Compare Nvidia Jetson products
Parameter | Jetson Nano | Jetson TX2 | Jetson TX2i | Xavier NX | AGX Xavier 8GB | AGX Xavier |
---|---|---|---|---|---|---|
AI performance | 0,47 TFLOPs | 1,3 TFLOPs | 1,3 TFLOPs | 14 / 21 TOPs | 20 TOPs | 32 TOPs |
GPU | 128-core Maxwell | 256-core Pascal | 256-core Pascal | 384-core Volta | 384-core Volta | 512-core Volta |
NVIDIA Tensor cores | 48 | 48 | 64 | |||
CPU | Quad-Core ARM Cortex-A57 | Dual-Core NVIDIA Denver 2 CPU and Quad-Core ARM Cortex-A57 | Dual-Core NVIDIA Denver 2 CPU and Quad-Core ARM Cortex-A5763 | 6-core NVIDIA Carmel ARM v8.2 64-bit CPU 6MB L2 + 4MB L3 | 6-core NVIDIA Carmel Arm v8.2 64-bit CPU 6MB L2 + 4MB L3 | 8-core NVIDIA Carmel Arm v8.2 64-bit CPU 8MB L2 + 4MB L3 |
Memory | 4 GB | 4 GB / 8 GB | 8 GB | 8 GB | 8 GB | 32 GB |
Storage | 16 GB | 16 GB / 32 GB | 32 GB | 16 GB | 32 GB | 32 GB |
Power consumption | 5 / 10 W | 7,5 / 15 W | 10 / 20 W | 10 / 15 W | 10 / 20 W | 10 / 15 / 30 W |
Wi-fi | No | No / Yes | No | No | No | No |
Connector | 69,6 mm × 45 mm 260-pin SO-DIMM connector | 87 mm × 50 mm 400-pin connector | 87 mm × 50 mm 400-pin connector | 69,6 mm × 45 mm 260-pin SO-DIMM connector | 100 mm × 87 mm 699-pin connector | 100 mm × 87 mm 699-pin connector |
Market launch | March 2019 | March 2017 | March 2017 | March 2020 | December 2018 | December 2018 |
We supply NVIDIA Jetson developer kits and production NVIDIA Jetson modules. We consult on the appropriate combination of carrier boards and chassis/boxes for mounting the modules. We have Jetson developer kits in stock for testing including cameras, power supply, SD cards.
Interesting links and information on Nvidia Jetson products
We thank very much Sven Stuewe from Nvidia for his support and provided information.
Jetson basics
- https://developer.nvidia.com/embedded/faq#xavier-faq
- https://developer.nvidia.com/embedded-computing
- https://developer.nvidia.com/embedded/community/ecosystem#machine_vis_cam_sens
- https://devtalk.nvidia.com/default/board/139/embedded-systems/1
- https://elinux.org/Jetson_AGX_Xavier
- https://elinux.org/Jetson_TX2
- https://elinux.org/Jetson_Nano
Other useful links
The Jetson TX1/TX2 4GB/TX2/TX2i have the same connector/interface and carrier board they connect to (which will provide the Jetson with the necessary connectors), so it can be the same for all these models.
Jetson Nano and Jetson AGX Xavier have their own carrier board (electrically and mechanically).
As of April 15, 2019, the SW platform is the same for all Jetson products – Nano, TX2/TX2i, Jetson AGX Xavier.
Main portal for Nvidia Jetson (Embedded) products:
The Download section contains a lot of technical information and specifications, including power and cooling requirements, etc.
JetPack:
JetPack is a simple installer that performs BSP installation and all other related SDKs (CUDA, cuDNN, TensorRT, VisionWorks). A software patch is available on the Jetson NX Xavier and is directly supported from JetPack 4.4 onwards.
Devblog on Jetson Nano: https://devblogs.nvidia.com/jetson-nano-ai-computing/
Jetbot: https://github.com/NVIDIA-AI-IOT/jetbot
Jetson FAQ, including information about TX2 vs TX2i MTBF: https://developer.nvidia.com/embedded/faq
Jetson module availability (lifecycle): https://developer.nvidia.com/embedded/community/lifecycle
NVIDIA cuDNN: https://developer.nvidia.com/cudnn
NVIDIA Tensor RT: https://developer.nvidia.com/tensorrt
Deepstream 4.0 : https://developer.nvidia.com/deepstream-sdk
Transfer Learning and pruning: https://developer.nvidia.com/transfer-learning-toolkit
Deepstream plugins: https://github.com/vat-nvidia/deepstream-plugins
Blog on Deepstream 3.0: https://devblogs.nvidia.com/intelligent-video-analytics-deepstream-sdk-3-0/
Nvidia Deep Learning Institute
The Deep Learning Institute (DLI) offers practical online and hands-on training for developers, data analysts or researchers who want to tackle challenges in artificial intelligence (AI). Among other things, it also offers a great Getting Started with Jetson Nano course(provided free by NVIDIA):
Code samples on Dustin github: https://github.com/dusty-nv
Tutorial on how to create a DL demo: https://developer.nvidia.com/embedded/twodaystoademo
Interesting presentations from GTC US 2017:
Testing
To test the performance and especially the deployment speed of ML and AI applications, we have the DGX Station system, 2× Tesla V100 and Jetson Nano developer kits including camera, power supply and SD card as part of the Test Drive program. If you are interested in our testing offer, please fill out this form.