On-device AI acceleration
According to Arm, the launch is a response to the “rapid acceleration” of research into running artificial intelligence workloads on the edge. AI is finding its way into a broad spectrum of applications outside the datacentre. These include smartphone apps with on-device AI capabilities and industrial sensors which use neural networks.
To cater to the growing demand, Arm’s devised new processors and software packages designed around edge AI. The company’s announcement includes two new processor series, each of which is optimised for a different facet of artificial intelligence. They’re designed as dedicated coprocessors that offer a “massive” efficiency improvement over regular CPUs.
The Arm ML processor is designed for machine learning operations and can process over 4.6 trillion individual operations each second. It also benefits from a low-power design which makes it ideal for use on the edge. For each watt of power consumed, the chip can complete three trillion operations each second.
The Arm ML is joined by the Arm OD series, which is intended for use in computer vision workloads. It can optionally be combined with the Arm ML to develop a complete system for detecting and recognising people in real-time. The chip processes Full HD video feeds at 60 frames per second. It’s claimed to be capable of simultaneously tracking a “virtually unlimited” number of objects or points.
High scalability
Arm’s announcement comes at a time when on-device AI and edge computing are becoming the forefront of developments in chip engineering. Last week, Intel unveiled its own new processor series intended for use on the edge.
According to Arm, its solution is designed to be more flexible and versatile than rivals from its competitors. The company reflected that current machine learning systems tend to be tailored for individual platforms or workloads. Project Trillium is a response to demand for broader solutions with improved scalability.
READ NEXT: Samsung uses AI to upscale video to 8K, optimise TV sounds
“The rapid acceleration of artificial intelligence into edge devices is placing increased requirements for innovation to address compute while maintaining a power efficient footprint,” said Arm.
“New devices will require the high-performance ML and AI capabilities these new processors deliver. Combined with the high degree of flexibility and scalability that our platform provides, our partners can push the boundaries of what will be possible across a broad range of devices.”
Arm will launch the Project Trillium suite in preview form this April. The company said general availability will commence during mid-2018. Arm also plans to demonstrate devices using the new processors at Mobile World Congress later this month. It’s already integrating the Arm OD chip into security cameras that need advanced object detection.
