Graphics chip specialist NVIDIA announced a slew of new products and partnerships at its annual GPU Technology Conference (GTC) in San Jose this past week. Among the most exciting was its teaming with leading semiconductor designer Arm Holdings to bring artificial intelligence (AI) to billions of Internet of Things (IoT) devices.
The IoT refers to the networking of physical objects through the use of embedded sensors and actuators. The scads of data collected about the objects can then be analyzed to improve products, services, and operations. Some of the IoT's burgeoning applications include many things "smart" -- such as homes, cars, cities, and factories -- and wearables for health and fitness monitoring.
NVIDIA and U.K.-based Arm, which was acquired by Japan's SoftBank (NASDAQOTH: SFTBF) in a $32 billion deal in 2016 and is now owned by the Japanese company and its Vision Fund, are partnering to bring deep-learning inferencing to the billions of mobile, consumer electronics, and Internet of Things devices that are expected to enter the global marketplace.
Deep learning (DL) is a type of machine learning (ML) within AI that aims to mimic the way the human brain processes data and creates patterns for use in decision making. NVIDIA's graphics processing unit (GPU)-based approach to AI is a deep-learning tech that's being rapidly adopted by companies for many different applications. Inferencing is the second step of the two-step DL process, following training, and involves a machine applying its training to new data. DL inferencing is primarily done in data centers, like training, but it's increasingly moving out of data centers and being performed by "edge devices," right where the data is being collected, which is often called "inferencing at the edge" or "AI at the edge." These edge devices include such things as cars and drones -- both of which NVIDIA has specific platforms for -- and, now with the Arm teaming, IoT devices of all kinds.
NVIDIA and Arm will integrate the open-source NVIDIA Deep Learning Accelerator (NVDLA) architecture into Arm's Project Trillium platform for machine learning. This will make it simple for IoT chip companies to integrate AI into their chip designs, which should help accelerate the growth of smart products. NVDLA is based on NVIDIA Xavier, the world's most powerful autonomous machine system on a chip, which NVIDIA introduced earlier this year. It's a free, open architecture that the GPU specialist touts is "scalable, highly configurable, and designed to simplify integration and portability." NVDLA is aimed at promoting a standard way to design deep-learning inference accelerators.
"Inferencing will become a core capability of every IoT device in the future," said Deepu Talla, vice president and general manager of Autonomous Machines at NVIDIA, in the press release. "Our partnership with Arm will help drive this wave of adoption by making it easy for hundreds of chip companies to incorporate deep learning technology."