TensorRT samples Cross-Compilation for NVIDIA Drive Platform

How does one cross-compile the samples from https://github.com/NVIDIA/TensorRT for the NVIDIA DRIVE AGX platform?
I have found information regarding Jetson and QNX, but nothing for the NVIDIA DRIVE platform.

The instruction for cross-compile is available in README.md under /usr/src/tensorrt/src/ on DRIVE AGX.
Can you check??

Thanks

Thank you, appreciate your help.
I succeeded to run it on NVIDIA DRIVE platform.
The README file doesn’t have enough explanations regarding ssd.prototxt file (there is no ssd.prototxt file in tensorrt folder at all), but I resolved this issue.

I would like to cross-compile SSD example in Nsight eclipse for NVIDIA DRIVE platform.
I succeeded to build and run it on host (eclipse installed on host), but still don’t know how to cross-compile it for NVIDIA DRIVE.
Is there any document that explains how to do that ?

If I take tensorrt from NVIDIA DRIVE with its Makefiles, and compile in Nsight eclipse on host PC, it builds executable for host, and not for NVIDIA DRIVE. I don’t understand why.

Thank you,
Slava

The main idea is to compile our Tesnor floor network using TensorRT (to get inference engine), and then to integrate the received file into SSD sample, and run it. Using SSD sample, we would like to do assessment to our Tensor floor network.
In order to do that, I would like to work on Nsight eclipse.
I copied from NVIDIA DRIVE platform the folder /usr/src/tensorrt to my host PC with Nsight eclipse. Tensorrt folder already include Makefile and Makefile.config that build properly executable for NVIDIA DRIVE platform. But on my host PC, it builds executable for my PC and not for NVIDIA DRIVE (builds in different format). What should I update in order to build it for NVIDIA DRIVE, on my host PC ?
Tensorrt example folder doesn’t have any cmake/third_party folders. I try to run the existing Makefile, Makefile.config.