The project, which was funded by an Innovate UK Smart grant, will also be demonstrated at the upcoming Flash Memory Summit where the chip, an accelerator solution for computer vision (CV) AI-enhanced image recognition, will be showcased.
A paper will also be presented at the Summit, describing the development of the RISC-V-based chip, and will discuss the achievement of acceleration by a factor of 16 to 128 times for processing image data, along with ultra-low power consumption.
When used in a security camera monitoring a rapidly-evolving active shooter situation, for example, the chip can enable real-time identification of different types of firearms to automatically trigger an alarm – a potential game changer that could save many lives.
The Cambridge Architecture has been developed to address the Von Neumann Bottleneck – the phenomenon that data transfer between the core and the memory has become the limiting factor in computational speed. As computing tasks grow more data-hungry, it overcomes a growing obstacle to computational efficiency, and it also offers huge energy savings by eliminating unnecessary movement of data.
“This is the first time that Blueshift Memory’s technology has been demonstrated in a real-life application, and the results are extremely promising,” said Peter Marosan, CTO and founder of Blueshift Memory. “We know that in more challenging, data-intensive use cases like servers for high-frequency trading, the Cambridge Architecture is capable of even higher levels of acceleration, up to 1000x or more, and this is the first step towards us reaching that market. This high performance will also be accompanied by dramatic energy savings, since moving large amounts of data around unnecessarily makes excessive demands on energy consumption.”
Recog.AI, based in Budapest, Hungary, will act as a distributor for the chips and accelerator modules, which it plans to use both in CCTV cameras and in the cloud.
“Recog.AI can set up a cloud-based Computer Vision as a Service (CVaaS) solution for its clients, powered by Blueshift Memory cloud inference cards," said Máté Hegedűs, CEO of Recog.AI. “The chip will offer the users considerable benefits such as faster processing, lower latency, and improved energy efficiency. Alternatively, by using the solution as a standalone chip we can offer enhanced real-time image and video analysis capabilities on the edge and integrate these devices into our own computer vision platform."
