Nvidia brings the computer in cars to the office


Nvidia pulls its Raspberry Pi-like on-board computers, named Orin, out of car trunks and labs. The Jetson AGX Orin 32GB development board is the first of four new AGX Orin development boards coming to desktop computers for the development of robotics as well as industrial and medical applications.

The computer – which looks like a motherboard – is a sibling of the Drive Orin computer that will power self-driving cars from Volvo and Mercedes. The Jetson board has an Orin SoC with ARM CPU cores and Ampere GPUs, which are also in Orin computers in cars.

Nvidia shipped its first Jetson development board, called TK1, in 2014, and it featured the company’s ARM-based Tegra K1 smartphone chip. The TK1 board was modeled around the Raspberry Pi, which spawned the do-it-yourself subculture and many innovative DIY hardware projects. Earlier Jetson cards have been used in local robots and drones.

The company has since shifted its GPU development to AI applications, and Jetson is now a hardware platform on which companies can test, deploy and run AI and graphics applications.

Nvidia already dominates AI and machine learning in data centers with its GPUs, but the company is also trying to capture AI inference and edge applications through hardware like Jetson cards.

Jetson Orin offers a higher level of performance than a Raspberry Pi, making it better for prototyping products that could ultimately be sold commercially, said Bob O’Donnell, founder of Technalysis Research.

“It’s their particular robotics thing. There’s a lot of people doing cool things in robotics, like AGVs (automatic guided vehicles) in warehouses and industrial robots,” O’Donnell said.

The Jetsons use older GPU architecture because of cost and power requirements, and devices like robots and drones don’t require the latest and greatest computer chips, O’Donnell said.

Nvidia said the new Jetson AGX Orins are up to eight times faster than their predecessor, called the Xavier. Nvidia recently introduced Ampere’s successor, called Hopper, but the company hasn’t announced any plans for GPU architecture on Jetson cards.

Jetson Orin cards offer server-class performance in a compact form factor, which should improve developer productivity and expand use cases, said Charles King, principal analyst for Pund-IT.

The Jetson AGX Orin 32GB delivers 200 TOPS of INT8 performance. It has 1,792 GPU cores running at a frequency of 930 MHz, an 8-core Arm Cortex-A78AE processor, 32 GB of LPDDR5 DRAM, and 64 GB of storage. Nvidia is selling the card for $999.

The Jetson AGX Orin 32GB module. Image courtesy of Nvidia.

In November, the company will ship Jetson AGX Orin 64 GB, which offers the 275 TOPS of performance. It has 2,048 GPU cores running at 1.3 GHz, a 12-core Arm Cortex-A78AE processor, 64 GB of LPDDR5 memory, and 64 GB of storage. The computer is based on hardware included in the $1,999 Jetson AGX Orin dev kit, which already ships and has a chassis.

GPUs and the power of graphics and AI separate Jetson boards from the Raspberry Pi and other development boards. The Jetson has HDMI and DisplayPort ports, and it supports camera input for computer vision applications.

The Jetson roadmap also includes the less powerful Orin NX boards, which look more like on-board parts than full-fledged development computers. NX cards will ship later this year with 1,024-core Ampere GPUs and up to 16GB of storage.

Nvidia also offers simulation software and tools for cards that help with product development, and that’s something other card makers can’t replicate, said O’Donnell of Technalysis.

For example, Isaac SIM allows roboticists to combine computer-generated synthetic data with real-world data to better simulate robots in the virtual world.

Development computers run on Linux, and applications can be developed with a dedicated software development kit called JetPack. The software includes the Ubuntu desktop environment and libraries for AI, graphics and computer vision, an Nvidia spokesperson said. Nvidia recently introduced the TAO AI Toolkit – whose standards train, adapt and optimize – which makes it easier to write and deploy AI applications. The new development boards will support TAO and other AI toolkits like Riva for conversational AI.

The Jetson roadmap includes a new version of the Nano card – which is popular with developers – for release in 2023.


Comments are closed.