Nvidia introduced the upcoming launch of the Jetson Orin Nano, a system-on-module (SOM) that may energy up the following technology of entry-level AI and robotics, throughout its GTC 2022 keynote immediately.
Nvidia says this new model delivers an 80x enhance in efficiency over the $99 Jetson Nano. The unique model was launched in 2019 and has been used as a bare-bones entry into the world of AI and robotics, notably for hobbyists and STEM college students. This new model seems to significantly up the ability.
A system-on-module (additionally known as computer-on-module) incorporates a single board with a microprocessor. It additionally has reminiscence and enter/outputs (IOs), and normally has a service board. It’s not the identical factor as a system-on-a-chip (SOC), thoughts you — an SOM is board-based and will have the house to incorporate additional elements; it may even embody an SOC. In brief, an SOM is a ready-to-use computing resolution, but it surely’s not a full pc.
With the technicalities out of the best way, let’s discuss Nvidia’s newest improvement, the Jetson Orin, arriving with six Orin-based manufacturing modules that have been made to deal with AI and robotics purposes at an reasonably priced value. Among them is the Nvidia Jetson Orin Nano.
Despite being the smallest kind issue Jetson SOM, the Jetson Orin Nano can deal with as much as 40 trillion operations per second (TOPS) of AI-related duties. The efficiency hits new heights with the AGX Orin, serving up 275 TOPS to be able to deal with superior autonomous machines.
Nvidia’s Jetson Orin comes with an Ampere-based GPU, an ARM-based CPU, and multimodal sensor help. It’s additionally absolutely suitable with Nvidia’s Orin NX modules, together with full emulation help that may allow Nvidia’s clients to design round a number of Jetson modules. Other perks embody help for a number of concurrent AI utility pipelines, full with quick inputs and outputs.
The Jetson Orin Nano modules can be out there in two variants, one with 8GB reminiscence and as much as 40 TOPS, and one with 4GB reminiscence and as much as 20 TOPS. In phrases of energy consumption, the SOM wants subsequent to nothing: The former requires between 7 watts and 15 watts whereas the latter solely wants 5 watts to 10 watts.
Nvidia foresees that the modules can be utilized by all kinds of consumers, from engineers coping with edge AI purposes to robotics working system builders. The low value level, beginning at simply $199, will make this expertise extra accessible to a wider vary of customers. Nvidia cites Canon, John Deere, Microsoft Azure, and extra as early adopters of Jetson Orin Nano.
“With an orders-of-magnitude increase in performance for millions of edge AI and ROS developers today, Jetson Orin is the ideal platform for virtually every kind of robotics deployment imaginable,” mentioned Deepu Talla, vice chairman of Nvidia’s embedded and edge computing division.
Nvidia claims that the Jetson Orin will supply an 80x enhance in efficiency over the earlier technology of Jetson SOMs. That’s a large step up at an affordable value. The modules can be out there beginning in January 2023.