6. Jtop Tool Installation and Use

Installation of Jtop

(1) Installing JTOP to check CPU usage


    sudo apt-get update
    sudo apt-get full-upgrade
    sudo apt install curl
    sudo apt install nano
    curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py 
    sudo python3 get-pip.py 
    sudo pip3 install jetson-stats
    jtop

Check the installed system components

(1)Jetson's OS image already comes with JetPack, cuda, cudnn, opencv, etc. It has been installed and has examples. The installation path for these examples is as follows


x
TensorRT /usr/src/tensorrt/samples/
CUDA /usr/local/cuda-11.4/samples/
cuDNN /usr/src/cudnn_samples_v8/
VisionWorks /usr/share/visionworks/sources/samples/
/usr/share/visionworks-tracking/sources/samples/
/usr/share/visionworks-sfm/sources/samples/
OpenCV /usr/share/opencv4/samples/

(2) Check CUDA Version 11.4 of CUDA has already been installed in Jetson orin nx, but at this time, if you run nvcc - V, it will not succeed. You need to write the path of CUDA to the environment variable. The Vim tool comes with the OS, so run the following command to edit the environment variables

Firstly, check if there is nvcc in the bin directory of cuda:


xxxxxxxxxx
ls /usr/local/cuda/bin

IF EXISTS，


xxxxxxxxxx
sudo vim ~/.bashrc Enter the configuration file; Add the following two lines on the last side：

Note: In vim, use Esc to return to command mode, and switch to the input module through I to enter text in input mode


xxxxxxxxxx
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

Note: After exiting the command mode through Esc, press: to start inputting commands, wq to save and exit, q to exit, q! For forced exit

Save to exit.

Then it needs to take effect under the source.


xxxxxxxxxx
source ~/.bashrc

After the source, execute nvcc - V again at this time, and the result is as follows

beckhans@Jetson:~$ nvcc -V

（3）check OpenCV

Jetson already has OpenCV4.5.4 installed, you can use the command to check if OpenCV is installed properly

pkg-config opencv4 --modversion

If OpenCv is installed properly, the version number will be displayed, and my version is 4.5.4

（4）check cuDNN

CuDNN has been installed in Jetson nano and there are examples available for operation. Let's run the examples to verify the CUDA above

Enter jtop at the terminal, press the right arrow key on the keyboard to select 7info, and you can see the version of cuDNN, as shown in the following figure: