Device Type

CHI@Edge Supported Device Types

This page details the Single Board Computers (SBCs) available for edge computing experiments. These devices serve as the primary compute nodes and hosts for peripherals.

machine_name (for reservation)	Description
raspberrypi5	Raspberry Pi 5, all variants
raspberrypi4-64	Raspberry Pi 4, all variants, including CM4
jetson-nano	Nvidia Jetson Nano
jetson-xavier-nx-devkit-emmc	Nvidia Jetson Xavier NX
jetson-orin-nano-devkit-nvme	Nvidia Jetson Orin Nano Developer Kit (NVMe)
jetson-agx-orin-devkit-64gb	Nvidia Jetson AGX Orin 64GB Developer Kit

Hardware Details

Raspberry Pi 5

• 8gb and 16gb variants

• https://www.raspberrypi.com/products/raspberry-pi-5/

Compute Module 5 (coming soon)

• https://www.raspberrypi.com/products/compute-module-5/?variant=cm5-104032

• Lower power draw, wide temperature range support, suitable for embedding into outdoor or rugged applications.

Raspberry Pi 4

Model B, 8gb

• https://www.raspberrypi.com/products/raspberry-pi-4-model-b/specifications/

compute module 4, 8gb

• https://www.raspberrypi.com/products/compute-module-4/?variant=raspberry-pi-cm4001000

• Lower power draw, wide temperature range support, suitable for embedding into outdoor or rugged applications.

Nvidia Jetson Nano

• https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-nano/product-development/

The latest Nvidia software supported on the Jetson Nano is:

• JetPack 4.6.6

• L4T 32.7.6

This is based on Ubuntu 18.04, Linux Kernel 4.9, and CUDA 10.2

You must use a container image targeting CUDA10.2 to work on these devices.

The NVIDIA SDK for these devices is unfortunately end-of-life, and we are rolling out support for the replacement "Orin Nano" devices, described below.

Nvidia Jetson Xavier NX

• https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-xavier-series/

The latest Nvidia software supported on the Xavier NX is:

• JetPack 4.6.6

• L4T 32.7.6

This is based on Ubuntu 18.04, Linux Kernel 4.9, and CUDA 10.2.

You must use a container image targeting CUDA 10.2 to work on these devices.

Support for these is limited, as NVIDIA considers them EoL. The replacement is the AGX Orin and Orin Nano, described below.

Jetson Orin Nano Developer Kit

• https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/

• https://nvdam.widen.net/s/zkfqjmtds2/jetson-orin-datasheet-nano-developer-kit-3575392-r2

The Orin Nano is Nvidia's current offering, replacing the Jetson Nano. CHI@Edge images target the NVMe variant of the developer kit (jetson-orin-nano-devkit-nvme);

The Nvidia software supported on the Orin Nano is:

• JetPack 6.x

• L4T r36.4

• CUDA 12.6

This is based on Ubuntu 22.04. The Orin Nano is currently running in the original (non-Super) power profile.

Container images must target CUDA 12.6 (or compatible) and linux/arm64.

Jetson AGX Orin 64GB Developer Kit

• https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/

• Jetson AGX Orin Developer Kit datasheet

The AGX Orin is the high-end Orin-series board. CHI@Edge currently supports the 64GB developer kit (jetson-agx-orin-devkit-64gb).

The Nvidia software supported on the AGX Orin is:

• JetPack 6.x

• L4T r36.4

• CUDA 12.6

This is based on Ubuntu 22.04.

Container images must target CUDA 12.6 (or compatible) and linux/arm64.

Hardware Summary Table

Device Name	CPU	Memory	GPU Arch	AI Perf [1]	Power (TDP)
Raspberry Pi 4B	4c A72 @ 1.8GHz	8GB LPDDR4 13.3 GB/s	VideoCore VI		3W - 7W
Raspberry Pi 5	4c A76 @ 2.4GHz	8GB LPDDR4x 27.3 GB/s	VideoCore VII		5W - 12W
Jetson Nano	4c A57 @ 1.43GHz	4GB LPDDR4 25.6 GB/s	Maxwell (128c)	472 GFLOPS	5W - 10W
Jetson Xavier NX	6c Carmel v8.2	8GB LPDDR4x 59.7 GB/s	Volta (384c+48T)	21 TOPS (Dense, Int8)	10W - 20W
Jetson Orin Nano (Dev Kit)	6c A78AE @ 1.5GHz	8GB LPDDR5 68 GB/s	Ampere (1024c + 32T)	40 TOPS (Sparse, Int8)	7W - 15W
Jetson AGX Orin 64GB	12c A78AE @ 2.2GHz	64GB LPDDR5 204.8 GB/s	Ampere (2048c + 64T)	275 TOPS (Sparse, Int8)	15W - 60W