Nvidia’s vary of Jetson forums don’t seem to be standard Raspberry Pi possible choices. Nvidia’s focal point is on entry-level and mainstream AI primarily based robotics, drones and cameras. Its newest board, the $499 Jetson Orin Nano ups the processing energy whilst nonetheless holding the package compact.
The Jetson Orin Nano improves at the Nvidia Maxwell GPU’s 128 CUDA cores with 1024 Ampere GPU primarily based CUDA cores. The additional cores and more recent structure signifies that the Orin Nano has as much as 80 instances the AI efficiency of the Jetson Nano. The six Arm A78AE CPU cores supply virtually seven instances the efficiency of the Jetson Nano. The similar AI structure that powers the Jetson AGX Orin module is used within the Orin Nano, however at a a lot more reasonably priced value.
Nvidia shipped me a Jetson Orin Nano pattern for overview, however because of some logistical problems, the unit arrived with little time for a complete overview, so right here I’ll supply a short lived creation to the board and observation on what I discovered. Sadly, I used to be not able to do any trying out of the principle use case for the Orin Nano, inference and gadget finding out, as a result of what I attempted in this beta-level instrument simply didn’t paintings for me within the restricted time I had it. We’ll post a complete overview with inference benchmarks in a couple of days, when we’ve optimistically been in a position to get a more recent construct of the Orin Nano’s instrument.
Word that JetPack instrument supplied with the board is a non-public preview and does now not mirror the general instrument that will likely be to be had to customers. As soon as the general instrument free up is made to be had, I’ll supply a complete overview of the Orin Nano, together with its robust AI features.
Contents
Jetson Orin Nano Specs
| Header Cellular – Column 0 | Jetson Orin Nano | Jetson Nano |
|---|---|---|
| CPU | 6-core Arm Cortex-A78AE v8.2 64-bit CPU | Quad-core ARM Cortex-A57 MPCore processor |
| 1.5MB L2 + 4MB L3 | ||
| GPU | Nvidia Ampere structure with 1024 Nvidia CUDA cores and | Nvidia Maxwell structure with 128 Nvidia CUDA cores |
| 32 Tensor cores | ||
| Reminiscence | 8GB 128-bit LPDDR5 | 4 GB 64-bit LPDDR4, 1600MHz 25.6 GB/s |
| 68 GB/s | ||
| Garage | Micro SD | 16 GB eMMC 5.1 |
| NVMe M.2 by way of Service Board | Micro SD | |
| Energy | 7W to 15W (5V Enter Voltage) | 20W (Max 5V at 4 Amps) |
| Dimensions | 69 x 45 x 21 mm | 69.6 x 45 x 20 mm |
Jetson Orin Nano Service Board Specs
| Header Cellular – Column 0 | Jetson Orin Nano | Jetson Nano |
|---|---|---|
| Digicam | 2x MIPI CSI-2 22-pin Digicam Connectors | 12 lanes (3×4 or 4×2) MIPI CSI-2 D-PHY 1.1 |
| M.2 Key M | x4 PCIe Gen 3 | |
| x2 PCIe Gen3 | ||
| M.2 Key E | PCIe (x1), USB 2.0, UART, I2S, and I2C | 1 x |
| USB | 4 x USB 3.2 Gen2 | 4x USB 3.0 |
| 1 x Kind C for debug and tool mode | 1 x USB 2.0 Micro-B | |
| Networking | Gigabit Ethernet | Gigabit Ethernet |
| RTL8822CE 802.11ac PCIe Wi-fi Community Adapter | ||
| Show | DisplayPort 1.2 | HDMI 2.0 and eDP 1.4 |
| GPIO | 40 Pin GPIO | 40 Pin GPIO |
| 12 Pin Button Header | ||
| 4 Pin Fan Header | ||
| Energy | DC 9-19V Barrel Jack | DC Barrel Jack 20W (Max 5V at 4 Amps) |
| Dimensions | 100 x 79 x 21 mm (Top comprises Orin Nano module and cooling answer) | 100 x 80 x 29mm (Top comprises Jetson Nano module and cooling answer) |
At a passing look, the Orin Nano and the Jetson Nano glance equivalent. What offers the Orin Nano away are a fan constructed right into a heatsink and the loss of HDMI port. The USB-C port replaces the micro USB of the Jetson Nano. The aforementioned fan is whisper quiet, even if we’re working on the complete 15W. We ran considered one of Nvidia’s prompt inference benchmarks and the fan stayed quiet, in contrast to different lovers we now have examined on SBCs.
Inference Checking out
Presently this segment is brief, and now not very candy. Nvidia’s claims that the Orin Nano delivers virtually 30 instances the efficiency of the Jetson Nano (that it hopes to give a boost to to 45 instances) I used to be not able to make sure.
The foundation reasons of this being a brief timescale and the personal instrument construct. I sought after to show the Hi AI International instance the use of a Raspberry Pi Digicam Module 2, however I bumped into digicam problems which noticed the instrument encoder now not detecting the digicam, regardless of it being indexed as suitable. Those problems were raised to Nvidia, and I’m hoping {that a} long term JetPack OS free up will get to the bottom of those problems.
The Desktop Revel in
Operating JetPack 5, a customized model of Ubuntu 20.04, the 8GB of LPDDR5 and six-core Arm CPU supply sufficient energy for basic desktop tasks. Then again, we wouldn’t suggest making an investment $500 on this board simply to make use of it as a desktop PC.
First boot used to be just a little slower than we was hoping, however Nvidia has mentioned within the reviewer’s information that ultimate manufacturing devices won’t have this factor. Every other factor we noticed used to be that most effective 6.3GB of RAM used to be to be had within the preview construct. The entire 8GB will likely be to be had to finish customers by way of a repair. The Ubuntu enjoy used to be delightful, with the minimum quantity of customization made to the desktop, wanting putting in gear explicit to the strengths of the Orin Nano.
Set up of Chromium took just a little longer than we’d have anticipated. It apparently put in the browser by way of Snap, Canonical’s most popular packaging platform. Name us old style, however we nonetheless have numerous love for APT.
With the set up whole, we opened Chromium after which went to YouTube to look at a few HDR and 4K movies. First used to be LeePSPVideo’s HDR video take a look at, which we set to fullscreen and at 1440p. Video playback used to be nice, as stats for nerds confirmed a tiny collection of frames dropped for the 1440p 30fps video.
If we hadn’t used stats for nerds, we’d by no means have spotted. The following video, a shuttle round Costa Rica and its flora and fauna used to be performed at 1440p fullscreen, however this 60 fps video fared worse. It dropped round 4% of the frames over its complete run, the overwhelming majority being at the beginning of the video. In spite of that factor, playback used to be nice.
Lacking from the Orin Nano is a devoted {hardware} encoder (NVENC). As an alternative, Nvidia provides a instrument encoder the use of the six-core Arm A78AE CPU. This turns out like a downgrade from the Jetson Nano, however in all probability the 2 further Arm CPU cores are there to make up for it?
The loss of a {hardware} encoder additionally affects how we use a digicam with the Orin Nano. There are two 15 pin CSI connectors at the left facet of the service board. Those have compatibility with CSI cables made for the Raspberry Pi 0. We hooked up a Raspberry Pi Digicam Module 2 to CAM0 and examined a handy guide a rough script to report video. Unfortunately this wasn’t to be with our preview construct of the OS. In spite of the IMX219 sensor of the Raspberry Pi Digicam Module 2 being suitable, we by no means controlled to get a picture.
The use of the GPIO
Symbol 1 of 3
The 40 pin GPIO of the Orin Nano is at the proper hand facet of the service board and here’s our first factor. What pins are we connecting to? At the Jetson Nano we had the board reference revealed as a silkscreen subsequent to the pins.
For the Orin Nano, we need to turn over the board and carry out a feat of psychological dexterity to bear in mind the place each and every pin is. This used to be compounded by means of the Python examples the use of a Broadcom (BCM) mapping (Raspberry Pi additionally makes use of BCM mappings in all its reputable tutorials) which required additional interpreting. The Python module is RPi.GPIO, a module that Raspberry Pi lovers will likely be in detail acutely aware of. Created by means of Ben Croston, this Python module has powered 1000’s of Pi tasks, and reasonably a couple of Jetson tasks too. The module has been tweaked to run at the Jetson forums and is as acquainted as ever to those eyes. To get across the BCM to BOARD pin mappings we selected the bodily (BOARD) pin mappings, regardless of our years of enjoy instructing Raspberry Pi primarily based content material.
It labored and we had an LED flashing. The GPIO pins additionally give you the same old plethora of conversation protocols. From easy virtual IO to UART, SPI, I2C and I2S. The GPIO of the Orin Nano isn’t the point of interest of the board, extra of an added characteristic for those who need to merge gadget finding out with robotics or a chain of sensors.
Nvidia’s Jetson Orin Nano developer package is to be had now for $499 by way of licensed vendors.
Supply By means of https://www.tomshardware.com/information/nvidias-new-orin-nano-developer-kit-like-a-raspberry-pi-for-ai