Nvcc: command not found after installing nvidia-jetpack

Hi all,

I’m setting up Jetson AGX Orin Devkit for the first time without host computer following the instructions mentioned here. After the installation I rebooted and quickly wanted to verify if all the packages are installed correctly. But I couldn’t find any set of instructions for that. I tried to invoke nvcc command from the terminal but I keep getting command not found error. I’m able to access nvcc on my Jetson AGX Xavier which is on earlier version ofJetpack.

Any idea why the error is occuring on my Jetson Orin?

Also is there anyway to quickly check if CUDA + CUDNN are installed on the board. Thanks in advance for your comments!

I was able to resolve one part of my issue. I can access nvcc via its full path. /usr/local/cuda-11.4/bin/nvcc --version. I was also able to verify the CUDA installation by running some sample programs using/usr/local/cuda-11.4/bin/cuda-install-samples-11.4.sh .

Is it necessary to add cuda to PATH variable?

Yes, you may add these two lines at the end of your user’s home directory file .bashrc:

# Add CUDA bin directory into $PATH so that NVCC and others tools can be found
export PATH=/usr/local/cuda/bin:$PATH

# Add CUDA lib directory into the list of places for searching dynamic libraries
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

Next time you’ll open a terminal using a bash shell, everything should be fine.

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.