performance advantage of using Infer X with our LPDDR4X chip has the potential to significantly expand AI applications by finally bringing inference capabilities to the mass market.”

To support the Infer X high-speed operation – 7.5Tops maximum – while keeping power consumption low, Flex Logix has paired the accelerator with the W66CQ2NQUAHJ from Winbond, a 4Gbit LPDDR4X DRAM with a maximum data rate of 4267Mbit/s at a maximum clock rate of 2133MHz.

To enable use in battery-powered systems and other power-constrained applications, the W66 device operates in active mode from 1.8V/1.1V power rails, and from a 0.6V supply in quiescent mode. It offers power-saving features including partial array self-refresh.

The Winbond chip operates alongside the X1 processor in Flex Logix’s half-height, half-length PCIe embedded processor board for edge servers and gateways. The system takes advantage of Flex Logix’s architectural innovations, such as reconfigurable optimised data paths, which reduce the traffic between the processor and DRAM, to increase throughput and reduce latency.

“The combination of the unique Infer X X1 processor and Winbond’s high-bandwidth LPDDR4X chip sets a new benchmark in edge AI performance,” said Dana McCarty, vice president at Flex Logix. “Now, for the first time, affordable edge computing systems can implement complex neural networking algorithms to achieve high accuracy in object detection and image recognition even when processing data-intensive high-definition video streams.”

The 4Gbit W66CQ2NQUAHJ is comprised of two 2Gbit dies in a two-channel configuration. Each die is organised into eight internal banks, which support concurrent operation.

The chip is housed in a 200-ball WFBGA package measuring 10 by 14.5mm.

Flex Logix provides products for making flexible chips and accelerating neural network inferencing. The company is headquartered in Mountain View, California.