Innovative NVIDIA Jetson AGX Xavier 8GB Module

Home / Categories / AI Products / AGX Xavier 8GB Module

AGX Xavier 8GB Module

Per piece

Product details

Jetson AGX Xavier 8GB is a lower-power, lower-price Jetson AGX Xavier offering full hardware and software compatibility with the existing Jetson AGX Xavier. It consumes a maximum of 20W for the full module while delivering up to 20 TOPS of AI performance.
The Jetson AGX Xavier 8GB Module is an accelerator for use in AI and deep-learning applications and is capable of up to 20 TOPS. The module is centered on the 384-Core NVIDIA Volta GPU, which includes Tensor Cores for accelerated computation of AIs designed around Tensor Flow.
Jetson AGX Xavier is intended to be designed into robots and other autonomous machines that require modern AI throughput capabilities while existing “on the edge”. Jetson AGX Xavier can be used in endless applications such as manufacturing, service, logistics, retail, agriculture, healthcare, smart city, and much more.

Product Name	Jetson AGX Xavier 8GB Module
GPU	384-Core Volta GPU with 48 Tensor cores 5.5 TFLOPS (FP16) 11.1 TOPS (INT8)
CPU	6-core Carmel ARM v8.2 64-bit CPU, 6MB L2 + 4MB L3
Memory	8GB 256-bit LPDDR4x 1333MHz – 85GB/s
Storage	32GB eMMC 5.1
DL Accelerator	(2x) NVDLA Engines 4.1 TFLOPS (FP16) 8.2 TOPS (INT8)
Vision Accelerator	7-way VLIW Vision Processor
Video Encoder	2x464MP/sec,2x 4K @ 30 (HEVC),6x 1080p @ 60 (HEVC),14x 1080p @ 30 (HEVC)
Video Decoder	2x690MP/sec,2x 4K @ 60 (HEVC),4x 4K @ 30 (HEVC),12x 1080p @ 60 (HEVC),24x 1080p @ 30 (HEVC),16x 1080p @ 30 (H.264)
Camera	16x CSI-2 D-PHY 1.2 (Up to 40 Gbps),C-PHY 1.1 (Up to 64 GB/s)
UPHY	3x USB 3.1, 4x USB 2.0 1 x8 or 1 x4 or 1 x2 or 2 x1 PCIe (Gen3)
Display	Three multi-mode DP 1.2/eDP 1.4/HDMI 2.0
Other	UART, SPI, CAN, I2C, I2S, DMIC, GPIOs
Connectivity	10/100/1000 RGMII
Size	100 mm x 87 mm
Shipment Weight	0.4 kg
Shipment Dimensions	10 × 8 × 6 cm

More than 20X the performance and 10X the energy efficiency of the NVIDIA Jetson TX2.
Supported by NVIDIA JetPack and Deep Stream SDKs, as well as CUDA®, cuDNN, and Tensor RT software libraries.
Low power mode.