NVIDIA Jetson Xavier NX SoM Overview
|
Hardware Overview
NVIDIA designed this platform emphasizing in the following three key features:
Small size The NVIDIA Jetson Xavier NX has a 70 mm x 45 mm SoC where the power of an NVIDIA Xavier SoC is packed into a module the size of a Jetson Nano. This small module combines exceptional performance and power advantages with a rich set of IOs—from high-speed CSI and PCIe to low-speed I2Cs and GPIOs. We are now able to take advantage of the small form factor, sensor-rich interfaces, and big performance to bring new capability to all our embedded AI and edge systems.
Huge Performance This platform delivers up to 21 TOPS at 15 W, making it ideal for high-performance compute and AI in embedded and edge systems with size and power-constrained requirements. With 384-CUDA/48-Tensor cores Volta GPU, 2 NVIDIA Deep Learning Accelerators (NVDLA) engines, 6 Carmel ARM CPUs, video encode/decode hardware accelerators, plus 51GB/s of memory bandwidth, makes it the ideal platform to run multiple modern neural networks in parallel and process high-resolution data from multiple sensors simultaneously.
Power efficiency The NVIDIA Jetson Xavier NX supports multiple power modes with high performance capabilities that delivers up to 21 TOPS at 15 W and 14 TOPS at 10 W. This makes this platform ideal for battery-operated systems that demands high processing resources and that share the power budgets with sensors and peripherals. This platform enables the entire NVIDIA software stack that runs all modern AI networks and frameworks with accelerated libraries for deep learning as well as computer vision, computer graphics, multimedia, and more. And it still leaves more of the available power budget for the sensors and peripherals required by today's popular embedded applications.
Processing Specifications
| Feature | Description |
|---|---|
| GPU | 128 Core Maxwell 0.5 TFLOPs (FP16) |
| CPU | 4 core ARM A57 @ 1.43 GHz |
| Memory | 4 GB 64 bit LPDDR4 25.6 GB/s |
| Storage | 16 GB eMMC |
| Encoder | H.264/H.265 up to 4K@30, 4 x 1080p@30, 9 x 720p@30 |
| Decoder | H.264/H.265 up to 4K@60, 2 x 4K@20, 8 x 1080p@30, 18 x 720p@30 |
Peripherals
- 1 x1/2/4 PCIE
- 1 USB 3.0
- SD/MMC Controller
- Display
- HDMI 2.0 or DP1.2
- Up to 2 simultaneous displays
- Camera
- 12 (3x4 or 4x2) MIPI CSI-2 DPHY 1.1 lanes (1.5 Gbps)
- Compatible with Raspberry-Pi v2 camera module
- WiFi/BT — equires external chip
- 1 x SDIO
- 2 x SPI
- 5 x SysIO
- 3 x GPIOs
- 6 x I2C
Package
| Feature | Description |
|---|---|
| Module Size | 69.6 mm x 45 mm |
| Connector | 260 pin SO-DIMM |
| Operating Temperature | Range from -25 – 80C |
| Power Input | 5.0V |
| Module Power | 5 – 25W |
Developer Kit
The developer kit comes with a MIPI-CSI camera connector to enable video applications, as well as an ethernet port, HDMI output, and four USB ports. A 5V/4A source is required to power the kit. Other features offered in the development kit are:
- GPIO
- I2C
- I2S
- SPI
- UART
- Display port
The Jetson Nano developer kit requires an external chip for WiFi and/or Bluetooth and unlike other devkits in the Jetson family, it does not support USB OTG.
Developer Tools
Software development can be done using NVIDIA's Jetpack, as in other NVIDIA Jetson boards. The image generated by Jetpack can be flashed into an SDCard to boot Jetson Nano. Jetpack provides Ubuntu 18.04 OS with kernel 4.9, find the information about how to install JetPack in Related Documentation. For AI and parallel computing applications, Jetpack provides CUDA, cuDNN, and TensorRT. Additionally, tools like nvpmodel to control the power/performance profile and tegrastats that provide CPUs and GPU stats, come with the filesystem.
The 128-core Maxwell GPU supports OpenGL 4.6, OpenGL ES 3.2 and Vulkan 1.1 as well.
Jetson Nano vs Jetson TX1
Nano and TX1 are not pin compatible, however, they share common features, like USB 3.0 and CSI. Both come with the same size of eMMC and memory. Encode/decode capabilities also remain the same. While TX1 has 256-core GPU, Nano comes with 128-core, both NVIDIA Maxwell GPUs. Nano CPU has a lower frequency of 1.43 GHz compared with TX1 CPU frequency of 1.7 GHz. The differentiator of the Jetson Nano with respect to other boards in the Jetson family is its low power consumption, it requires an input voltage of 5 V while TX1 input voltage vary between 5.5 V and 19.6 V.