NVIDIA Jetson AGX Xavier

General

CPU	NVIDIA Carmel ARMv8.2 @ 2.26GHz
Cores	8
RAM	32GB
USB2.0 Interfaces	4
USB3.1 Interfaces	3
Ethernet	10/100/1000 MBit
PCIe	1x8 + 1x4 + 1x2 + 2x1 Gen 4.0

Note

The above table describes the specification of the NVIDIA Jetson AGX Xavier Developer Kit.

Benchmarks

GigE Performance

Test setup

Additional Settings Applied On The System

To improve the data transfer between the camera and the ARM device the following Kernel parameters have been modified:

In /etc/sysctl.d/62-buffers-performance.conf:

Note

You may have to create this file!

Setting	Value	Description
net.core.wmem_max	16777216	Maximum memory size of a socket buffer for sending in Bytes
net.core.rmem_max	16777216	Maximum memory size of a socket buffer for receiving in Bytes
net.core.netdev_max_backlog	10000	Maximum number of packets which can be buffered if the Kernel does not manage to process them as fast as they are received

NIC Settings

Setting	Value
MTU	8000 Byte

The importance of setting these parameters as above is explained here: Network Performance Settings.

Results

The following scenarios have been tested:

1. When de-Bayering is carried out on the host system: The camera delivers Bayer8 image data to the host system. The Bayer8 image data then get de-Bayered to RGB8 format on the host system. This setting results in a higher frame rate but a higher CPU load as well.
2. When no de-Bayering is performed: The camera delivers Bayer8 image data to the host system. No de-Bayering is performed. This settings results in a lower CPU load and a higher frame rate. The behavior is identical to monochrome cameras.

Camera	Resolution	Pixel Format	Frame Rate [Frames/s]	Bandwidth [MB/s]	CPU Load
mvBlueCOUGAR-X102mC	1600 x 1104	BayerRG8 (on camera) -> RGB8 (on host)	68	120.32	~22%
Camera	Resolution	Pixel Format	Frame Rate [Frames/s]	Bandwidth [MB/s]	CPU Load
mvBlueCOUGAR-X102mC	1600 x 1104	BayerRG8 (on camera) -> BayerRG8/Raw (on host)	68	120.32	~20%