About setting up NVIDIA DRIVE AGX XAVIER

NVIDIA DRIVE AGX XAVIER setup videos (https://www.youtube.com/watch?time_continue=11&v=FhyjhC5rD5g) I have a question about 8:44. I am editing text using the vi command, but the file can not be found. How do you proceed to the next step?

Dear s18,

If you mean drive8-step.txt, the file is the presenter’s own file for this tutorial.
So for compilation setup for Drive AGX, please refer to Compilation Setup part in https://developer.nvidia.com/drive/learn

Thank you very much.

It is a new question to continue

  1. Is the host PC still needed as a development environment after initial setup?
  2. If possible, please tell me the specifications required for the host PC.
  3. About the scp command near 13:20 of the setup video (https://www.youtube.com/watch?time_continue=11&v=FhyjhC5rD5g), how can I know the address of nvidia@10.31.174.152?

Dear s18,

Could you please refer to “Getting Started” part in DriveWorks doc?

Q1, Q2,
Getting Started
NVIDIA® SDK Manager automatically installs on the NVIDIA DRIVE™ target platform and host systems the following items:
• NVIDIA DRIVE ™ 5.1 SDK/PDK
• NVIDIA CUDA® Toolkit version 10.0.166
• cuDNN 7.3.1
• TensorRT version 5.0.3.2
• cmake version 3.5
• gcc version 4.9
• NVIDIA® DriveWorks
• Libraries upon which the above items depend
NVIDIA DRIVE Platform Prerequisites
On the hardware side, you will need one of the following:
• NVIDIA DRIVE™ platform with the latest PDK flashed in the system and/or
PC (x86 architecture) with a NVIDIA GPU (Pascal or Volta based GPU).
x86 Host System Prerequisites

The Release Notes describe the software dependencies. SDK Manager installs these dependencies.
Note
You may also need to install (using apt-get install) the following packages:
• libx11-dev
• libxrandr-dev
• libxcursor-dev
• libxxf86vm-dev
• libxinerama-dev
• libxi-dev
• libglu1-mesa-dev
• libglew-dev
• libgles2-mesa-dev
Desktop development relies on NVCUVID for video decoding, which is included with the NVIDIA drivers. In general, the cmake build scripts can find NVCUVID installation. However, if this fails, you must set a symbolic link /usr/lib/nvidia-current pointing to your NVIDIA driver lib, for instance to /usr/lib/nvidia-410.

Q3, If you setup DriveAGX using ethernet connection, you can access your DriveAGX via ssh.

Thank you very much.
I do not know the Ethernet port connected to NVIDIA side well. Which Ethernet port should I connect?

Dear s18,

The IP address is just for the tutorial so you can’t access NVIDIA internal side.
I meant you could setup DriveAGX + ethernet port and access your DriveAGX via ssh from your host pc.

So, while connected via USB, should I check the IP address using SSH commands and rewrite it as setup videos?
I’m sorry I couldn’t understand it, but I’m sorry.

Dear s18,

If you connected via USB cable HostPC and DriveAGX, you can access via minicom tool.
Could you please provide your development environment like below?
(Ethernet <->)HostPC <-- USB cable --> DriveAGX(<-> Ethernet)

Internet <---- Wi-Fi ----> Host PC <---- USB Cable ----> Drive AGX Xavier
Currently, it is such a development environment.

Dear s18,

According to your development environment, there is no ethernet connection on NVIDIA Drive AGX Xavier.
So you should access HostPC via minicom tool from DriveAGX.

https://docs.nvidia.com/drive/active/5.1.0.2L/nvvib_docs/index.html#page/DRIVE%2520Linux%2520HYP%2520PDK%2520Development%2520Guide%2FUtilities%2Fminicom.html

I understand. Thank you very much.