Cross compiling in jetson nano

Autonomous Machines Jetson & Embedded Systems Jetson TX2
#1

I want to build the cross compiling environment in the ubuntu linux system. I had build the CUDA cross compiling. But can not build the tensorRT cross compiling environment.When i watch the NVIDIA TENSORRT DOCUMENT–Sample Support Guide, it say(web: Sample Support Guide :: NVIDIA Deep Learning TensorRT Documentation)：

  1. Install the TensorRT cross-compilation Debian packages for the corresponding target.
    Note: If you are using the tar file release for the target platform, then you can safely skip this step. The tar file release already includes the cross-compile libraries so no additional packages are required.
    QNX AArch64
  • libnvinfer-dev-cross-qnx
  • libnvinfer8-cross-qnx
  • libnvinfer-plugin-dev-cross-qnx
  • libnvinfer-plugin8-cross-qnx
  • libnvparsers-dev-cross-qnx
  • libnvparsers8-cross-qnx
  • libnvonnxparsers-dev-cross-qnx
  • libnvonnxparsers8-cross-qnx
    Linux AArch64
  • libnvinfer-dev-cross-aarch64
  • libnvinfer8-cross-aarch64
  • libnvinfer-plugin-dev-cross-aarch64
  • libnvinfer-plugin8-cross-aarch64
  • libnvparsers-dev-cross-aarch64
  • libnvparsers8-cross-aarch64
  • libnvonnxparsers-dev-cross-aarch64
  • libnvonnxparsers8-cross-aarch6
    Linux SBSA
  • libnvinfer-dev-cross-sbsa
  • libnvinfer8-cross-sbsa
  • libnvinfer-plugin-dev-cross-sbsa
  • libnvinfer-plugin8-cross-sbsa
  • libnvparsers-dev-cross-sbsa
  • libnvparsers8-cross-sbsa
  • libnvonnxparsers-dev-cross-sbsa
  • libnvonnxparsers8-cross-sbsa

and i can not find the deb packages. I had create a topic, and the staff replied can use SDKmanager to install. But when i use the SDKmanager, i also can not find corresponding deb packages. What should i do to build the TensorRT cross compiling environment! Please give me a hand!

How to cross compiling on jetson nano for cuda,cuDnn,tensorRT
#2

and my target is jetson nano 4G

#6

Hi,

Since TensorRT depends on much more packages than a CUDA app.
We usually recommend users compile it on the Jetson directly.

Please note that the TensorRT engine doesn’t support portability.
The engine is required to be generated on the Nano directly.

If cross-compiling is essential for you, please try it with docker that mentioned in the below commands:

Thanks.