Author: Zen Kuda Nixjoen
Artificial intelligence is evolving all around us. What used to be a speculative topic found only in sci-fi movies is now all around us. AI technology is evolving and tech companies are trying their best to keep up, even companies focusing on small SB computers. Unfortunately, these advanced single-board computers and microcontroller-enabled development kits are not yet capable to handle AI applications simply because they lack the right hardware. NVIDIA has exploited this gap in the market with its “Jetson” branded products. In this review, we will be speaking about the powerful Jetson Nano Developer Kit.
WHAT IS THE JETSON NANO?
The Jetson Nano is a small, compact powerful computer specifically designed to power and handle entry-level AI applications and devices. With the Jetson Nano Developer Kit, Nvidia has given developers a capable system to create, test and deploy their AI applications. Plus, all this is possible without having to connect to the internet or rely on the cloud. The NVIDIA Jetson Nano opens the door to a new world of embedded Internet-Of-Things (IoT) applications, including home robots, network video recorders, and other applications with great analytical capabilities. The NVIDIA Jetson Nano Developer Kit is the kit to begin your journey into the world of artificial intelligence and robots. Thanks to a powerful board, step-by-step projects, and an interactive developer community the learning process is easier than ever.
Let us take a look at what the Jetson Nano has to offer:
|GPU:||128-Core Maxwell 472 GFLOPs (FP16)|
|CPU:||4 core ARM A57 @ 1.43GHz|
|RAM:||4GB 64 bot LPDDR4 25.6 GB/s|
|VIDEO ENCODE:||4K @ 30 |4x 1080p @ 30 | 8x 720p @ 30|
|VIDEO DECODE:||4K @ 60 | 2X 4K @ 30 | 8X 1080P @ 30 |16X 720P @30 | H.264/H.265|
|CAMERA:||12 (3×4 or 4×2) MIPI CSI-2 DPHY 1.1 lanes|
|DISPLAY:||HDMI 2.0 or DP1.2 | eDP 1.4 | DSI (1×2)|
|UPHY:||1 x 1/2/4 PCIE|
1 USB 3.0
|SDIO/SPI/SysIOs/GPIOs/12C:||1x SDIO/2x SPI/5x SysIOs/13x GPIOs/6x 12C|
GETTING STARTED WITH THE JETSON NANO
Setting up and getting started with the NVIDIA Jetson Nano is quite a simple and straightforward process. Once you take the Jetson nano out of the packaging, you will come across the included Quick Start guide which will direct you to an online video. The video will show you exactly how to download and install the latest Linux4Tegra Operating System onto the Micro SD card.
JETSON NANO DEVELOPER KIT VS RASPBERRY PI 3
It’s always good to compare products with their competitors so you can always get the best products available. One competitor of the NVIDIA Jetson Nano is of course the Raspberry Pi 3. With a quick glance at the two boards side by side, we can see that the Jetson Nano is significantly larger in size. On the software side of things, the Maxwell-based GPU in the Jetson Nano will definitely produce more heat than the Pi 3 and it will require you to purchase a quality heatsink to help cool it. The heating issue is to be expected because the Jetson Nano developer Kit is SIGNIFICANTLY faster than the Raspberry Pi and is capable of handling much more taxing applications as well as more AI-specific processes that the Raspberry Pi 3 would not be able to handle.
Let’s bring in another competitor of the Jetson nano, the Coral Board with Edge TPU. The NVIDIA Jetson Nano’s Cortex A57 based processer is a whole lot more powerful than the ARM Cortex A53 based CPUs in the Raspberry Pi 3 and the Coral Boards. To top it off the Jetson Nano has 4GB of RAM whereas its competitors have just 1GB. Whether you’re running AI applications or regular everyday apps, you will get more speed and power from the NVIDIA Jetson nano Developer Kit.
JETSON NANO DEVELOPER KIT HARDWARE
Compared to other SBCs like the Raspberry Pi 3, the NVIDIA Jetson Nano has a much more sophisticated architecture that supports a much higher bandwidth. The Jetson Nano incorporates a comprehensive I/O set. This gives the engineers the chance to include high-speed networking interfaces such as Gigabit LAN ports and USB 3.0 which are not found on a lot of single-board computers.
Of course, when talking about development, not only the hardware matter but also the software library support, compatibility, and development support. Thanks to NVIDIA’s vast amount of experience in this area, the NVIDIA Jetson nano Developer Kit answers all those questions.
Linux4Tegra is the official OS for the NVIDIA Jetson nano and other boards in the same family. It is actually a version of the Linux Ubuntu 18.04 OS specifically designed to run on NVIDIA’s hardware. The overall appearance and feel of the desktop when using the Jetson Nano is the same as on a standard Ubuntu Linux OS, so for Linux users, this is an added familiarity bonus. There is also a massive ecosystem of tried and tested Linux applications – from games to development software, it is all available on the bundled software app. You should know that the AI demos and the AI software that is recommended in NVIDIA’s documentation are not available in a store but instead, you have to download, compile and invoke the software directly through the command line. It would be great for novice users if the software was already preloaded on the OS but I must add that the process of using the command line to download and install programs is typical in the Linux development environment.
JETSON NANO DOWNSIDES
The one big downside of the NVIDIA Jetson Nano Developer Kit is the lack of onboard wireless connectivity. All the Raspberry Pi boards in the Jetson Nano bracket have built-in Wi-Fi and Bluetooth options. Unfortunately for Jetson Nano users will have to make do with an Ethernet connection, only writing offline programs or attaching a third-party USB Bluetooth/Wi-Fi dongle, which can be quite a tedious task.
Rather than having to spend your rent’s worth of money on a third-party development kit compatible with NVIDIA’s machine learning platform; you can pick the NVIDIA Jetson Nano Developer Kit up for under $199 and be a part of the Jetson platform. You will still be getting a highly capable developer kit and you will be able to attempt all sorts of fun, challenging, and interesting machine learning tasks. Of course, you could always go out and buy a Raspberry Pi for much cheaper but that is up to your requirements and your budget. If you want a faster, more powerful machine that can deliver you hardware-accelerated machine learning performance that is unheard of in most raspberry pi boards, then the NVIDIA Jetson nano Developer Kit is perfect for you.
You can get your Jetson Nano Dev Kit here.
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.