Nvidia’s Jetson Orin Nano to be integrated in edge computers by Impulse Embedded
Announced at this year’s GTC 2022 conference by Nvidia, the Jetson Orin Nano series will be available in Q1 2023 and will be integrated into a range of rugged edge computers from Impulse Embedded.
The Jetson Orin Nano module targets the growing demand for real-time processing capabilities in AI devices such as smart cameras, handheld devices, smart meters, and service robots, where integrators often require a lower power and a more cost-effective solution, said Impulse Embedded.
The Jetson Orin Nano offers up to 80x the AI performance of its predecessor, to boost on-device, low-latency processing without increasing power consumption or cost.
Two versions, a 4Gbyte and an 8Gbyte model will be available, with LPDDR5 memory. The 4Gbyte version features a 512-core Nvidia Ampere GPU (graphics processing unit) with 16 Tensor cores. The 8Gbyte version features a 1024-core Nvidia Ampere GPU with 32 Tensor cores. Each version shares the same max GPU frequency of 625MHz, delivering 40 trillion operations per second (TOPS).
Both versions of the Jetson Orin Nano will include a six-core Arm Cortex-A78AE CPU, Ampere architecture GPU, video decode engine, ISP, video image compositor, audio processing engine and video input block.
The modules also feature high speed interfaces including up to seven-lanes of PCIe Gen3, three 10Gbits per second, USB3.2 Gen 2, eight lanes of MIPI CSI-2 camera ports and a range of supported sensor I/O.
The Jetson Orin Nano features the same 70 x 45mm 260-pin SODIMM footprint that is used in the Nano, TX2 NX, Xavier NX and Orin NX modules. This can reduce engineering costs when scaling up, said Impulse Embedded.
The Jetson Orin Nano marks the first time that a single Nvidia GPU architecture (Ampere) will span the entire Jetson range, from the entry-level Orin Nano, through to AGX Orin.
Impulse Embedded offers support for customers developing an industrial AI computing solution. The company said it can create reliable, repeatable and robust systems to help reduce customer’s costs and development time. It has a team of in-house engineers and specialists, all with decades of experience, to deliver fully deployable embedded edge AI computing.